Roche Holding-br Stock

My name is Leonard and I'm the technical operator for today's call. [Operator Instructions] One last remark, if you would like to follow the presented slides on your end as well, please feel free to go to roche.com/investors to download the presentation. At this time, it's my pleasure to introduce you to Bruno Eschli, Head of Investor Relations. Bruno, the stage is yours.

Thanks, Leonard. And could I have the first slide, please? So welcome to our second IR went in 2026, focusing this time on the Phase III ALLEGORY results for Gazyva in SLE, which just got presented last Friday at SLEuro. Today's call is scheduled for 60 minutes, and let me quickly take you through today's agenda.

Following my introduction, the first speaker for today will be Jay Garg, our Vice President and our Global Franchise Head of Nephrology. Jay will provide us an update on our pipeline in immune-mediated kidney and rheumatological diseases. This includes the Gazyva development program, but he will also quickly touch on other B-cell depleting agents currently in development.

Our second speaker for today will be Dr. Richard A. Furie. Richard is an MD and the Chief of the Division of Rheumatology at Northwell Health. He's a Marilyn and Barry Rubenstein Chair in Rheumatology and the Professor of Medicine at the Donald and Barbara Zucker School of Medicine at Hofstra in Northwell.

He is the lead principal investigator of the ALLEGORY study and the lead author and he has also presented the results last Friday at SLEuro. Today, Dr. Richard Furie will take us again through the Phase III ALLEGORY results for Gazyva in SLE.

After the presentations, we will have roughly 30 minutes of Q&A. In addition to our 2 speakers, we'll be joined for the Q&A by Stephen Wright, our Lifecycle leader for rheumatology and nephrology, which includes Gazyva and autoimmune diseases, but also sefaxersen in IgAN.

Could I have the next slide, please? So before we get started, I just wanted to quickly summarize here, take the opportunity to summarize the opportunity for Roche in immune-mediated kidney and rheumatological diseases as we have previously communicated with you the Gazyva peak sales opportunity in immune-mediated diseases to be around EUR 2 billion. This includes SLE, LE, membranous nephropathy and INS. And as you can see here on the left side, the U.S. EU5 Japanese market for SLE, lupus nephritis, membranous nephropathy and IgAN is expected to more than double from 2050 to 2030 from USD 5 billion to around USD 11.4 billion, driven by the launch of various new and innovative medicines with different underlying MOAs.

On the right side, I just wanted to provide a quick update on consensus expectations. You have seen this analysis back in September 2025 at Pharma Day. Since then a lot has happened and as we just got our latest consensus in last week, I wanted to provide here a quick update where we stand and what has been or has not been captured by the sell-side models. So let me start with Gazyva. Gazyva sales for 2030 are now forecasted to reach EUR 1.7 billion. That's EUR 0.7 billion more than 2025 and what you can see here is that the EUR 2 billion Gazyva immunology opportunity is still not really captured. Sefaxersen in IgAN, which is shown here in red on the right side and the potential upside has been, in the meantime, included in most analyst models with sales of EUR 400 million forecasted by 2030. Just to remind everyone, we have indicated here a peak sales potential in the range of EUR 1 billion to EUR 2 billion.

And there are more NMEs I quickly need to cover here as they have become material to the Roche investment case. Sales expectations for giredestrant have now increased from EUR 0.9 billion in 2029 back in September to now EUR 3.2 billion in 2030. So the adjuvant and metastatic opportunities are starting to get captured. And finally, sales expectations for fenebrutinib in multiple sclerosis have hardly moved since last September despite positive Phase III readouts in RMS and PPMS with sales expectations only increasing from EUR 0.7 billion for 2029 back in September to now EUR 1.1 billion by 2030.

So overall, we can see some initial adjustments to the peak sales consensus expectations. But clearly, there is additional upside once latest results will get presented at medical conferences. And as we are nearing the launch phases, but also as we will have additional data, for example, for giredestrant in first-line endocrine-resistant ER-positive breast cancer in 2027. So that's the pionERA study I'm referring to, where we have enriched for up to 40% in ESR1 muted patients. And where giredestrant is combined with the physician's choice of a CDK4/6 inhibitor.

And with that, let me quickly move to Slide 7. This is just the latest immunology pipeline overview. Jay will touch here on many of these programs, including the Gazyva program, sefaxersen and also the various other B-cell depleting strategies, including various bispecifics and CAR Ts which we are pursuing in parallel to hopefully achieve B-cell research for various chronic autoimmune diseases. Maybe 1 other program here to quickly call out that's afimkibart, our TL1A antibody, where we expect Phase III readouts in UC and CD in 2027.

We will be the second company coming here to market, with this novel MOA. And you also see that we test the afimkibart already in additional indications, including rheumatoid arthritis, atopic dermatitis and MASH and what you also see here that we have the first TL1A bispecific, the p40 x TL1A in Phase II development. And we just initiated another NME so far undisclosed NME and IBD last quarter, which is now in Phase I.

Before I close, one other molecule here to be called out. This is our NLRP3i inhibitor in asthma. And as you know, we have broader development programs ongoing in the NLRP3 space with studies also ongoing in neurology with Parkinson's disease and also in CVRM so in -- for cardiovascular diseases. And with that, I will hand over to the next speaker. Jay, please, over to you.

Yes. Thanks, Bruno. So my name is Jay Garg. I lead the nephrology, rheumatology group here in the late-stage organization.

Next slide. So as you know, immunology is one of our core therapeutic areas of interest at Roche, and immune-mediated kidney and rheumatologic diseases are a core part of that. So our ambition, our vision in this space is to introduce long-lasting remission in this area. And we did this in 3 ways: one, becoming a global leader in this space, really focusing on the highest unmet needs and transformational disease outcomes, delivering patient and health system value and then achieve this by using differentiated therapeutic approaches really with the goal of enabling long-lasting remission and freedom from toxic therapies. And this is really going to be led by our current portfolio in B-cell targeted therapies and complement inhibition.

You can go to the next slide. So I just want to emphasize that this -- these areas are not new to Roche. So we've actually been in this space for a long time, getting back to the introduction or approval of CellCept in 1995 for the prevention of organ transplant rejection. This was followed by Rituxan approvals in RA as well as GPA, MPA, Mircera for the treatment of anemia due to chronic kidney disease; ACTEMRA for the treatment of RA giant cell arteritis as well as SSc-ILD and then most recently, Gazyva in lupus nephritis last year.

You can go to the next slide. So in addition to Gazyva being approved in lupus nephritis, we actually had 3 additional positive Phase III studies. One in generalized lupus; one in membranous nephropathy; and one in childhood-onset nephrotic syndrome. In addition, we had a fourth positive Phase III readout with our anti-C5 monoclonal antibody crovalimab in atypical HUS.

Next slide. And so I just want to call out that CKD is really a core interest of ours. It has a significant unmet need, both to the patients as well as to the health care system. So CKD is forecast to become the fifth leading cause of death globally by 2040, and you can see the numbers there. In addition, it is a significant impact to health care systems. It takes about 24% of the annual U.S. Medicare budget and really cost EUR 140 billion annually in Europe. Immune-mediated kidney diseases are really a significant contributor to CKD and actually on a per patient basis have one of the highest risk for progression to dialysis or need for transplant.

Next slide. So lupus, which is what we're here today, it's really a chronic disease that affects nearly every organ system. So most patients will have mucocutaneous and musculoskeletal manifestations and renal involvement will actually develop in around 50% of patients. It's a disease of young women, mainly of color that causes irreversible damage due to really these chronic immune flares. And then the current treatment or the current aim is to really induce long-term remission with low-dose steroids. LN specifically can be considered the most severe organ manifestation of lupus. And it's not really that first episode of LN because that will cause damage to the kidney, but it's also subsequent flares.

So the goal of treatment is not only to induce remission or response in those patients, but really to prevent subsequent flares. So up to 3.4 million people are affected by lupus worldwide, again, with about 50% having renal involvement. And then about 30% of patients will actually progress to CKD if they do not achieve a complete renal response.

Next slide. And so when you look at the highest unmet need, so this is a survey done for KOLs. And so their thoughts are that what they need to see from a therapy, a really long-term efficacy and safety which is clear as well as steroid sparing. Now at the Euro lupus or at least the SLEuro conference this past weekend, there was a patient group that actually present their own survey and they actually queried care patients. And what they found was the biggest unmet need from a patient perspective was a treatment of fatigue, but also getting rid of steroids as background. So pretty significant for them and then comes efficacy, especially in the kidney realm as well.

Now biologics are expected to rise in the future, at least the use of them, really, this is driven by guidelines. So the guidelines now are taking more of a treat-to-target approach, where you are trying to get to disease remission. In lupus nephritis specifically, the recommendation now is to start with triple therapy, which is really steroids, mycophenolate mofetil and then a biologic therapy, either BENLYSTA or advanced therapy, BENLYSTA, voclosporin and now obinutuzumab, has also been included in those guidelines. You can go to the next slide.

So Gazyva. So the reason we believe Gazyva has been so successful in lupus and other autoimmune diseases is really because of its deep B-cell depletion that we've seen. So it's really the first glycoengineered anti-CD20 that's been evaluated in autoimmune disease. And what does this mean? It's a type 2 anti-CD20. So it binds differently than type 1 anti-CD20 antibodies like rituximab or ocrelizumab. So this leads to increased direct cell death, decreased complement-dependent cytotoxicity and reduced internalization. The Fc receptor has been glycoengineered. So it has greater affinity for effector cells, which then will lead to increased antibody-dependent cellular cytotoxicity and antibody-dependent cellular phagocytosis. There's actually good preclinical data in human data suggesting that this binding confirmation and the inhibition or the depletion of these B-cells is actually greater than therapies like Rituxan.

You can go to the next slide. So in terms of LN, just as a reminder, our initial Phase II data were in a study called NOBILITY where we showed superiority to placebo on the background of MMF and steroids. And this was then confirmed in our Phase III REGENCY study, where we showed superiority and complete renal response at week 76. In addition, we showed a greater proportion of patients getting to CRR on low-dose steroids as well as showing in an exploratory analysis, a significantly reduced time to lupus nephritis flare, where we had a 56% reduction in the risk of lupus nephritis flare.

As I mentioned, this has actually been included in guidelines so far. And the treatment regimen, just to remind you, is 2 doses at day 0 and week 2, repeat at week 24 and then week 26 and then it'll be 1 dose every 6 months thereafter. Importantly, we also got approval in both the U.S. as well as most recently, the EU for short duration infusion. And what this means is that after the first infusion if the patient has tolerated that well, they can then receive Gazyva and a 90-minute infusion as opposed to the typical 4-hour infusion, which we think will be a significant impact for both physicians and the health care system as well as the patient.

You can go to the next slide. So Rich will talk about this in more detail, but since it's already been communicated both at SLEuro as well as in the New England Journal simultaneous publication. As you can see here, this is a trial design where we compared Gazyva in a similar dosing regimen as we did with lupus nephritis versus placebo on top of background immunosuppression in patients with moderate-to-severe disease activity. It achieved its primary endpoint of SRI-4, and you can see here was a 23.1% delta, which was highly significant. It hit off 5 key secondary endpoints, which Rich will go into in more detail.

And I think importantly, it actually hit on some prespecified secondary endpoints around some higher bar endpoints like DORIS remission, which requires achieving a clinical SLEDAI of 0 on low-dose steroids 5 milligrams or less as well as LLDAS, lupus low disease activity. So the remission rate for Gazyva was 35% versus, I think, 14% for placebo. So kind of the highest we've seen so far in a global pivotal trial.

You can go to the next slide. Most recently, we had a press release on the Phase III study, MAJESTY, which evaluated Gazyva versus tacrolimus as calcineurin inhibitor and primary membranous nephropathy, which is known to be an autoantibody-driven disease. This was a 2-year endpoint, looking at really a high bar response for a complete remission, which was achieving a UPCR of less than 0.3 with stable renal function. And these are patients that typically start around 6, 7, 8 grams of proteinuria.

So pretty significant reduction. This is the most common cause of nephrotic syndrome in patients who are nondiabetic and about 1/3 of these patients will eventually progress to ESRD and require dialysis. Additionally, we reported last year in a press release around our Phase III study in childhood-onset nephrotic syndrome, where we met the primary endpoint. And again, the first global Phase III to study a targeted therapy in this disease. It actually received FDA breakthrough designation and has been accepted as a late-breaking oral presentation at the World Congress of Nephrology at the end of this month.

You can go to the next slide. So we -- our thought is now, given what we've seen with both Rituxan and now with Gazyva and seeing the success of Gazyva where Rituxan wasn't as successful, especially in a disease like lupus that B-cell depletion truly could be a very effective approach in this disease. External data, again, just case reports, case series with both T cell engagers as well as CAR T suggest that if you can get greater B-cell depletion that actually might lead to even greater responses and potentially even remittance.

We actually have a pretty strong pipeline in this area that will help us test the hypothesis. So we have a few T-cell engagers, T-cell bispecifics. We have Lunsumio, which is approved for follicular lymphoma already. This is a CD20, CD3 that can be given as an outpatient. The safety profile in follicular lymphoma seems to be appropriate such that we decide to take it into lupus and lupus nephritis. We've actually run a Phase Ib study, which we previously presented at EULAR last year, which showed an acceptable safety profile and did show B-cell depletion in this population. So we are now in a Phase II study with that molecule. In addition, our colleagues in pRED, our early development organization are evaluating an NME of a CD19, CD3, which is currently in Phase I in lupus as well. In addition to this, we have an allogeneic CAR T therapy targeting CD19, CD20 that we have initiated a Phase I study in both lupus and in MS.

Next slide. And so I'm happy to announce that we also have announced positive results of our anti-C5 monoclonal antibody in atypical HUS. So this is a devastating disease in children that can definitely lead to high morbidity and mortality. And it showed positive results. We have not -- this will be presented, I believe, later this year at a medical conference. And you can see that we have data both in adults and in pediatrics.

You can go to the next slide. And I think lastly, I just want to talk to you about our Phase III study in IgA nephropathy with our complement Factor B inhibitor. So sefaxersen is an antisense oligonucleotide that actually inhibits production of complement Factor B. So it inhibits it in the liver and actually had some local effects in the kidney as well based off of what we've seen in preclinical models. And our thought is that by actually inhibiting complement Factor B, we could have a greater effect in IgA nephropathy in terms of the complement-driven inflammation that this has. We actually have Phase II data that has been recently published in Kidney International reports where we showed that there was in a single-arm study reduction of 24-hour proteinuria.

And then this led us to conduct the Phase III IMAGINATION trial where we're comparing sefaxersen, against a placebo in background on optimized background standard of care with a 9-month endpoint for proteinuria to hopefully lead to accelerated approval and then a 2-year endpoint on eGFR. You go to the next slide.

And with that, I will turn it over to Rich.

All right. Thanks, Jay. I'm Rich Furie, I'm the Head of Rheumatology at Northwell, which is a 30-hospital health system based in New York. My passion is to take care of lupus patients and in the early '90s got involved in couple of lupus trials because that's all we had back in the '90s. And now we have a huge number of studies. All right. I have 6 minutes to present this, but I understand today, I can take a little bit longer, and I started out just to try to break the ice talking about ALLEGORY being a misnomer because remember, go back to your high school English days and ALLEGORY was a fictional piece of work. It was a fictional story with a lot of symbolism. And this is the polar opposite of that, pure facts.

All right. Next slide. Disclosures, yes -- so just a little bit of background. And Jay went into this. I mean I could spend an hour on this because it's such an interesting story. If you go back to our early efforts to deplete B-cells and see how patients with lupus and lupus nephritis did. What we learned is that we weren't depleting enough. More is better. And so obinutuzumab, and I'll go over this again, is a far more potent B-cell depleting antibody than our predecessors.

Next slide. And I really like this slide. It's so simple. If you look at the left, you'll see the modifications made to the molecule. So it is glycoengineered and so what does that mean? It means it was defucosylated, so these fucose groups were removed. And the end result there is that there's less steric hindrance and greater enhancement of binding of the Fc to an Fc receptor. But there are also modifications done to the front end of the molecule, a so-called type 2 binding confirmation. And then if you shift over to the right side of the slide, you'll see the results. We have greater antibody-dependent cellular phagocytosis. And at the bottom, antibody-dependent cellular cytotoxicity is increased as well. Because of the type 2 binding configuration, there's greater direct cell death and less dependence on complement-dependent cytotoxicity. This all adds up to just greater B-cell depletion.

Next slide. Here's the design of ALLEGORY similar to other studies done in lupus. So remember, when we study lupus, we usually have a split between SLE and lupus nephritis. So different treatment algorithms and somewhat different treatment designs. So in ALLEGORY, there were 303 patients who were randomized 1:1 to either receive obinutuzumab or placebo. On top of their standard of care therapies, and we'll go into the standard of care in a little bit. There were a total of 4 infusions over the 52 weeks. The first 2 infusions constituted 1 course and that was done at baseline, and then there was second course, it was done at 6 months.

With each infusion, whether a patient got obinutuzumab or placebo, they received a methylprednisolone intravenously 80 milligrams to prevent any kind of infusion reaction. The primary endpoint was the SRI-4 at 1 year. And we could go into that maybe afterwards if you're not familiar with the SRI-4. The SRI-4 made its debut, I actually presented, as I recall at EULAR, some almost 20 years ago, but it's been along with BICLA as an endpoint, the regulatory endpoints for our SLE trials.

There are a lot of key secondaries. I'm not going to go through them right now. I'll show you on the next slide when I show you results, but we're not ready for the next slide. I'm sorry. I just want to point out a few more things. So prednisone. Prednisone has been a confounder for our studies. And we know in clinical practice, it's so important to taper prednisone. So we do that as part of our trials and patients who are on prednisone, which is a good chunk of these patients were required to try to taper to 5 milligrams or less by week 24, but then from the -- from week 40 till the end of the study, so the last 12 weeks of the study, they cannot change their prednisone dose. And probably one of the most important features of this particular slide is on the bottom right. This study enrolled very high disease activity patients. So what is a high disease activity patient? We required a SLEDAI of 8 or greater.

Almost all other studies, the bar was at 6. We required a BILAG A or 2Bs. So just another measure of disease activity. So the patients had to achieve -- had to have both and then, of course, a Physician Global Assessment of 1 or greater. But in addition, they had to be autoantibody positive. So that meant an ANA or an anti-DNA antibody or yet another antibody called Smith. And what was very unique to this study was the requirement for a low complement level, a low C3 or low C4 or low CH50. So these requirements added up to a very active group of patients.

Next slide. Baseline characteristics. I'll just go through a few of the points here. If you look at what's bold, mean SLEDAI score of a little over 13%. That's a very active population. Most of our SLE trials, the average SLEDAI score has been around 10, maybe up around 11. And then what's a little different from other studies also bold, background therapies, background immunosuppressives, about 70%, usually in our studies about 50%. And in fact, I think, the percentage using hydroxychloroquine was a little bit higher. In most of our studies, it's around 2/3 to maybe 70%. This was a little higher.

Next slide. All right. So the primary endpoint, the SRI, and it is defined on the left-hand side. So if I have some time, I will go through this. The SRI in order to be a responder, and it is a dichotomous endpoint. One has to have a 4-point or greater reduction in the SLEDAI score. And they can't acquire new activity in the form of a BILAG A, which is the most active domain score that one can achieve and/or they can have more than BILAG B. There can't be deterioration in baseline Physician's Global Assessment and obviously, no occurrence of any kind of intercurrent events like changes in medicines and so forth.

So is that a rigorous endpoint? It really is. When we first developed this, no one really understood what an SRI was, but it turns out it's a very rigorous, hard to achieve endpoint. But if you look at the right-hand side, you see that 76.7% of patients on obinutuzumab achieved this endpoint whereas in the placebo group, a little over 53% achieved this endpoint, accounting for a treatment difference of 23 percentage points. How does 76.7% compare to other studies? It's the highest we've ever seen. And I must say I can editorialize now when I saw those numbers, my first word was wow.

Next slide. Let's go through the secondary endpoints. There were 5 key secondary endpoints that are listed on the left. The short of it is that every single endpoint and we've never seen this. Every single key secondary endpoint hit statistical significance with treatment differences ranging from about 22 percentage points up to about 30 percentage points. But let me just run through them since I do have some time to explain them.

BICLA is our other regulatory endpoint and you can see about a 22 percentage treatment difference there. Glucocorticoid reduction to 7.5 milligrams or less that was sustained for the last 12 weeks of the study. I emphasize the importance of steroid reduction, also significant. The SRI-4 at week 40 sustained, so that not only speaks to efficacy, it speaks to durability. Now you might counter or people counter, but we know anti-CD20s can improve complement, they can improve DNA antibodies. And if you normalize both of those, you lose 4 points off your SLEDAI. So was it just about complement and anti-DNA antibody? No, because if you look at the next 1 down, SRI-6 requires a 6-point reduction. So now you have to introduce clinical improvement as well. And then the last one, time to first BILAG flare.

I mean the name of the game in lupus is to prevent damage. And where does damage arise? The damage arises from flares, from activity. So it's very important for a drug to be able to reduce flare rates. And there's a whole slide dedicated to that, that I'll show you. And we talked a little bit about additional secondary endpoints. These are not in the testing hierarchy, but I have a slide, and we'll go through DORIS and LLDAS in just a couple of slides.

So next slide. Time to first BILAG flare and that BILAG flare is defined as Ia or IIb that arises. And this, again, is very important. And you can see a hazard ratio of 0.58, which translates into 42% risk reduction of flare. And again, I can't tell you how important flares are because that's how patients accrue damage. And for each 1 point, we're not going to go through the SDI damage index, but for each 1 point increase in damage, there's about a 30% increase in mortality risk. So damage is very important.

Next slide. SLEDAI scores over time. So the main driver of SRI response is a reduction in SLEDAI. SRI-4 means a 4-point reduction. And this just gives you an idea of the kinetics of response. Sustained reduction in SLEDAI occurred about week 24, but you see there are certainly some separation even before week 24.

Next slide. All right. DORIS and LLDAS. So what are they other than cute names? Well, when patients not even that long ago, 6 or 7 years ago, asked me, "Doctor, am I in remission?" I said, we don't really use that word in lupus. We talk about maybe control of disease. But now we use that word. And so DORIS remission has a definition. It's pretty stringent. And you can see that about 35% of the patients who received obinutuzumab achieved DORIS remission by week -- or at week 52 compared to about 14% on placebo.

That's -- I can tell you that, that's a very high number, 35%. It's a very stringent definition. LLDAS, lupus low disease activity state, achieved by a little over 57% in the obinutuzumab group versus 25% in the placebo group. LLDAS is not as stringent as DORIS And it may be worthwhile just taking a couple of moments to talk about, well, how is DORIS and LLDAS. How are they different than SRI or BICLA? So SRI is improvement in activity by, as I mentioned, 4 points, whereas DORIS and LLDAS are absolute, you have to hit a particular state. And the state that of relative well-being is associated with a lot of benefits. Reduction in mortality, a reduction in damage accrual, less health care utilization and the list goes on and on. So to clinicians, DORIS and LLDAS are far more important in practice than say to SRI and BICLA. I mean, they're all important, but DORIS and LLDAS really do stand out.

Next slide. B-cell depletion in peripheral blood. An interesting slide that compares the degree of B-cell depletion in this particular study to what we've seen in the EXPLORER study, EXPLORER, we did many years ago. It was a comparison of rituximab to placebo on background therapy. In this particular slide, dark blue are the obinutuzumab-treated patients and gray represents the rituximab-treated patients. And every single pair shows that more patients are achieving a threshold of B-cell depletion, which was 10 cells per microliter with obinutuzumab than with rituximab.

So not only a greater percentage depletion across those patients, but it would seem to be more sustained. But this doesn't tell the whole story. We need to know what's going on in the tissues and now that wasn't evaluated in this particular study, it was evaluated in the lupus nephritis regency trial and the data were quite impressive that we were achieving deep B-cell depletion in the tissue of kidneys from patients with lupus nephritis.

All right, next slide. What about safety? So there were a total of 4 deaths during the first 52 weeks, one in the obinutuzumab group, 3 in the placebo group. And I could spend a lot of time going through this, but let me just kind of shorten it, and we'll go through some of the bold items. Infections were, in fact, more common in the obinutuzumab group than in the placebo group. So we've kind of seen mixed results across the 3 studies. If you go back to the NOBILITY lupus nephritis trial, there are actually more infections in the placebo arm. The REGENCY was the opposite and here we did, in fact, see more manageable infections like urinary tract infections, respiratory infections, pneumonia. So more common, but very manageable.

We can jump down to infusion-related reactions, more common with obinutuzumab, but these are like nausea and headache and lastly, drug-related neutropenia. What we've been finding with basically, any strategy to deplete B-cells is that patients, and we don't quite understand why can't become neutropenic, there were slightly more patients in the OB group that developed neutropenia than in the placebo arm. And let's go to the next slide, which is the last slide. I'm not going to belabor all these conclusions because I just went through them.

But the most important one is at the bottom. This was a very active cohort of patients. So to be able to achieve a relatively high rate of DORIS remission to show that the drug is steroid sparing and that it reduces flares. This is what is so important to the clinicians and also to the patients.

Next slide. So I did want to thank all the study participants, not only in this study, but this has been a journey that started about 20 years ago with the rituximab studies, the ocrelizumab studies in lupus, and we really do owe it to the study participants to their families and the study investigators get a little credit and their staff. And this was -- simultaneously appeared online at 1:45 Lisbon time on what was it, Friday, so you can look at the article. It has a little more information than what I just went through. And that's it for me.

[Operator Instructions] The first questions go to Simon Baker from Redburn.

A couple of questions. Firstly, Dr. Furie, looking at the subgroup analysis by SRI-4 scores, it looks like a very consistent result across the piece. So 2 questions. Firstly, are there patients where you would not consider using this. And secondly, the one area where there seemed to be an enhanced performance was in people of Hispanic or Latino ethnicity. I just wondered if you had any thoughts on why that was?

And then probably a question for Jay. On cell therapy, there's been a lot of activity in the broader autoimmune CAR T space. Just intrigued to see why you're going straight for allogeneic rather than autologous. Is that -- any specific reasons? Or is this really about scalability of the opportunity that allogeneic gives you? Any thoughts? Much appreciated.

All right. So first question, who would I not use it on? Well, there's still a lot of data that we need to see. I mean lupus is not 1 disease. Lupus is about 1,000 diseases. And we treat by interviewing the patient, I mean, we're so old fashioned in rheumatology. So we interview, we examine and we look at labs. And some of the major manifestations of SLE are skin rash and arthritis. I haven't seen the domain-specific data for OB just yet, nor the kinetics of domain-specific responses. But I must say Anifrolumab works very well for skin. It works very quickly. So in all honesty, I think it's going to be hard to compete with Anifrolumab in patients with very active skin disease. But again, that's without seeing the OB data.

Other than that, patients who are serologically active with arthritis, patients who have a little bit of nephritis and patients who have a lot of nephritis, they'll get obinutuzumab. So I can't really think of contraindications to giving someone obinutuzumab. Plus it's very convenient to infusions, and we have a problem with noncompliance and nonadherence. Patients don't take their pills. And so to be able to have them come in for an infusion and the infusion duration will be decreasing and then just have them come back 6 months later. Now it's hard to beat that.

You asked about ethnicity. Well, you asked about subgroup analyses. So I have such mixed feelings about subgroup analyses because they're hypothesis generating. And if you took them at face value, you probably won't use a lot of these drugs, belimumab, Anifrolumab or obinutuzumab in certain groups. So I can't really explain why the differences. It certainly would be interesting to find out. But again, I think, we just have to think about mechanisms, but not take them at face value.

I'll answer the question, maybe Bruno, you can chime in as well. So it's a good question around allogeneic versus autologous. And if you remember the first slide of what we're trying to achieve for both patients and the health care system. An allogeneic off-the-shelf therapy if we can -- if it does demonstrate the efficacy, we are expected to see with autologous or what we have seen that would be just from a convenience or accessibility, more so a factor that it would seem like it would trump the autologous. So that's our main goal. And the data that we've seen with the allogeneic suggests that this is something that we might be able to achieve.

Next question would go to Luisa Hector from Berenberg. Luisa dropped off for some reason. Let's quickly then continue with Graham Parry.

So two on Gazyva, and then I've got 2 on persevERA that we just been it would be good to address. So on Gazyva, just thoughts on commercialization with an IV administration only given there is a subcut BENLYSTA. I think it's about half of patients actually get that. And AstraZeneca has obviously refiled now its subcut formulation there. So is there an advantage to subcutaneous administration? And perhaps Dr. Furie can actually chime in on that one as well.

And then secondly, do you expect to see the remission data on label, the DORIS remission data on label, given again, Dr. Furie indicating the importance of that? And then just the 2 on persevERA, the question we've been getting all day was just was that missed due to powering. So did you beat the target 0.85 hazard ratio just missed on [indiscernible]. So trial -- powering the trial was fine? Or do you just see a smaller effect size with a 0.85 hazard ratio? And do you have a sense of whether that fail was due to not enough naive patients in the study that were endocrine sensitive? Or does this validate the theory that CDK4/6s negate the advantage of an oral [indiscernible] in combination?

Okay. Maybe I'll do this in between oncology question first, Graham. It's the big topic of the day. So yes, what we've seen -- we had powered when we were looking for getting a hazard ratio of 0.85 and the effect size was below. We have indicated in the press release, there was a numerical benefit. There is, I think, no more I can say right now at this point, but I think let me also remind you that there was a limited number of ESR1-mutated patients in this study, so around 5%. So as said, there was a bit of a benefit seen.

Does it -- what does it mean in terms of CDK4/6 combination development? I think we have indicated in the press release that we believe that there still is a viable path forward in the adjuvant setting. And we -- I would not agree -- what you have asked me basically, does this mean that there is no additional benefit to be gained in combining giredestrant with CDK4/6? Is no, I would not draw this conclusion from the data we have seen. And did I miss anything on your -- what else, Graham?

No, I think you covered there. Thanks, Bruno.

Okay. Then let's go back to Gazyva.

I've forgotten the questions. I guess the first one was subcu versus IV. Is that right?

Yes, exactly. So it's just the importance of -- potential importance of having a subcu just given I think it's about half of the BENLYSTA patients subcu and AstraZeneca filed for an indication because it has only got IV. And so if you could just talk through the dynamics of that would be great.

Yes. So I'm spoiled. We have an infusion unit. It's about 50 feet from where I'm sitting right now. So we can offer it either. And it's kind of open discussion, shared decision-making as they say. But IV every 6 months, I think, is pretty easy, and it guarantees that the patient gets the drug. I don't know about adherence and compliance to subcu. But I could say advantages to -- I mean I can argue both sides for this. So I don't really have a definitive answer for you.

It's more about how effective the drug is and what the -- I mean that's -- and then we have other variables. We had to deal with insurance companies and what they're telling us, how to give it and then resources as well.

And maybe I can take the remission question because this is something we'll discuss, right, with health authorities. I do think just based off of what Rich was saying, the importance of remission to physicians and to patients and in the guidelines seem to be something that should be shared, right, in the label. So -- but that's a discussion we'll have health authorities. So it is not Type 1 error control. So it wasn't of the key endpoints, but it is -- it was prespecified.

Yes, I was just thinking of one question, which came here in the chat from Michael Leuchten from Jefferies, maybe also to access and the question is about would you expect that there would be a step through, BENLYSTA first and then Gazyva. Or how would you expect this to develop?

Well, we can distinguish a lupus nephritis from SLE. I'll start with lupus nephritis. No, I don't see any step through at all there. And for SLE, I don't see that either unless it's dictated to us by insurance companies. But clinicians will want to use obinutuzumab upfront without having to fail other drugs.

Next one is James Quigley from Goldman Sachs.

My question, I've got 2, please. So firstly, maybe for Dr. Furie. So as you think about your typical lupus population given the inclusion criteria and activity levels, what proportion of your SLE population would this trial represent? And similarly, when you think back to other studies, many of which have failed in the past. Did you see any differential benefits between the higher risk groups in the SLEDAI versus the lower risk group. It doesn't seem to be the case here, but the cutoff used in the subgroup analysis in the New England Journal was greater than or less than 12. So interested to see if you see any differences there with other trials, which maybe had lower activity, including criteria?

And secondly, for Roche on the bispecifics, you've got one targeting Lunsumio, CD20, you also got CD19. So what are your expectations here for differential B-cell depletion and then also the -- why you got different approaches for each of those assets? And what are your expectations of differentiation between Gazyva, Lunsumio and then your CD19, CD3?

Okay. Let me see if I remember the question. So the first question was what percentage of, say, my cohort here in practice? Has that kind of SLEDAI, you mean a SLEDAI of 13?

Yes, yes.

Very few unless they have lupus nephritis. But that's one of the problems with SLEDAI. It doesn't tell the whole story. You could have someone with terrible, terrible arthritis or terrible, terrible rash. So the arthritis patient picks up 4 points and the rash patient picks up 2 points. You might want to use this drug in those patients, even though they don't have the entry criteria that ALLEGORY had. So sometimes they're held to this by insurance companies, for example, with belimumab. They go, well, the patient -- in order for you to get belimumab approved, the patient has to have a SLEDAI of 6. And then we argue back, that was for the studies. We're talking about practice.

So not everybody that we see has low complement. That was part of the criteria to get a high disease activity group. Anyway, so there's discordance between what we actually can see in practice and call severe versus what is severe in a study. I hope that kind of gets to your question. Then you had one about studies with lower disease activity.

Yes. So did you see differential efficacy between the lower disease activity groups and the higher disease activity groups in other trials or maybe the entry criteria was for lower disease activity. Did you see the -- do you see benefits in the higher disease subgroup of those trials?

Well, this trial, as I recall, I mean, you can look at the subgroup analysis in the New England Journal, but as I recall SLEDAI. So the partitioning threshold was -- for entry was a SLEDAI of 12. And as I recall, those above 12 and those less than 12, performed about the same. I mean I have to go back and look at the paper, but that's how I recall it. For other trials, with average SLEDAI, I say, 10 to 11, I have a lot of data swirling in my head. I'd have to go back and look. But so look at the subgroup analysis for this study, I think your question will get answered. Jay, help me out.

No, you're right about our data. I don't know the other data as well. So just a question on our T-cell bispecifics. So it's a good question. So we don't know what the best target antigen is in all honesty. So we know that targeting CD20 works as we've demonstrated with Gazyva. Are the hypothesis we're testing with Lunsumio is whether greater B-cell depletion targeting CD20 will get you great efficacy? So we have preclinical data head-to-head showing that a CD20-CD3 is superior to a monoclonal antibody targeting CD20 in terms of tissue B-cell depletion. So that's the question here that we're trying to answer with Lunsumio.

Again, it's approved. We know what the safety profile is. We have some initial data on lupus. So we feel interested in understanding what it can do. But the question around CD19, CD3 is a different target antigen that has a broader B-cell expression profile So this was the data that Georg Schett has presented with this CAR T CD19. It does hit some plasma cells that a CD20 might not hit. So that's a question of what can that do? Can it actually induce some more remission-like profile than what you could get with CD20, CD3. So we're testing kind of 2 separate hypotheses there and understanding what each of these can do. I hope that answers your question.

Next one is Luisa Hector from Berenberg.

I'd love to hear from Dr. Furie. Just thoughts on why the patients are still flaring, what is the biology and how we might eventually get better control on that. Does it link to simply the B-cell suppression or that alone is not enough? And in the New England paper, there was some commentary around the response rate to some extent being attributed to modern refinements in trial design. So I'm curious to know what that is. And does that mean the older drugs run with the same refinements would also look better perhaps?

You all ask difficult questions. So for the flare, I don't have an answer for you, but that's something that will be analyzed. I mean -- you have to -- you could look at the definitions in BILAG. You can go through what Ia means and b means. But I don't have an answer for you for these data, but hopefully, we will soon.

All right. The modern refinements is actually very interesting. I mean lupus is one of the hardest diseases to study. I mean, we go back to the belimumab Phase II, that's where we learned our lesson. Not all the patients were serologically active. So each study thereafter, patients had to be serologically active. And then we learned something more than that in the next study and the next study. So we have made major refinements that I think allow us to say that our patients who were enrolled in the study probably truly do have lupus and that the their activity indices are not overinflated. And then there are some other things that we've added like steroid taper and so forth. So it's a very interesting question. I mean, if we were -- if we went back and studied rituximab again, would we see better results? I don't have an answer, and we're not going to do it.

So Rich, I just want to chime in because the argument I would make because one, you participated in EXPLORER. So you know that was unfortunately, pretty negative, where we missed all primary and secondary endpoints. I think the question here, especially with the low complement population where Rituxan actually depends on more complement-dependent cytotoxicity. We have this type 2 confirmation, which deals with -- or depends less on complement-dependent cytotoxicity. I think in this particular population, it might be hard to say. You may have seen something in some patients with Rituxan. The question is, would you have seen what we're seeing with Gazyva? I don't know. But it seems less likely.

But it is very possible that we had successful drugs or "successful drugs" in the past, but the study design didn't allow us to see that.

Could I just ask then, so I get more excited to see drugs with even deeper B-cell depletion or perhaps new mechanisms or combinations, like combinations for the patients might get more -- or raise the efficacy bar more than the B-cell depletion?

Well, that's a good question. So yes, if you go back to the EXPLORER and the LUNAR trials, the message there seemed to be more B-cell depletion is needed. So that came from some observational studies up in Leeds by Ed Vital in lupus and in RA and then post hoc analysis of the LUNAR trial, anyway, there was a lot of evidence accumulating that more is better. So how do you achieve that? Well, there were a series of studies done with sequential belimumab and rituximab. And there were 2 strategies behind them. When we did the Phase I study with belimumab and by gosh, we're going back over 20 years. What we saw was actually an increase in memory B-cells after the delivery of belimumab. And so some of those combination sequential studies were based upon giving belimumab first to mobilize these B-cells come in with rituximab to get better B-cell depletion.

There was yet another strategy that was employed to justify the sequential belimumab-rituximab studies. And that was when you wipe out B-cells, and it doesn't matter how you do it, rituximab, obinutuzumab, cyclophosphamide, you still have a lot of BLyS or BAFF being made, but it has nowhere to bind. But as the B-cells return, it's now a high BAFF environment, which may lead to a lot of autoreactive B-cells. But those studies were done. And they were done in SLE. They were done in lupus nephritis and nothing sort of very major. I mean there's a little bit of pharmacodynamic benefit. I think the last one, I forgot what it was called BLyS, B lupus or something...

Believe or...

No, it's less believe, but the one after that done by the fellows in London, I think that was -- had a lower -- saw a lower flare rate. So when I lecture about this, I do talk about those 4 studies, but then I say I introduced obinutuzumab. And you have B-cell depletion data. You have at least 2 studies that I know about. There was one that the folks at Genentech conducted in patients undergoing kidney transplant who had received obinutuzumab because they had been previously sensitized. Not lupus nephritis just -- and as I recall like some 20-some-odd patients, only 2 patients had -- or 2 to 3 patients had B-cells in their lymph nodes that they harvested at the time of kidney transplant. And then you have the 64 patients in the REGENCY lupus nephritis trial who underwent second biopsies. And they are very impressive data, and you really should take a look. They were presented at ACR, ASN and then again in Lisbon last week.

And we got the final person in the queue, this is Steve Scala from Cowen.

Three short questions. First, doctor, given the strength of these data, will you switch your BENLYSTA patients to Gazyva. Secondly, should the oral products be successful in Phase III, such as RINVOQ and Sotyktu. How will they fit into the treatment scheme? And then lastly, for Bruno, could you tell us if the persevERA trial showed an OS trend?

Tough, very unanswerable questions. Will I switch a beilimumab patient who is doing well? No. Will I switch a beilimumab patient who is not doing well? Absolutely. You asked about how will the orals affect. I mean it's a problem. We're sort of entering the same kind of dilemma that rheumatoid arthritis group has had. We were yearning for all these new drugs, and now we have them, we need to know which drug, which patients.

So it's going to be dictated by -- I don't think it's going to be dictated by comparing SRI and BICLA responses. I think we'll want to see domain responses. We'll think about safety. We need some kind of way to predict who's going to respond to what? And then on top of that, we have insurance companies who are going to tell us what the high -- at least in the United States, what the hierarchy will be. So I really don't have an answer for you. And then we have the patients who are prescribed pills and don't take them. That's a big problem.

So with hydroxychloroquine, we've had available hydroxychloroquine blood levels in the United States for a lot of years. I can't tell you how many patients. It's like Nathalie Costedoat told us years ago. Patients don't take their hydroxychloroquine. So if they're not taking their hydroxychloroquine, they're probably not taking all their mycophenolate or their deucravacitinib or their upadacitinib and you can go down the list. So there's something to be said for a parenterally-administered drug.

Doctor, in response to your answer, could you tell us what percent of your BENLYSTA patients are not doing well. And Bruno, if you could answer the persevERA question, that would be helpful.

Unfortunately, I have to tell you, I also have no answer for you, Steve. So you're going to have to wait for the data presentation that we plan to have the data at ASCO.

I don't have data for you. I would say maybe 30%, 40%. I mean, with BENLYSTA, you have to be very patient and tell the patient that as well. So I won't declare someone a failure until they've been on at least 6 months and up to 12 months. But yes, I should probably take a look at our switch numbers.

Very good. I think with that, we are at the end of today's event. I would like to thank again all of our speakers for their time and also their efforts in exploring here new treatment options for patients with severe autoimmune diseases.

Let me also thank the IR team members who prepared today's event after call out here, Julia Brauer and Sabine Borngraber and also Melanie Wolf for event organization. I hope the event was helpful to provide a timely update here on our immunology franchise. The next event, which we plan now will be for fenebrutinib, the FENhance 1 and 2 data in RMS, which will be presented at AAN mid of April. We'll send out the invite as soon as we have a date fixed.

And with that, I think, if there are any remaining questions, then please reach out to the IR team. We are always happy to assist you and follow up.

My name is Henrik, and I'm the technical operator for today's call. Kindly note that the webinar is being recorded. [Operator Instructions]. If you would like to follow the presented slides on your end as well, please feel free to go to roche.com/investors to download the presentation.

At this time, it's my pleasure to introduce you to Bruno Eschli, Head of Investor Relations. Bruno, the stage is yours.

Thanks, Henrik. Could I have the first slide, please? So welcome to our post-IR event in 2026, focusing on the Phase III FENtrepid results for fenebrutinib and PPMS, which just got presented last Saturday at ACTRIMS. Today's call is scheduled for 60 minutes. However, in case we would need a bit more time, we can go over just to make sure that we can take all your questions. Let me quickly take you now for today's agenda.

I will start today by providing a quick one slide wrap-up on our Ocrevus franchise, before we will dive into our MS pipeline and the latest exciting data presented at ACTRIMS. Following my introductory remarks, the first speaker will be Hideki Garren, Senior Vice President and our Global Head of Product Development for Neurology. Hideki will provide MS pipeline update, focusing on fenebrutinib and how this molecule differentiates from other BTKs in late-stage development.

He will also provide a summary on the Phase II RMS data for fenebrutinib, which have been very consistent with the Phase III FENhance 2 data, which we plan to present at an upcoming conference once we have FENhance 1 results in-house. Hideki will finish his presentation with highlighting also our early-stage clinical pipeline in multiple sclerosis, where we explore new drug modalities and new drug targets.

Our second speaker for today will be Dr. Amit Bar-Or. Amit is an MD, FRCPC, Director of Center for Neuroinflammation and Experimental Therapeutics and Chief of the Multiple Sclerosis Division of the Department of Neurology Perelman Center for Advanced Medicine at the University of Pennsylvania. He is a member of the [indiscernible] Study Steering Committee, and he is the lead author and presented the data last Saturday at ACTRIMS. Dr. Bar-Or will take us again for the Phase III FENtrepid results for fenebrutinib in PPMS.

Afterwards, we will have a Q&A session. In addition to our two speakers, we will be joined for the Q&A session by Alexandra Goodyear She is an MD and the lifecycle leader for fenebrutinib within the global product strategy team for neuroscience. Could I have the next slide, please? So before we get started, I just wanted to quickly summarize the latest information provided on our Ocrevus franchise. I thought this makes sense as with fenebrutinib, we are now really excited that we potentially will have another medicine available for multiple sclerosis patients. Following Ocrevus, which was the first ever CD20 launched back in 2017 and which revolutionized the MS treatment paradigm.

So as you can see on the left side, Ocrevus remains the #1 new-to-brand medicine with 450,000 patients being currently treated. It's important to keep in mind that Ocrevus is the first and only twice yearly anti-CD20 approved for both RMS and PPMS and is available for IV and subcutaneous administration. When Ocrevus was launched in 2017, it redefined the treatment paradigm by establishing high efficacy B cell depletion as a foundational standard throughout multiple sclerosis. And one additional comment here as Ocrevus is the only approved CD20 antibody in PPMS and since the Phase III FENtrepid trial compared fenebrutinib with Ocrevus.

This is really the only clinical data that generated directly comparing a BTK with CD20 antibody and therefore, really delivers unique insights on how these two different mechanisms of actions compare to each other and might get positioned in an evolving future to treatment paradigm. I'm sure we'll touch on this later in the Q&A session. On the right side of the slide, I quickly wanted to summarize the latest developments for our Ocrevus franchise. As you know, at year-end 2025, we had more than 17,500 patients on Ocrevus Zunovo, and we saw a key acceleration in the uptake in Q4, something we have been waiting for since we got the J code in April last year.

As you know, Ocrevus Zunovo expands use via a 10-minute subcutaneous injection, which is more convenient and more convenient delivery option for patients and practices. We have previously communicated that we see Ocrevus Zunovo as a EUR 2 billion incremental sales opportunity by 2029. But of course, there will be more than EUR 2 billion in Ocrevus Zunovo peak sales as there will be also switching from IV to subcutaneous occurring.

What really excites us about the momentum we have seen now in Q4 is that more than 50% of the total global growth was driven by Ocrevus Zunovo and that more than 60% of the subcutaneous volume is coming from community practices, with around 50% of subcutaneous starts being new-to-brand. That's exactly what we had hoped for to see an increased demand outside of the larger IV treatment centers where Ocrevus has well established as the leading medicine. And now we are looking one step ahead in terms of future development steps for Ocrevus. We are now working at on a high concentration formulation for Ocrevus, which would significantly reduce the current injection volume of 23 millimeters so that it can be combined with an innovative on-body injector.

This on-body injector is currently tested in a type of PK/PD bridging study and should allow for home administration and potentially even for self-administration twice yearly. We expect to launch this Ocrevus high concentration formulation with the on-body injector in 2028. With that, I'm at the end of my opening remarks on our exciting emerging MS pipeline. You will hear more from Hideki and Amit. And with that, I hand over to Hideki please.

Thanks very much, Bruno. And as Bruno mentioned, I'll go over our really exciting neurology pipeline globally as a whole. And then I'll talk a little bit about fenebrutinib. If you can go to the next slide. As you can see, this is our pipeline, which is really exciting in terms of the breadth and depth of the portfolio, which expands not just MS, neuroimmunology generally, neuro muscular diseases and neurodegeneration. I'll highlight a couple of things here. First of all, as you will hear in great detail for fenebrutinib. We've achieved significance in PPMS as well as positive readout in the first RMS study. We also had a positive readout with [indiscernible], which is an anti-IL-6 receptor, monoclonal antibody and mild AD.

Furthermore, we announced Elevidys 3-year data, which is the gene therapy for Duchenne's muscular dystrophy and the sustained effect we've seen there. And then lastly, we've advanced the Phase III 2 molecules, prasinezumab for Parkinson's disease and trontinemab for Alzheimer's disease, which is our BrainShuttle tagged antibody against a beta. So if you go ahead to the next slide, please. Focusing on our MS portfolio.

We -- as you can see, there is a huge unmet medical need in multiple sclerosis. About 30% of patients remain on oral or low efficacy treatments. 30% of patients continue to progress. It really, therefore, underlines the need for new treatments, which address both the relapsing and progressive biology. And that's described on the right-hand part of the slide where there is more of a biological definition of multiple sclerosis with both relapse and biology as well as progressive biology to progression, which underlines the whole course of MS from the very beginning.

And with our leading portfolio in MS, we are well positioned to provide treatments, potential treatments for options across the spectrum of disease from relapsing to progressive biology. And you'll hear a lot more about fenebrutinib in a minute, which we think is one of the first that can address that.

If you can go ahead to the next slide, please. Underlying all this is our strategy within multiple sclerosis, which is our ambition to cure MS, with our vision to stop reverse and prevent multiple sclerosis with capabilities underlying that, with development excellence, biomarkers with our unique diet collaboration and as well as target discovery with our 2 REDs, 2 research houses, pRED and gRED that provides the pipeline to fulfill this strategy.

Turn to the next slide, please. So a little bit about fenebrutinib before I hand it over to Amit. It has potential for first-in-class and best-in-class in both RMS and PPMS. We have -- we'll go into detail in terms of the detailed data for FENtrepid, which is [indiscernible] PMS study as well as we've announced the FENhance 2, which was our first relapsing MS study to readout, which show substantial efficacy. Now BTK is a molecule BK inhibitors molecule that can address both relapsing and progressive biology, the relapse in biology driven by acute inflammation of B cells and progressive biology driven by macrophages and microglia. And the reason that's important is that this can address disability accumulation in addition to the relapsing biology. So this BTK inhibitor dual mechanism action really has potential to impact both relapsing and progressive disease.

And as mentioned, the readouts for FENtrepid, we'll go into details, FENhance 2. We had a positive readout and FENhance 1, we will have a readout in Q1 of this year. I'll go on to the next slide. Fenebrutinib was designed as a best-in-class molecule. It's the only non-covalent BTK inhibitor in MS, and it's highly selective. The selectivity is shown in the right panel. And therefore, it really has a unique opportunity for both benefit and to reduce risk in this patient population. If you go to the next slide, please. This has an optimized -- fenebrutinib has an optimized pharmacokinetic profile, as shown in this slide, compared to the other BTK inhibitors in development, and it also has a very high CNS penetration, as shown on the right slide.

And it's the only BTK inhibitor that is just near maximal inhibition of B cells and microglia in the CNS. Finally, next slide, please. I'll just remind you of our FENhance, the FENopta Phase II study that read out. And we -- because of this readout, we're very confident of the positive results for FENhance 1 for [ study ] that remains to be read out in Q1. And as you can see in the left-hand panel, we had a reduction in relapse rates to the point where there's approximately 1 relapse, I mean 17 years. And we also had a near complete suppression of [indiscernible] lesions as shown in the right-hand slide, acute inflammatory lesions. And so this is why we're very confident in the FENhance 1 readout, and the results we've seen in FENhance 2 are consistent with what we saw in the Phase II FENopta study.

So one more slide, please. So that's our fenebrutinib molecule. We also have molecules in development, which are very exciting and can add to the patient options for multiple sclerosis treatment, including our BrainShuttle CD20. On left-hand side on Phase I, we have a CAR T against CD19/CD20 also in Phase I. And then finally, on the right hand side, we have a MAGL inhibitor in Phase II in combination with Ocrevus. So we are really excited about our portfolio and options for multiple sclerosis patients.

So with that, let me hand it over to Amit.

Thank you, Hideki. So this is the same slide deck that I had the privilege to present at the ACTRIMS meeting on Saturday on behalf of all the participants, the investigators as well as our colleagues at Roche Genentech. And you can see the title slide here. So this will be the Phase III data of fenebrutinib versus ocrelizumab in PPMS. This is the FENtrepid study. And just the disclosures briefly, I won't belabor the introduction, Hideki did a terrific job. I'll just highlight perhaps the third bullet that there really is tremendous excitement in the field about the prospect that a single agent may impact both the relapsing disease biology in a meaningful way as well as the nonrelapsing progressive disease. And here, of course, we're focusing on primary progressive MS, which has been the toughest nut for us to crack in our field.

And so FENtrepid is a Phase III multicenter randomized clinical trial. You can see on the left that it recruited individuals 18 to 65, noting that 65, by the way, is higher than other trials in progressive MS, whether secondary or primary, an EDSS range of 3 to 6.5. And of course, diagnosis of primary progressive MS, Individuals were then randomized one-to-one in a double-dummy parallel group, double-blind trial, either to fenebrutinib at 200 milligrams by mouth, twice a day or the ocrelizumab standard clinical dosing followed by an open-label period. This just shows you some of the readouts over time, and it's noteworthy that while this was an event-driven study designed as such, each individual has to have at least a 120-week treatment period before the trial was reaching its conclusion. The slide here points out to the primary endpoint.

This is a composite of confirmed disability progression at 12 weeks and so, of course in our field, we want to make sure that we look at changes in neurological status to assess this ability developing over time and confirming that it is indeed likely to be related to progressive disease as opposed to relapsing disease by having a confirmation of that worsening at 12 weeks. And this is a composite, which allows people to meet their primary end point in one of three different ways, either achieving confirmed disability progression on EDSS, the standard increment in EDSS are 1 point or more.

If your baseline EDSS is less than 5.5 or 0.5 increase in the EDSS if your EDSS at baseline is greater than 5.5. But you can also reach the primary endpoint, if you have a 20% or higher increase from baseline in the time to 25-foot walk as well as the 20% or more increase from baseline in the 9-Hole Peg test. This was designed as a noninferiority study as shown on the right. And the FDA has two approved methods, the synthesis method and the fixed margin method. The synthesis method is the one that was used here, but it's noteworthy that the same outcome was achieved also with the fixed margin analysis for noninferiority. And so the baseline characteristics are shown here and essentially quite well-balanced groups.

You can see at the top, in terms of age and sex, very similar across groups. And the age indeed is just under 50. In terms of time from symptom onset and from PPMS diagnosis, just between 9 to 10 years from the symptom onset in just under 5 years from the diagnosis. So well-established primary progressive MS population, approximately 25%, 1/4 to 1/5 of individuals had some prior disease-modifying therapy use. And importantly, as shown here with respect to the [ gadolinium ] just above is that you have a very low level compared to what we've seen in other trials of 10% to 11% of presence of focal inflammatory lesions as assessed by the enhancement with gadolinium at study entry. And just to compare to the ORATORIO study of ocrelizumab versus placebo and PPMS essentially double, if not a little bit more than double, seen at the study entry.

So this really is a population of primary progressive MS patients with a well-established progressive disease biology and little actually of the focal inflammatory disease that we think of it as relapsed biology. And at the bottom now highlighted are the different components of the composite CDP, and you can see that they're very well balanced across groups, which is important in terms of subsequent analysis.

This is the primary endpoint. So fenebrutinib indeed achieved noninferiority relative to ocrelizumab using the composite confirmed disability progression. What you can see on the left-hand side is that the curves are together initially, but they separate at 24 weeks, and they continue to remain separated throughout the remainder of the study. This achieved noninferiority by the synthesis statistic as noted. And it's also noteworthy that nominally from the time of separation throughout those individuals on fenebrutinib did a little bit better than those on ocrelizumab that results in a 12% risk reduction of confirmed disability progression.

So this is an exciting observation for us given that the ocrelizumab, of course, is the only currently approved treatment for primary progressive MS. And I must say that many of us in the field felt that this was a pretty brave study to do pitting fenebrutinib as a BTK inhibitor against ocrelizumab. So a very exciting primary outcome measure. Now it's interesting to reflect on how the different components of the composite CDP primary outcome measure contributed to people reaching the primary outcome. And what you'll see on the right-hand side is that, in fact, the greatest contributor to achieving the CDP 12 was the Timed 25-Foot Walk.

The second contributor was the EDSS and the third was the 9-Hole Peg test. Yet when you look at the relative contribution of each of these to actually establishing the noninferiority between fenebrutinib and ocrelizumab, the Timed 25-Foot Walk was not particularly helpful. It was really the 9-Hole Peg Test and the CDP 12 that the non-inferiority was established by primarily.

And it's interesting to reflect that in the O'HAND study, the ORATORIO-HAND study, which was another PPMS population looking at ocrelizumab versus placebo, individuals who, in many cases, were already experiencing a lot of difficulty with their gate, perhaps with wheelchairs, but the recognition that they too may benefit from an intervention. And the decision there was to have a 2-part composite that did not include the Timed 25-foot walk, but included the CDP 12 for the EDSS and the 9-Hole Peg test.

And that study was a positive study. If you applied the same 2-part composite to the FENtrepid data, you would in fact achieve a hazard ratio of 0.78 without crossing unity, and that would have been a 22% risk reduction of fenebrutinib as a superior agent to ocrelizumab. So it's interesting to reflect on that. And I can tell you as an aside that our field is learning from every clinical trial and learning that it's very important for us to think of how to ideally adjust and match our outcome measures to the population that enters the study.

Moving on now to the question of how different subgroups behave with respect to the primary outcome. And what you can see here is that pretty much all with the exception of one point estimate or left of unity, meaning favoring fenebrutinib over ocrelizumab. And this is true whether you look at age, whether you look at baseline disability by the EDSS. And very importantly, when you consider those who did or did not have GADOLIN enhancing lesions, again, as a measure of focal inflammatory disease, it is, in fact, the case that the group that did not have a focal inflammatory disease had the best point estimate with respect to the comparison between fenebrutinib and ocrelizumab. And again, as a side note, in studies of progressive MS, where there is some presence of focal inflammatory disease that we think biologically represents relapsed biology under the surface.

When you have a benefit of your drug, it is sometimes difficult to interpret or to attribute that benefit to an impact on actual nonrelapsing progressive disease versus an impact on the relapsing disease biology that is now we now present even in people who clinically have a phenotype of primary progressive MS. So this really is a cohort with little GAD lesions to start with and an effect that is present very clearly in those who do not have the GAD lesions. And the other at the bottom is you can see baseline B cell count. People may ask, what happens if you consider this treatment in people who come off of an anti-CD20 such as Ocrevus.

And you can see here that if you consider individuals at entry with less than 10 B cells per micro liter, which is pretty low and likely reflects people who have been or had been on anti-CD20 that they too benefit from this introduction even compared to ocrelizumab. What about the safety profile? So overall at the top, you can see that there is no imbalance in terms of any adverse events, and there was no imbalance in terms of serious adverse events. There were, however, more fatal events in the fenebrutinib arm, 7 individuals compared to 1 in the ocrelizumab arm.

And as far as adverse events leading to withdrawal from treatment, approximately 14% in the fenebrutinib arm compared to 5% in the ocrelizumab arm. The great majority of those withdrawals were based on protocol mandated discontinuation of withdrawal because of liver enzyme elevations. We'll now look in more detail at the reported fatalities. And so I'm going to go through each one of these individually. I think it's important. You can see at the top, the first individual is someone who succumb sadly to COVID-19 and very unfortunately, was non-vaccinated to SARS-CoV-2. This was a young 37-year-old women with a moderate EDSS, 51 days into the study. So very soon after entering the study.

The second person, also very unfortunate insulin pump failure and diabetic ketoacidosis, which resulted and then passing away well into the study. One individual on FEN with myocardial infarction. There were no other identified risk factors in that individual or in the family history, a sudden death in an individual who had a preexisting cardiac arrhythmia. There was one completed suicide in the FEN arm. This was really a very short treatment duration, 15 days into the trial, very hard to attribute that to the drug.

There was a suspected suicide in some who did have a history, both of generalized anxiety and insomnia, one pulmonary embolism in the FEN group in an individual with known chronic cardiac failure and hypertension. And then the individual with a fatal outcome in the ocrelizumab arm or someone who died of lung cancer metastases with known background tobacco use. So there's an imbalance here for what's worth, the investigators assessed all of the FEN associated fatalities as unrelated to the study drug. So not related to fenebrutinib. One always looks for patterns. One tries to understand whether there's anything here that jumps out as a pattern. I would comment that these are all deaths that one can see, certainly in the normal, certainly in the MS population as well.

And there's no obvious pattern here, either in terms of the cause or in terms of the timing. Just a little lag here as I click on the slides. Let's talk a little bit about the liver enzymes. So you can see with fenebrutinib that there was a higher frequency of liver enzyme elevations as assessed by the fold rise above upper limit normal of the ALT.

That is included in the lower elevations although those are ones that we're not as concerned about, we're more interested in the ones that are particularly high elevations and particularly those that may need Hy's Law criteria, you can see that above 20-fold upper limit normal, 0.6% of individuals on the FEN arm experienced that compared to 0.2 in the ocrelizumab arm.

So small difference, very low percentages, but small difference. And then where you had the upper limit normal above 3 as well as bilirubin elevations, this is what meets the Hy's law at least biochemically. There were no actual confirmed Hy's Law cases, as you can see at the bottom in either arm. Biochemical Hy laws, which was seen in 2 patients on FEN and one in ocrelizumab are defined as individuals who reach these lab test elevations but actually have very plausible other reasons for developing this as was indeed the case here. And importantly, the ALT elevations attributable to fenebrutinib essentially occurred during the first 22 weeks of treatment and all of the cases resolved. This relates to BTKI-related adverse events, one is of special interest, and they were pretty comparable overall between the treatment arms.

From the standpoint of infection looked at the top, there were no differences in all infections and infestations. And when serious infections and infestations were considered, if anything, a slightly higher proportion in the ocrelizumab arm compared to the fenebrutinib arm. A known potential risk, especially with the early BTK inhibitors is bleeding. And while bruising occurred at a higher frequency in the FEN arm compared to ocrelizumab, these were essentially all mild.

There were no serious emerge imbalances between the 2 groups. And as far as neoplasms, again, no particular imbalance in terms of malignancies and then non-melanoma skin cancer is perhaps a very small signal as shown there at the bottom. And so the conclusions that we have from the FENtrepid trial is that it achieved its primary outcome of noninferiority to ocrelizumab in reducing disability progression based on the composite and in fact, had a numerically favorable outcome starting all the way through from week 24 to the end of the study, including importantly, those who did not have the majority of patients who did not have GAD positive lesions at baseline and again, of interest, the strongest treatment effect observed on the 9-Hole Peg test.

The rates of the AEs and the SAEs were comparable, including infection. There was a higher incidence of liver enzyme elevations and then an imbalance in the fatal adverse events in fenebrutinib compared to ocrelizumab. And emerging data, as you've heard, will hopefully provide us insights in terms of preventing disability accumulation by targeting both the relapsing and the progressive biologies of MS.

And I will essentially stop here, making the point, I think, and in my mind as well, that we very much hope to see fenebrutinib emerging as the first oral and second ever therapy to demonstrate efficacy as shown in FENtrepid, not just in PPMS as a CNS penetrant and highly selective BTK inhibitor, but perhaps also in the relapsing remitting spectrum as predicted by the Phase II study, the FENopta study that Hideki commented on and is emerging at least in the FENhance 2, and we're looking forward to the FENhance 1 results. So again, I'll stop here. You'll be hearing more in the future. And just to acknowledge again all of the participants, the site investigators, the members of the steering committee and the colleagues from Roche and Genentech. And I think we'll open it up to questions.

[Operator Instructions] And the first questions go to Sachin Jain from Bank of America.

A few, please. If I could just kick off with a general one for Dr. Bar-Or. Just based on this data, what percentage of PPMS patients would you use this in relative to CD20 and why and just your perspective as a clinician on liver risk seen in the study? And then two, I guess, specifically for the Roche folks. What's your visibility on liver enzyme elevation in the second RMS study FENhance 1? Just do -- you are able to comment relative to the commentary, I think, Bruno, you put this one in the FENhance 2, liver enzyme elevation was in line with a [indiscernible]. And then on the PPMS data, the presentation refers to resolution of liver enzyme elevation, but I think that's only greater than 5x. So I assume that means that liver enzyme elevation at less -- greater than 3x or 3 to 5x didn't all resolve. Do you have any color on that and why that's not a concern?

So a good number of questions, all excellent. Maybe I'll jump in to start. So I think that imagining the approval of fenebrutinib for primary progressive MS, I would frankly have no reason to think of it as a treatment that couldn't be applied to anybody with the correct diagnosis of PPMS. I think the issue of liver enzymes, which we'll talk more about, we're all trying very hard to figure out who the very few individuals are who may be at risk for serious liver outcomes. And of course, if we could identify them and exclude them and that at the time would be the own population, I would think should be excluded if we could.

But otherwise, really no particular reason, whether they do have focal inflammatory disease based on GAD lesions or new enlarging T2 lesions or not, I think that this would be an agent that would be certainly up for discussion across the broad population of PPMS. The liver enzyme abnormality of course, an imbalance between fenebrutinib and ocrelizumab and across all the fold elevations of the ALT. Again, as I alluded to, we are less concerned about the lower range. And what we've seen, in fact, if you consider the HERCULES Trial program, looking at the nonrelapsing SPMS versus the GEMINI program in relapsing MS, there also was a difference there in terms of elevations in the lower range, and it's not surprising to see elevations in the older patient population as, of course, it was true for Hercules versus the GEMINI populations.

And in this case, we're dealing with a population with the entry criteria of up to the age of 65, which is higher than the HERCULES and criteria of up to 55. So I think we're looking forward to seeing these type of subgroup analyses. We haven't seen them, I think yet from either company, but these are being worked on. And I suspect that what we're going to see is age as the likely explanation for the imbalance in the lower range, which again, we are less concerned about than the upper range.

Thanks, Amit. To jump in on the questions about the RMS studies and the liver enzyme elevations we can see in there. The comment was correct that in FENhance 2, the liver enzyme elevations are much more similar to what we've seen with Gemini and also very similar to what was seen in [indiscernible]. FENhance 1 is, of course, still blinded, so we can't comment on that. What we can say is overall in the blinded population, that there's a no reason to expect alternative data coming out of that study.

I think, Sachin, you had one more question, which was on the resolution of the elevated liver enzyme levels in the range of 3 to 5.

Yes. The slide sort of notes, denote so the resolution refers to above 5. So just obviously, drives the obvious questions.

Very happy to respond to that, that we have seen 100% resolution of liver enzyme elevations across the program, no matter the level.

The next questions go to Graham Parry from Citi.

Just a follow-up on liver enzyme elevation, actually. You commented that liver enzyme elevations similar to GEMINI and to or PEN studies on cross-trial comparison and Aubagio in FENhance 2. Could you just comment there on the high levels of elevation and how they compare when you're looking at a cross trial that they look more like GEMINI? Should we think about this as being a very different patient population to defend trepid? And then secondly, I asked on the full year '25 call how confirmed disability progression versus annualized relapse rate might be for differentiating fenebrutinib. And the question was answered saying that annualized relapse rate is a very good endpoint. So just wanted to clarify what Roche sees as the differentiating feature or unmet need that fenebrutinib is addressing? Or is it more just that it's an oral dosing option for patients?

Would you like me to quickly address the liver enzyme question, then we can hand it to Amit for the clinical one. Regarding the high elevations, those are also lower in the RMS population.

And just to clarify the question. Are you asking about annualized relapse rate in the context of FENtrepid?

So it was looking at as you down to what the differentiating feature of the product would be. Does Roche see as being the annualized relapse rate reduction in RMS? Or is it confirmed disability progression, because there's a brain penetrant molecule, then there's the potential here, and we obviously haven't seen the data that you could get a CDP benefit. So the question is, what is the -- what makes this different to Ocrevus other than just a being an oral dosing option?

Well, I would say, first of all, that what we had learned from both evobrutinib and tolebrutinib is that they are at best modestly effective against relapsing disease. And any profile that they might have, and this would be the case, not evobrutinib, which is not seen as penetrant tolebrutinib. And HERCULES and GEMINI programs showing what I think is an impact on CNS compartmentalized inflammation. But with a profile that would be viewed as rather modestly effective certainly not superior to teriflunomide, which we think of as a modestly effective agent. Fenebrutinib, based on the FENopta and the first of the 2 Phase III enhanced trials, I predict is going to be highly effective against relapsed disease, and that's great, because one of the questions, there are people over time who are not going to tolerate anti-CD20.

There are people who may not want to or may not have access to infusions are not like injections. And so I think there'll be a lot of uptake for an agent as an oral agent that could impact importantly, both the relapsing and the non-relapsing disease components. And as the community was thinking about tolebrutinib and whether it would be approved, one of the things we kind of struggle with is if you have somebody who's on a high-efficacy therapy, but progressing. So the biology of progression is there. What happens if you stuck that high efficacy therapy against relapses and start one that is best modestly effective.

And that would be a concern, and not everybody would be comfortable doing that, especially for patients who had evidence of a lot of focal inflammatory or relapsed disease activity prior to going on high-efficacy therapy. So to me fenebrutinib is an agent that I would be delighted to think of as having a similar profile, if you like, on both relapsing and progressive disease, but it's a great -- an important alternative or as a first in PPMS.

And if I can just add, fenebrutinib is really the first and only oral high efficacy treatment, potential treatments for both PPMS and RMS. And FEN, along with Ocrevus ID, Ocrevus subcu, it has potential to offer a leading portfolio for patients and choice for patients, high efficacy treatments.

Next question will go to Luisa Hector from Berenberg

So for Dr. Bar-Or, I wanted to check did Ocrevus perform as you would have expected. And what proportion of your primary progressive patients are being treated with Ocrevus today? And if not, why not? And maybe just understand some of the wording. So you have the primary endpoint, the 3-part composite. I know in the press release, you talked about a reduction in disability progression. If can that be claimed? And is there anything you can claim on the two parts, the composite endpoint where you showed us versus [indiscernible], the outcome there? Like can you sort of drill down and make any claims where the data looks more favorable? .

Maybe I'll start at the end. So I don't think one can make formal claims about that, given the redefined statistical plan for noninferiority. That said, the data is the data. And we always want to see the data and you see the point estimates how the curves separate. If the curves were separate at the beginning, which sometimes happens, then maybe an imbalance in patients coming into the study, and it's very difficult to interpret.

But they separated at week 24 not before, and then they remain separated with Fed consistently having a favorable impact relative to the ocrelizumab. And to me, that is something that we -- is noteworthy. We are interested to see what will happen, of course, in the FENhance trials as well. With respect to the question of -- did I see what is expected? I wasn't sure initially whether you're referring to within the clinical trial or sort of in practice, and I'll start with practice first. Of course, in the clinical trial, the very fair assumption is that it's behaving in PPMS as it was shown to do so in the ORATORIO study.

So modest but important impact on limiting progression or disability in individuals with primary progressive MS. Our clinic, which follows approximately 6,000 individuals with MS. So that's a fairly large clinic, one of the larger ones in the U.S. We have over 3,000 people on anti-CD20, the majority of whom at the moment are on ocrelizumab. I'm always very reluctant to say in clinical practice, I see an effect. People do get worse. But I think that the population that we followed on Ocrevus with PPMS has generally been recognizes that they are on the best available, the only available treatment and that, too, is reassuring being proactive about their care.

Again, very difficult to gauge when people are getting worse, whether they would have gotten worse even more so without the treatment. But of course, that's what the clinical trials are for. So did that answer the questions? Are we also asking in the context of the trial?

Can you still hear me?

Yes.

Yes. Yes, and also within the trial, so the performance of Ocrevus within the trial.

I mean I am suppose -- nominally not as good as fenebrutinib, right? But one can't comment about how much it did other than relative to fenebrutinib.

The next questions go to Michael Leuchten from Jefferies.

A few questions for Dr. Bar-Or, please. One, just going back to the shape of the curves. If this is a composite of the disability measurements, it's good to see the curve separating early, but why wouldn't it separate further given it's a composite not a single estimate? Just your views on that, please? And then a question on how do we get so many ALT above 3x, if there was monitoring place. Just trying to understand how those patients sort of weren't detected earlier? Does it just happen too quickly to catch them? Just wondering about the timing around those? And then lastly, I think this is for the Roche team. Can you explain to us who determines where the Hy's Law cases or Hy's Law case? Is that the guys, the way you look at the data? Is it a regulator? Like who makes a call whether the comorbidities, I can call it that make it a non-Hy's Law case? .

So maybe I'll start with respect to the composite. So if you hit one of the three, then you hit your endpoint, and you're not contributing to further. So if you hit by one of them the fact that you hit subsequent. And that's an interesting analysis that one would want to do. When you look at these reverse Kaplan-Meier curves, you always have to be cautious about what you see towards the end, just because if you look across the bottom, there are fewer or fewer and eventually really, really few people contributing to the comparison. So if there were an opportunity to help people avoid the progression of visibility on multiple counts over time, that's an interesting question. To me, what we've learned as a community from the trials, and we had a very detailed discussion about this on Saturday at ACTRIMS, the question again on how to best choose the outcome measures with respect to these individual measures or different combinations of composite to match the population. So it kind of makes sense in the O'Hand that if people can't walk, there's no point in using a walking measure.

Here, apparently, people also have difficulty with that test discriminating the two. And there are potentially good reasons, a 25x Foot Walk, for instance, can change quite a bit. If you are not as bad as you were, you may -- or if you get worse, for instance, you may then need an ankle foot orthosis that you didn't need before. And now with the ankle foot orthosis you might actually work faster than you did before. So that could contribute to noise in the assessments. And of course, when you have a more noisy measure, it becomes much more difficult statistically to see it separate even if there's a potential separation or comment on noninferiority there.

I'm happy to take the part of the question regarding liver. Amit, if that works out for you.

Sure. Fine.

Absolutely. So there was a question about the percentage of greater than 3x the upper limit of normal and how are we catching those. And it's important to recognize that the every 2-week testing was actually introduced quite late into the program. So for the liver enzyme elevation data that we're looking at approximately I would say the vast majority of that was captured with monthly testing with which we started the program. After we introduce every 2-week monitoring, we prevented any further Hy's Law cases as well as any high elevations.

The goal of liver enzyme testing is to prevent severe liver enzyme elevations, not to prevent the lower ones because, of course, around 3, 3 to 5x the upper limit of normal can occur due to many things in life, infections, something you eat, how you're working out. So we are following the liver enzyme elevations in order to prevent the high ones. And we saw that with the introduction of every week testing, we have brought down the elevation seen so that you are making sure that you're keeping patients safe while they're on therapy.

And then the second part of that question was who is determining, what is considered a Hy's Law event. And thank you for that question. We do have an independent adjudication committee for a Hy's Law events.

Next one in the row is Colin White from UBS.

It's Colin White from UBS here. Two questions from me, please. The first is on the positioning of fenebrutinib in the treatment of PPMS and RMS. Given unexpected REMS similar to what we've seen in the clinical studies and the requirements for liver monitoring, and partly in PPMS, do you -- does this position at in patients that are maybe not doing well on an anti-CD20? Or does it possession in a broader population?

Similarly, in RMS, with those lever monitoring requirements, does it position it ahead of anti-CD -- and as a similar first treatment option with anti-CD20s or perhaps afterwards. The second question I have is looking ahead to the RMS data that will be presented and thinking about the benchmarks for relapse rate and lesion reductions versus Aubagio. Is something like an anti-CD20, like [indiscernible] Ultimate trial, where we saw a 50% reduction in relapse rate versus Aubagio and greater than 90% reduction in lesions. Is this something -- is that -- are these appropriate benchmarks? Or should we be thinking about it differently than that as we head towards the data?

Maybe I'll start, and Alexandra, you can perhaps follow. So at the moment, I'm optimistic that fenebrutinib will have an indication in primary progressive MS and relapsing MS. Of course, we need to confirm that with the FENhance 1. I think that an important point raised by Hideki in his introduction is that we really think differently now about the biology that we're trying to target. .

And while the regulatory agencies are still going to require to call it something that is sort of in the clinical spectrum. And as an aside, the same international committee that is charged with the new diagnostic criteria, that have come out, as you probably know recently, is now charged with revisiting the clinical course descriptors. And we will probably one day get rid of SPMS as an entity because of this notion that people start with relapsing biology. And then when magical day transition into secondary progressive MS is now understood to be simply incorrect biologically, that both relapsing biology and nonrelapsing progressive biology occur in an individual patient at the same time across probably decades of their experience. And we now have direct measures that capture at least some of the presence of progressive disease biology as early as you diagnose patients and even in ones who are referred to as radiologically isolated syndrome, they had an incidental finding of lesions that turn out to be consistent with MS and further testing that confirms a MS without having had a symptom.

So that, of course, is not necessarily going to be relevant for the here and now, but not that distant future. And the point is that as a academic clinician, there is presence of progressive disease biology, even in people who clinically you'd call relapsing remitting or relapsing MS. And if I have an agent that I believe works on both of these biologies, I'd be excited to start it. And of course, if it's considered to be well tolerated, which this agent really is, the safety monitoring, I think, if we can mitigate I would be very comfortable offering this treatment for anybody in the MS spectrum in whom I think there is both some element of relapsing disease and progressive disease biology. So that's really just to kind of share with you some of the evolution and sort of the conceptual framework of MS by the community, and we need to help the regulatories sort of catch up with that.

But once these are approved, I think that there will be a pretty substantial uptick as people recognize the potential benefit on these 2 biologies across a very broad spectrum of patients.

And the follow-up question was on expected RMS data?

Yes. It was on benchmarks, appropriate benchmarks to think about in terms of relapse rate and in terms of lesion reductions, percent reductions versus Aubagio, like what should we expect?

As Hideki started off with and then Amit commented on, of course, we can't share the totality of the FENhance 2 data yet, but we were very happy to see that it's very consistent with the FENopta data. So if you play that forward, then we could anticipate a really profound reduction in both ARR and GAD lesions.

Next question is go over to James Quigley from Goldman Sachs.

I have 2 for Dr. Bar-Or, please. So firstly, from a monitoring perspective, with some of the current oral therapies, including Aubagio, they already have a liver monitoring requirement. So typically, how long would you monitor with other therapies? And how burdensome would you characterize every 2 week monitoring for 20 weeks or so? And is there anything in the data to suggest that patients could stop monitoring after a certain period of time.

So if there's nothing very limited increases after 10 weeks, is there a chance you could stop those early? And then second question, and again, maybe asking you to somewhat speculate here, but we've seen 2 covalent BTK inhibitors not show any benefit in an annualized relapse rate in RMS noncovalent BTK show a benefit in both PPMS and one RMS trial. And again, we need to see the rest of the data for FENhance 1. Hideki showed some data on plasma availability and greater inhibition in the CNS. So given what we've seen so far in the clinic, how convincing do you think it is that noncovalent BTK inhibitors could have greater efficacy than the covalent binders, or is it more down to the dosing and therapeutic window of each asset.

Sure. So with respect to the liver enzymes, depending on the totality of the data with fenebrutinib once the FENhance 1 and 2 come out, if it's basically consistent with what we've seen up to now, I would be comfortable with the requirement of biweekly testing for the first period and then going on to the same frequency that we do with other agents.

Keep in mind that neurologists are actually pretty familiar with monitoring for liver enzymes. We deal with medications for epilepsy, for headache over the years, et cetera. So this is not new to us, including in the MS field, of course, as you pointed out, with Aubagio. This would be an increased burden initially compared to what we do, no question. But I think that the prospects is borne out of what this treatment has to offer for an upfront increased intensity of monitoring would be embraced by most patients and clinicians.

I do think that we will look to Roche Genentech to sort of help us with that, right? I think that's going to be important. I suspect the regulatory agencies will want to see something in place. They will feel confident we'll keep people safe as well. The very interesting question about whether it is indeed the reversibility, the noncovalent reversibility of fenebrutinib that contributes to the differences. I think first, that the data to date already suggests to me quite strongly that fenebrutinib is different from tolebrutinib, evobrutinib in the context of the impact on relapsing disease biology.

The extensive work that I understand went into the choice of dose regimen for fenebrutinib included a very comprehensive assessment of how a noncovalent reversible binder may differ from the covalent irreversible. And one of the important points is that you're less concerned about perhaps the Cmax of your therapeutic and more about the area under the curve.

And it's very interesting to speculate that an area under the curve that is more comprehensive may, in fact, result in different biological responses that could have relevance to disease biology. An important insight, at least already from FENopta is -- looks as though there is some degree of B-cell depletion with fenebrutinib, which is not seen or at least has not been reported in the same way with tolebrutinib and evobrutinib. And that may be important because it may again underscore biological differences between the agents that would confer to fenebrutinib a much higher efficacy on relapsing disease activity than the others out there that we've heard about to date.

Yes. And just on top of that, fenebrutinib is a unique BTK inhibitor, not only is it cov -- noncovalent irreversible, this very high point seeing very high selectivity. And what's important to keep in mind is the the optimal PK profile that we shown on that slide that I showed earlier. So the optimal PK could contribute to the FCC as well. .

That's what I was referring to is the area under the curve in terms of exposure, for sure.

I was going to comment on the -- what is Roche doing to contribute to identifying patients at risk. And also, how are we thinking about reducing burden for patients and physicians. And I think in this space, we're very lucky that MS and neuroimmunology is considered in end-to-end disease area in Roche. So we really can draw from the support of the entire organization. And so as we're getting this data out of the Phase III, so we are able to loop that back with our translational medicine colleagues, with our early research colleagues. And of course, now with our growing AI and machine learning capabilities. So that if anyone was going to be able to identify a predictive measure to predict patients that are at risk, we have a good shot at being able to do that.

And then on the other side of the commercialization engine, we are really focused on what can we do to reduce burden for both patients and physicians. And we are actively working through that. And of course, during -- throughout the filing process and as we have dialogue with the FDA, we will determine what is necessary -- and we are very committed to bringing fenebrutinib to patients in a way that is easy and on burdensome to them.

Next one in the queue is James Gordon from Barclays.

James Gordon for Barclays. A couple of questions, please. First one was on PPMS. So have you had a chance to discuss this data with FDA, particularly with relation to the agent's liver profile. And in terms of how different this is seen as being versus tolebrutinib, which obviously CRO due to the liver profile, because the key investor question I've had today is how confident are you that this isn't going to be seen as the same given the response on the liver profiles?

So if you could talk about that lease? And the second question was just on timing. So I think that the second RMS trial, we're going to get to a mid-H1. So could it theoretic would be possible that we would get data from the 2 RMS trials, both AAN in late April or 18 to 22nd of April? Or is that a bit ambitious? We might have to wait until later this year to understand what this really looks like in RMS.

I'm happy to address both of those, if that's okay. Perfect. Regarding the path to approval and especially how is Roche thinking about that in light of the tolebrutinib CRL. We think that the CRL was written very clearly and transparently and really elaborated on the FDA's view on benefit risk and what's necessary to achieve. So -- and I believe that's very much in line with Roche thinking as well. So it has to do with benefit, then risk and then, of course, risk mitigation measures. .

So with that, we have shared all of our data with the FDA. We are in dialogue with them. And based on what both benefit risk and what we think we can do for risk mitigation, we do think we have a path to approval. And specifically in terms of mitigation measures, the most important being what we talked about before, that going from monthly monitoring to every 2-week monitoring, we've been able to prevent high elevations and any further Hy's Laws. So as -- although in the CRL, they discussed the weekly monitoring for tolebrutinib being not adequate to capture high elevations. For fenebrutinib, we have not seen that. So we think we have a different molecule. We have -- and the FDA thus far seems to be appreciating that as such.

And regarding timing, you are -- you are correct. We are looking to see data kind of mid-H1, end of Q1. And so if everything goes well, we would be very excited to see that data at AAN.

And we move on and the next one is Steve Scala from Cowen.

A few questions. First for Dr. Bar-Or, will you switch any CD20 patients to fenebrutinib in RMS based on the data to date? Secondly, for the Roche folks when modeling fenebrutinib in the commercial market, do you assume there will be other BTK competitors on the market? Or will fenebrutinib be alone. And curious where are the majority of the patients or opportunity, I should say? Is it in the U.S. or OUS? And then the last question, you mentioned the numerical trend favoring fenebrutinib over Ocrevus. But did Roche actually test for superiority to Ocrevus after testing for noninferiority and did it indeed fail superiority?

May be I'll start. So I think the first question was, would I switch people and if so, who, who are on Ocrevus for relapsing MS to fenebrutinib. So first, for individuals who are having issues with anti-CD20, which again is overall very well tolerated and people do very well from that perspective on Ocrevus. But nonetheless, there are some who will have difficulty with it. There are some who will develop some infections, more frequent infections, run of the middle ones, but more than they were used to. And then there will be some who will have their IgG levels go down to levels that are quite reduced and even below lower limit normal, and we know that there's a small increased risk of more serious infections.

So anybody who needs to come off for a consideration of coming of Ocrevus, I would be comfortable switching to fenebrutinib, again, assuming that it will emerge as having the high efficacy on relapsing disease that we're hoping it will. The question of whether I would switch someone who is doing well on Ocrevus. And here, I think, again, it depends on how much one leans on the FENopta data suggesting that fenebrutinib may actually be a little bit better than Ocrevus. And I think while, again, statistically one cannot say that formally, recognizing that the Timed 25-Foot Walk really did not help here, it kind of hurt in a sense, that those who did have an opportunity to contribute did so with the 9-Hole Peg test and EDSS.

And of course, this would be viewed as post hoc and sort of cherry picking, and we want to avoid that. But if faced with a situation in the clinic where somebody is concerned about progressing in the face of Ocrevus, and we do have more and more people with relapsing MS, who run Ocrevus for relapsing MS, but nonetheless have emerging progression. That would be a reasonable context in my mind to discuss fenebrutinib, which would not be the case with the other profiles of BTK inhibitors we've seen to date.

Thanks, Amit. Regarding the question about our commercial modeling, of course, this is a very dynamic space. So our models have adjusted over time as we get data from the competitors. And at the moment, as you all know, there are many BTK inhibitors in development for multiple sclerosis. So it perhaps it would be premature to presume that we would -- that fenebrutinib would be the only one there. So we're considering multiple different models as we go forward. .

For the -- where do we think we will get the patients from? Of course, the U.S. is an important part of the MS market, but we think there are patients that would benefit from fenebrutinib globally, of course. And what type of patient segments, as Dr. Bar-Or has talked about, there's many different types of patients that could benefit from a drug -- an oral drug like fenebrutinib. We think that there are perhaps, especially in Europe, there are a significant segment of the population is still on low-efficacy orals.

And we as a company and I think in partnership with our steering committee are really believers that all patients have they deserve to have high-efficacy therapy. So there is that segment of the market. There is also the part of the market as Amit discussed, in which patients have been on anti-CD20s for some time now and it might need to switch to something else. And so fenebrutinib could offer an opportunity there. And then we also do not forget that even today in 2026, about 40% of MS patients are on no therapy at all. And this is really shocking to find out, but we're really hoping with having an oral efficacy agent that could treat both PPMS and RMS across the spectrum. We could perhaps continue to grow the market and extend high efficacy care to those patients that have not received it at all yet.

I just want to jump in because I'm the only one of the 3 here that have to jump off, so apologies for that. But I appreciate the questions around how we discussed with FDA because I just want to state there. Clearly, that patient safety is the #1 priority for Roche, And we're going to work closely with the FDA to get this to patients as quickly as possible and protect patients and keep them safe as well. So I appreciate those questions. But I do have to jump off apologies for that.

I think Steve, did we answer all your questions?

It was just the one about superiority, did you test for superiority?

Did we test the superiority?

I can hand this to Amit, but also very happy to take that as well. We did also test for superiority. Fenebrutinib did not meet statistical significance on superiority.

Next one in the queue is then Richard Vosser from JPMorgan.

A couple of questions, please. Dr. Bar-Or, you mentioned at ACTRIMS about the signal to noise ratio about the composite CDP end point. And I just wondered whether EDSS aligned would have -- I mean, clearly, it would have changed the result here. So the question was what composite CDP endpoint was actually used in the FENhance trial. Is it the same? Or is it a different CDP? And have you had the chance to use EDSS there? Because that seems like a more less signal-to-noise ratio on the endpoint. So that's the first question. Second question was just on the proportion of patients that were actually subject to the biweekly testing during the ramp-up phase or the initial initiation of treatment during those first 20 weeks across FENhance and the Fentrepid trials, just so you can show the success of testing. It would be nice to understand the proportion there.

Go ahead, Alexandra, let you take [indiscernible]

Regarding the proportion at this time when we haven't yet to read out all of the studies, I think giving precision there is going to be really challenging. So at the moment, we can just say that the vast majority of patients had their -- their first 20 weeks of therapy with monthly testing. And we'll be digging into that, but we were able to see with following the blinded data across all 3 studies that once the every 2-week monitoring was in place, we were able to prevent higher-level elevations. But getting more precise for that, we're going to have to wait until we have the readouts of all 3 studies, apologies.

And then the disability outcome, Alexandra and FENhance.

Yes. Apologies. The disability outcome measure used in the FENhance studies is also the composite CCDP. Of course, we are capturing EDSS. As you likely well know, for EDSS alone, it's very tough to power Phase III studies individually in RMS to be able to detected treatment difference for CDP, which is why most MS studies, if not all, have identical or sister studies in which EDSS is pooled across both of them. We saw that going back to the Ocrevus trials, same thing the other anti-CD20s that have been used. So we will be looking at both the composite in the individual RMS studies as well as the pooled EDSS across both RMS studies.

And I might add that the discussion in the community is ongoing as to the utility of composites versus not in. Again, I mentioned before that we keep learning. Maria Pia Sormani was one of the main statisticians in the MS field. Although not published, she's presented this data a couple of times and it's okay to share.

She did a meta-analysis comparing outcome measures of facility that use composites and use EDSS alone. The conclusion from all of that is that EDSS is no worse than the composites. But she acknowledges that as a meta-analysis, what we lose there are the contextual issues, which relate to different populations that depending on their status and they are most likely progress with clinically there may actually be an advantage of thinking carefully about the value of composites, which allow more people to hit an endpoint than if you just limited to one measure.

EDSS, of course, is not sensitive to a lot of progression of disability from a patient's perspective. And that gives you more events and more events can be helpful in terms of your comparisons. But of course, if they introduce more noise, they're not that helpful.

Next is Simon Baker from Redburn.

A couple of questions for me. Firstly, Dr. Bar-Or, really going back to a things that's been talked about repeatedly. But I'm going to ask it a slightly different way. With fenebrutinib ultimately approved, let's assume -- what are the choices you would make in deciding whether a patient received Ocrevus or fenebrutinib, just giving us an idea of the process that one would go through there.

And then secondly, looking at baseline GAD-enhancing lesions, is the ratio of present to absent in the study typical of your broader clinical practice? And related to that, and I appreciate this has got no regulatory legitimacy. If you reanalyze the data excluding those baseline lesions present. What does that do to the noninferiority and superiority of the study?

I wonder given you're so close on the upper confidence interval to one, would the removal of that low responding group actually tip it, so it was a superior result. I know it's purely nominal, but it would be interesting to know if that analysis has been done.

Right. So while the point estimate of the hazard ratio for the small population -- for the large population that did not have GAD lesions was favorable for fenebrutinib versus ocrelizumab, the confidence interval just marginally kind of clipped the edge of unity. So you might say that as a single subpopulation analysis that didn't meet the mark.

But I think that the relative contribution is important to note, and it really reinforces the notion that this drug is acting on the biology of interest that we're trying to target, which is the nonfocal inflammatory vision. Of course, when you see GAD lesions there, you can say that they're there. When you don't see them, you can't exclude the possibility that a patient still harbors that biology and may manifest with it during the trial, and we'll be interested in seeing additional subgroup analyses, including people on trial, we do or don't develop new lesions.

With respect to the question, the first one, so I think that the -- if indeed, we will see that fenebrutinib has an important impact on relapsing disease activity relative to teriflunomide. I would be very comfortable positioning this as a first-line therapy.

We now in a program are very much about high efficacy upfront. But again, when you combine the understanding, there's a biology there that is not just the relapsing biology, an agent that has a profile of that as well. Would be viewed as favorable. And one of the things that we now can do, not everywhere, but more and more places are doing this, is you can actually ascertain whether someone does or does not have chronic active lesions based on the presence of these paramagnetic rim lesions to indicate that the biology is there.

That's only one aspect of progressive disease, and we hope to have better additional biomarkers. My sense at the moment is that with a high efficacy profile in relapsing disease and what we've learned from FENtrepid in terms of non-relapsing progressive biology that this would be absolutely in the discussion in terms of front line therapy. And then we monitor people as they go along. And if there's any reason to think that this is not the right medicine for them, we can adjust, but it would be a great medicine for people to be able to start with as well.

We have one final question. Luisa Hector from Berenberg.

Was the tri-composite endpoint mandated by the FDA or could you choose? And then I just wanted to check on the imbalance of debt. So once that was observed within FENtrepid, did that trigger a look across the whole program, i.e., can we be confident that there is no imbalance in the FENhance trial, the FENhance 1.

I'm happy to start on the CCDP question. As we were thinking about putting together a PPMS study. Of course, it's -- one of the challenges is that disability events accrue very slowly. And we, as an organization, of course, want to develop drugs as quickly as possible and get them to patients who need them as quickly as possible. So EDSS has been very challenging in that in that it is not that sensitive to change.

So that is what has prompted the MS community to be looking for more sensitive markers for disability progression. Therefore, we put together the EDSS plus 9-Hole Peg test and Timed 25-Foot Walk.

Hoping the Timed 25-Foot Walk would be a much more sensitive measure to pick up disability progression sooner. So that was not mandated. It was us really going out and trying to develop drugs for progressive MS quickly. Unfortunately, as Dr. Bar-Or explained in some detail, we certainly captured more events with time 25-Foot Walk. But unclear if those were meaningful or if that was just capturing a lot of noise.

And I might respond to the other question. Obviously, we're going to wait to see what happens with the FENhance program in terms of fatalities. But it's worth noting that there have been in the order of 1,500 individuals who've been exposed to fenebrutinib in other contexts, and a good 1,200 or so in nonmalignant context.

So conditions that include rheumatoid arthritis, systemic lupus and chronic spontaneous urticaria. And there were no imbalance in AEs with fatal outcome for fenebrutinib in all of those contexts and in people who you might imagine are perhaps additional therapies or sicker than average MS patients. So that I think noteworthy in the context of fatalities. I don't know if you have anything else to add to that, Alexandra.

Just that we, of course, are working very closely with our IDMC and sharing all data, they, of course, have access to the unblinded data from across the Phase III program throughout. So we've been in close contact with them and, of course, with the FDA as well.

Very good. So I think with that, we can close the Q&A session. Let me quickly thank again, our speakers for their time and their efforts exploring new treatments here, treatment options for MS patients. Let me also thank the IR team members who worked on the slides and prepared this event. So I have to call out here [indiscernible] and also [indiscernible] for the event organization. I hope the event was helpful in providing here a timely update on the MS franchise. If there's any remaining questions, please reach out to the Roche IR team. We're happy to follow up. And I think with that, we close the call. Bye.

Thanks, everyone. Bye-bye.

Ladies and gentlemen, welcome to Roche's Full Year Results Webinar 2025. My name is [ Henrik, ] and I'm the technical operator for today's call. Kindly note that the webinar is being recorded. [Operator Instructions] One last remark. If you would like to follow the presented slides on your end as well, please feel free to go to roche.com/investors to download the presentation. At this time, it's my pleasure to introduce you to Thomas Schinecker, CEO of Roche Group. Mr. Schinecker, the stage is yours.

Thank you very much, and good morning, good afternoon and good evening. I'm really, really excited to share with you the update for the full year because we had an amazing fourth quarter, not only in terms of financial results, but also in terms of pipeline news. So let me get started on the normal overview slide. So group sales in 2025 grew with 7%, Pharma at 9%, Diagnostics at 2%.

Again, this was due to the China healthcare pricing reforms. Without that, Diagnostics actually grew with 7% last year with a very strong operating performance with a core operating profit of plus 13% and a core operating margin plus 1.9 percentage points and core EPS plus 11%. So you may ask what's the difference between EPS and OP? Why is there a deceleration there? This is basically mostly driven through higher taxes.

On the full year LOE impacts, we had impact of about CHF 700 million. Now I come really to the exciting part. We had truly an outstanding Q4 when it comes to pipeline news. From a pharma regulatory perspective, the EU approval for Gazyva in lupus nephritis, U.S. and EU approval for Lunsumio as subcut solution in third-line plus follicular lymphoma. And U.S. filing for giredestrant in post-CDKi ER-positive HER2-negative metastatic breast cancer.

Now come the many positive readouts that we had. Phase III, FENtrepid and FENhance, so 2 positive studies in fenebrutinib, 1 in PPMS, the other one in RMS, a positive Phase III study just already this year in Enspryng in MOG-AD, positive Phase III lidERA giredestrant study in adjuvant ER-positive HER2-negative breast cancer, a positive Phase III PiaSky in aHUS and a positive Phase III in Gazyva in INS and another positive Phase III Gazyva in SLE and positive Phase II in CT-388 in obesity.

And I know Teresa will go through a lot of these details with you, but you can see with very, very busy newsflow, and I think there are more exciting things to come also this year. On the diagnostic side, regulatory approval of the Elecsys dengue test, Matt will talk about that, the cobas BV/CV test and also the mass spectrometry extension of our menu. So exciting launches there as well. There is significant newsflow ahead in 2026. We are awaiting the second fenebrutinib study in RMS. We are awaiting persevERA, so the giredestrant in first-line ER-positive, HER2-negative metastatic breast cancer.

We have other studies reading out in -- for Itovebi, but also for our divarasib KRAS medicine and further in Lunsumio and Gazyva. So again, I think very exciting. And we have a number of Phase II readouts coming. So it could again be a very busy year in terms of transition from Phase II into Phase III. And of course, everyone is talking about it, our next-generation sequencing solution. And we promised it for many years. Now here it is.

So I'm super excited also personally that we are now coming with this very exciting solution to the market. And yes, I think it will cause a couple of [ waves ] in the market. Now let me go through the growth rates. And I don't think I have to cover too much on the left-hand side. But what you can see on the slide is that we have consistently strong growth in the last 2 years. And even before that, when we had the washout of COVID-19 tests and the medicine, we had a good underlying growth.

So we always said we will deliver and we delivered. Now on the 2025, Pharma kept growing at 9%. Diagnostics, we did have the healthcare pricing reforms impact in China. Without that, Diagnostics would also have been growing consistently at 7% over this time period. Again, here, we just look at the full year, 9% growth, again by Pharmaceuticals division, 2% by Diagnostics. Again, without the China effect, it's 7% and the Roche Group has 7% growth. And this is really driven across our entire portfolio.

I think Diagnostics, I already explained, and I know Matt will go into that further. Vabysmo, we have continued strong global growth. We do expect that we will see even more uptake in the next year when it comes to the U.S. market now that we are also supporting more on the co-pay assistance foundations, and Teresa will go into that. Also Xolair keeps growing significantly. Gazyva, we have now launching in lupus nephritis, but you will see also the other indications really contributing to the growth in this business.

Oncology growing well at 6%. For Phesgo, we are now at the global conversion rates above 50%. Tecentriq, returning to low single-digit growth, Alecensa growth driven by U.S. and Japan. On the Hematology side, Polivy strong now. We are reaching a U.S. patient share of 36%. Columvi/Lunsumio, growth driven by second-line plus launch and third line plus DLBCL, strong third-line growth in follicular lymphoma. So you can see also in the Ocrevus franchise, we see now a strong uptake of the subcut solution, as we also discussed, and Evrysdi is the leading SMA solution and medicine now with more than 21,000 patients on treatment.

So overall, we've achieved the upgraded guidance. On mid-single-digit sales growth, we had 7%. And you may remember in the QR -- in the IR call in Q3, you asked why are we not upgrading the guidance on sales? And my answer was because 7% is still mid-single digit, and we delivered 7%. So we delivered on what we also communicated. On core EPS, we upgraded the guidance from high single digit to high single digit to low double digit. And why did we do that? Because we already knew that we would land in double-digit range.

So we wanted to include it in that range, and that's why we upgraded it at the time. And we further increased the dividends in Swiss franc. So I can say we, for the second year, upgraded the guidance during the year, and we ended up on the upper end of these 2 guidances every time in '24 and in '25. Now what's super exciting is the growth outlook. And the growth outlook has fundamentally changed based on some of the Phase III readouts that we have seen in Q4.

Giredestrant, our SERD has had 2 positive readouts, one with evERA and the second one with lidERA, lidERA reading out early. We expected the readout only next year. But based on the very significant results, it read out early already this year. We're now waiting for the persevERA data, but we clearly see that we have a very active and very good molecule in hand. Not only is it going to replace the standard of care and it's going to become the new backbone in this field but it's so safe and tolerable that we do believe that this will become the standard of care.

Fenebrutinib in MS, we had 2 positive studies here in RMS and in PPMS, and now Teresa will go into that. We're still waiting for the second, the FENhance 1 study in RMS in order to be able to file this. You will see the data in not-too-distant future. But again, we are very excited with the data that we have seen. Gazyva, the same. We've had already positive data in lupus nephritis. Now we have 2 more additional Phase III data, again, something that will give us momentum in the midterm.

On vamikibart and in Enspryng, we had 2 trials each. And each of the 2 trials, we always had 2 arms and 2 different doses. In one trial, in both of those, it was fully positive trial. In the second trial, one dose was positive, the other one was negative. Now this was always a very close call. And based on the conversations we had with the FDA, we do believe that the FDA will support filing here. So this is the midterm. And in the long term, what's also exciting is that we had 10 NMEs moving into Phase III. That's a record for us.

We've never had 10 NMEs moving into Phase III. And these are all NMEs with substantial revenue and patient impact attached to them. Now we know the other part or a topic that's on your mind is our agreement with the U.S. government. Now the agreement we reached with the U.S. government is not an LOI, it's a contract with the U.S. government. And based on this contract with the U.S. government, which is terminated for the next 3 years, we get an exemption from tariffs, and we get an exemption of the demo projects.

In return, we agreed to the Medicaid rebates in some of our portfolio. We agreed to the encouragement of other wealthy nations to reward biopharmaceutical innovation and to pay their share to contribute to innovation. And we also support the direct-to-patient medication access with our influenza portfolio, Xofluza and Tamiflu and also future medicines with which we can go direct to patients. Also, we agreed to invest in the United States $50 billion over the next 5 years. This includes R&D and PP&E investments.

And for example, on the red dots on the right-hand side, you can see where these investments are happening. In North Carolina, Holly Springs, we are investing $2 billion in manufacturing for our CVRM portfolio. In Indianapolis, we're investing in CGM manufacturing. And in Boston, we're investing in our research hub as well as we're going to invest more in Genentech in San Francisco. But we already have a very strong manufacturing and R&D network present in the United States. Now it's all about delivering on the next innovation cycle.

So I think we have good growth momentum. We do believe we can sustain the good growth momentum and how do we keep that beyond 2030. Well, one is we've made substantial progress on our initiatives in order to prepare the company for future success. For the last 3 years, we've been talking about high-performing organization, about delivery and about execution. And I do believe we've delivered on that. We've introduced the bar so that we focus on those medicines with the highest impact.

We have an intentional focus in terms of TAs and also -- and disease areas. And we look at the portfolio as an overall investment portfolio with certain risk and reward profiles that we want to balance to have the right portfolio. We're expanding into new technologies, not only in Dia, but also in Pharma, and we're implementing AI across the value chain. And we've made good progress in R&D. Now more than 60% of the NMEs are post bar, 66% of the late-stage projects have best-in-disease potential.

We want to get to 80% and 60% more average peak sales per pipeline project. If you actually take not end of '23, if you take end of '22, it's even higher. And the same for the total portfolio value. Since end of '23, it's about 45%. If you look at end of '22 as a comparator, we are more than 60% higher, and this is a risk-adjusted value. So we do believe we've made a significant move when it comes to our pipeline in terms of higher rewards and also manageable risk.

We have a strong on-market portfolio. In the midterm, we have great readouts that are going to help us sustain our growth, but we have many, many NMEs that will read out before the end of this decade. And many of them can be significant contributors to the sales of our company. And here, you can see the prioritization that we have done over the years in our portfolio, really taking out high-risk, low-value projects and adding higher-value projects with very strong data as a foundation that gives us more confidence into Phase III.

And this has resulted in a significant shift in our portfolio value. As mentioned, this is year-end '23 as a baseline. If you take year-end '22, it's even significantly higher. And on the Diagnostics side, as I mentioned, we have been experiencing the headwinds in China from the healthcare pricing reforms. These headwinds will become a lot less in 2026 and will be gone in 2027. We will continue to grow significantly. This year, we expect mid-single-digit growth, but then back to mid- to high single-digit growth.

And these products that you see here can each contribute additional CHF 1 billion when it comes to sales per year. And again, very excited about the sequence of that's coming. Now let me go through the outlook. For 2026, we again have an exciting year when it comes in terms of pipeline readouts. But then after that '27, '28, '29, we will have many different NMEs that will have Phase III readouts. In '26, I just want to highlight a couple, persevERA for giredestrant and fenebrutinib, Itovebi, divarasib, a number that can also continue to drive our growth in the next years.

And of course, we have a number of Phase II readouts in obesity, 2 of them already have been positive this year. So we had a very good start. So the year ended well in terms of readouts and the year started well in terms of readouts. So we are very positive in terms of the momentum that we have and that we can continue this momentum. And these are the 19 medicines that we can launch by the end of the decade.

Now it's clear that not all 19 ultimately will make it. But these are really substantial contributions that they can deliver to the future business of our company and to patients out there. So I'm very excited about what's ahead for us. Now let me talk about the 2026 guidance. In group sales, again, we expect mid-single-digit sales growth; in core EPS, high single-digit core EPS growth, and we do believe that we can continue to further increase the dividend in Swiss francs.

With that, I hand it over to Teresa or to Alan, yes.

Yes, welcome from my side as well. As Thomas said, great pipeline progress. And that really then combined with great financial results. I think that's really -- that's a great setup, and thanks to the whole Roche team for delivering that. So let's go into the results right away. And here's the overview. And I will focus on the right-hand side, so really the changes in constant rates. Sales plus 7%. Matt and Teresa will go into this, 9% on the Pharma side, 2% on the Dia side.

As said, I think on the Dia side with the impact from China. Matt will highlight this. The core operating profit, plus 13%. So you see really good cost containment, but at the same time, we have invested where it really matters. And I will lead you through this. And then Thomas mentioned it, slower momentum on the core net income and the core EPS, yes, driven by a higher tax load, roughly CHF 600 million, CHF 579 million more taxes that we had to pay, which brought the momentum a little bit down.

IFRS net income up 58%. And you know where it comes from. It comes from a base effect from last year. We had 2 goodwill impairments accounted for CHF 3.2 billion negatively. If you adjust for that, I think it would be a plus 20%, which, in my opinion, mirrors very well the operational performance. Well, now I come to the cash flow. And we can now debate quite a bit about the cash flow. I think you see it, CHF 16.2 billion, down from CHF 20.2 billion. Let me say here, we had really a very strong December when it comes to sales and quite an increase in accounts receivables.

They will convert into cash. So automatically, what I'm saying, I'm expecting quite a strong cash year 2026. The other piece here is in the net trade working capital inventories. We had the situation that we had to deal with the tariffs. So we brought inventories up a little bit. That contributed to this. And last but not least, we have invested more into intangible assets, roughly CHF 600 million. I come to this, but I think that explains the number very well. As I said, I think we will recover quite well on the cash flow side in 2026.

Let me say, I will also highlight this net debt came down at the same time. I will lead you through this. Good. I think when you look really at the bridge here for the sales, in constant rates, 7% up, as you can see. When you go from the left-hand side to the right-hand side, it's a plus 2% because we have the currency impact in, which is quite significant with minus 5 percentage points. Let me focus on the middle. You see Pharma and the loss of exclusivity of minus CHF 745 million, in total, a plus 9%, as mentioned.

And then you see the situation in Dia, where we have a 7% growth, excluding China. And we have then an impact of minus CHF 579 million coming really from the healthcare pricing reform in China, which means a minus 24% of sales in China itself. Good. With that, let's go through the P&L. I talked about the sales growth. Other revenue, I think on one hand, we had a lower milestone income compared to last year, minus CHF 87 million, but we also had higher royalty income. That's why this is very, very stable and really looks good.

Cost of sales, both divisions increased their cost of sales by 7%, but with very different dynamics. I think we in -- from a volume point of view in Pharma, plus 13%. So very, very clearly, I think here a driver for the plus 7% growth on the cost side. I think that growth was also a little bit supported by royalty expenses.

And the Dia division, plus 4% on the volume side compared now to a plus 7% growth on the cost side. So what is that triggered by? Well, very clearly, the healthcare reform plays a role here, so China once again, but also higher costs related to placing of machines, which certainly fuels our future growth of diagnostics, so well invested here. We had certainly investments into the new technologies like CGM, Lumira, et cetera. And last but not least, we had a tariff impact on the Diagnostics side of CHF 64 million half year impact.

So it's also something we potentially have to deal with in 2026. Core operating profit up 13%, a nice margin increase. When you look at the margins itself, I think really in constant rates, overall, plus 1.9 percentage points for the group. You see a nice progression on the Pharma division. You see a decrease of minus 1.1 percentage points in CER on the Diagnostics division side, which is explainable given that here, a significant portion of the sales in China went away.

When you look at the core financial results, and really here, interestingly, I think really in CER, we have an increase. When you look at Swiss francs, we have a decrease. And I think really what that speaks for is, well, the U.S. dollar is weak, hurts us a little bit in the P&L, but helps us with the financial result. That's one conclusion here. You see the equity securities with minus CHF 88 million. I think the market has recovered, but we have some investments in the Roche Venture Fund where we wait for data.

So hopefully, a better year ahead. Net interest income, we had less cash available, led to less income here. Interest expenses was plus CHF 69 million; in concentrate, it would be rather flattish. But let me say here, certainly, the weak U.S. dollar helped. And then we have really here other, and these are predominantly less hyperinflation impacts compared to last year. So with that, let's go to the tax rate. And Thomas mentioned it, I mentioned it already. Let's focus on the middle of the slide.

First, you see really the effective tax rate full year 2024, excluding the resolution of tax disputes, 18.1%. And then you see really the effective tax rate full year 2025, excluding the resolution of tax disputes, 19.5%. So you really see the increased momentum here. Both years were really mitigated by the resolution of tax disputes. In 2024, that was a positive of CHF 263 million, representing a minus 1.4 percentage points.

And in 2025, it was a lower effect, plus CHF 185 million in constant rates, resulting in minus 0.9 percentage points, leading to the effective tax rate full year 2025 with 18.6%. Well, for 2026, I think we will hover more to a 20% tax rate. Core EPS. On the core EPS side, this is the bridge here. I think for me, the most important point to say is it's driven by operations. The increase in core EPS is driven by operations.

And I think there's nothing better to say here. Look, I think the product disposals, you've seen in the P&L, I think less income coming from that. Financial income and expenses are negative in constant rates, the slides in constant rates, that's a negative. It's roughly CHF 60 million. And then effective tax rate changes, as said, this is the once again mentioned increase on the tax side that we have seen here. So operations was the major driver. All other effects on that slide worked against us.

When you look at non-core, and the IFRS income, I've mentioned the IFRS income, see it on the right-hand side, plus 58%. Core operating profit, I've mentioned as well with plus 13%. Perhaps 2 points to mention on that slide. One is the global restructuring plans. You see really the restructuring charges have increased by roughly CHF 300 million. I would argue that's a positive because that gives us savings in the future.

And then I think you see the impairments of intangible assets, as mentioned, not a lot of impact in 2025. In 2024, with CHF 4.6 billion negatively, an enormous impact, very much driven by the 2 goodwill impairments for Spark and Flatiron accounting both together, minus CHF 3.2 billion. With that, let's go to the cash. And here's the story. And look at the left-hand side, CHF 20.2 billion, I've mentioned that already in 2024. And then you see in constant rates, full year 2025 with CHF 17.7 billion.

Well, you see really when you go back to the left-hand side, the operating profit, net of cash adjustments. I think that's the positive momentum coming from operations, really positive. And then you see the net working capital movement. And out of that net working capital movement's, minus CHF 2.2 billion comes from net trade working capital. And as I said, predominantly driven by accounts receivables. Let me mention here that has nothing to do with extended payment terms for certain products and nothing to do with Vabysmo.

This is really we have for -- especially for Vabysmo payment terms for a longer period for the last couple of years, and we have not changed them. So this is really because we had a strong December and that brought the accounts receivables up. We have the higher inventories on the Pharma side. That explains the minus CHF 2.2 billion. We have the investments in PP&E, placements in Diagnostics for the positive site investments that we have done.

And then we have the investments in intangible assets, the increase of CHF 645 million, very much driven by Zealand and the deal we have done there. Then foreign exchange kicks in with minus 8 percentage points quite significantly leads us to an operating free cash flow of CHF 16.2 billion. Well, I think when you look at the margins, I think that tells the same story here. When you look for it, I can just really point back to the fact that we will recover in 2026.

Now when I talk about cash, I have to talk about the net debt development and debt in total as well. And as said, I think interestingly, when you look really at net debt at the end of 2024 with minus CHF 17.3 billion. If you compare that on the right-hand side with the net debt position at the end of 2025, we have reduced by CHF 1.1 billion. So you might ask yourself, okay, the cash flow was not so strong. How is that possible? Did they invest less, especially on the M&A side? No, we didn't.

I think that the numbers are really on the lower part of the slide. As you can see on the right-hand side, you see what we've invested in intangible assets and in M&A is pretty equal, CHF 4.6 billion in 2024 and even more in 2025 with CHF 5.1 billion. One driver here is the weaker U.S. dollar. And you see it really when you look really on this bar, minus CHF 10.7 billion, dividends, M&A and alliance transactions and other, there is the currency translation point with plus CHF 1.8 billion. That is a major driver here and helps us on the debt side.

By the way, we have decreased gross debt by CHF 3.1 billion. And certainly, as 70% of our gross debt is in U.S. dollar, the U.S. dollar helps in that sense to bring the debt down, at least when we report in Swiss francs. Good. With that, a quick comment on the balance sheet. Not too much to say. When you look really on the left-hand side, where we have the assets, I think cash and marketable securities went a little bit down. Nothing to mention here, paid the dividend, all of that.

When you look at other current assets, well, higher trade receivables, as mentioned already, mostly coming from the Pharma side. And when you look really at the noncurrent assets, that was driven by higher intangible assets of CHF 4.1 billion, mostly acquisitions accounting for plus CHF 2.5 billion. On the right-hand side, you see the liabilities and the equity. The current liabilities increased mainly due to the higher accounts of payables and bank creditors, some loans here.

And the noncurrent liabilities decreased slightly due to the decrease in long-term debt. I've mentioned the CHF 3.1 billion already. Leads us to an equity increase quite nicely and now an equity ratio of 38%. Good. Leads me to the currencies. Well, yes, I think really, the volatility is certainly disturbing. And the weak U.S. dollar is something we're fighting against. You see really the result for the full year, minus 5 percentage points on sales and minus 8 percentage points on core operating profit, a minus 7 percentage points on core EPS. To be honest, if you apply today's rates, that would be basically the same picture for 2026 if we keep all that rates from today until the end of the year 2026. So I think it would be basically the same impact. When you compare at year-end 2025 and you keep these currency rates stable until the end of 2026, you see it really on the box down low. The impact would be minus 4 percentage points on sales and minus 6 percentage points on core operating profit and core EPS.

This topic remains in 2026. Good core EPS. I think really we want to set the base right for you for the core EPS and what is the starting point because we have currency effects in the core EPS. So let me lead you through this. You see on the left-hand side, the CHF 19.46 per share as reported. And then I think really, we adjust for CHF 0.37 to get to the CHF 19.83 per share, which would be the starting base for your calculations in -- or if you like, for the core EPS in 2026.

So let me explain now the CHF 0.37. What you see here, these are the exchange rate effects. This is a result of dividing the 2025 currency losses of minus CHF 273 million as well as the 2025 losses on net monetary position in hyperinflationary economies of minus CHF 48 million. This is shown in Note 4 of the consolidated financial statements on Page 64. So on Page 64, you find these 2 numbers. Net of taxes and noncontrolling interest by the number of diluted shares of 803 million, and this number is outlined in Note 29 of the finance report on Page 126, just to confirm that, which is, I think, quite a positive because it sets the higher bar for us, so to say.

So hopefully, a little bit of a positive for the projections for 2025. Here's the guidance again. Thomas has alluded to it. Let me mention to it, the loss of exclusivity impact that we have estimated or actually that we expect of roughly CHF 1 billion for 2026. So relatively narrow to what we had for 2025 and well in line between the CHF 1 billion and CHF 1.5 billion that we've indicated to you that we will have on an ongoing basis.

And with that, I have the pleasure to hand over to Teresa.

Fantastic. Thanks, Alan. So I'm going to hand it back to Alan.

You hand it back to me. Teresa -- it's hard to bring it to you. But let me make a last comment here. We have a change in our income statement presentation for 2026. Let me say very clearly, has nothing to do with the group in itself. So really, when you look at sales, group core operating profit and the core EPS, the metrics are unaffected. This topic is between corporate and the divisions. And basically, what we're doing is we are centralizing the legal department.

That's what we're doing. And that has quite an impact. When you look at Pharma, we reduced SG&A costs in Pharma by CHF 250 million, roughly, I think, certainly in constant rates, which represents roughly 0.5 percentage point in core operating profit margin. So you should adjust for that in your calculations. On the Diagnostics side, that's roughly CHF 50 million less for the SG&A costs, which represents 0.4 percentage points in the core operating profit margin in CR.

Corporate equally then would increase by CHF 300 million. Just to point that out, as I said, for the group accounts, nothing changed. This is between the divisions and corporate. Just to remind you when you project your margins forward for 2026, especially for the divisions. And now, yes, I have the pleasure to hand over to Teresa.

I'm going to move forward really quickly in case they try and take it away from me again. So let's jump right in. So as Thomas shared the group perspective with you a little earlier, I wanted to provide some additional color on the key priorities for the Pharma division, starting with our focus on delivering the on-market portfolio.

Q4 2025 marks our eighth consecutive quarter of growth. So today's on-market portfolio continues to deliver strong performance with 16 blockbusters across our 5 therapeutic areas. We expect this momentum to continue until 2028. And thereafter, we expect that sales become stable to fully compensate for generic erosion. Importantly, we do not expect a patent cliff. As Thomas mentioned, though, in the near term, we are expecting to deliver multiple key launches, which come on top of today's on-market portfolio.

This includes Gazyva in immunology, giredestrant in breast cancer, fenebrutinib in MS, vamikibart in UME and Enspryng with additional indications in both neurology and ophthalmology. We expect that these products are going to continue to extend our growth momentum until well beyond 2028. We are currently in the process of updating our mid- to long-term outlook, and we'll be sharing that with you a little bit later this year. And so while we're very excited for these upcoming launches, we are just as excited about the progress that we've made on our pipeline.

As Thomas mentioned, through R&D excellence and rigorous application of the bar, we have successfully rejuvenated our pipeline. With our post-bar NMEs, we have the potential to enter new disease areas like Alzheimer's and obesity, and we aim to bring multiple new transformational medicines to patients. As I mentioned at Pharma Day, all of our activities are really driven by 2 key tenets: discipline in the business and rigor in the science. Discipline in the business, we remain committed to keeping our COP at least stable.

And rigor in the science, optimizing how we spend our R&D budget and applying our bar criteria to each and every asset progressing in the pipeline or entering it, including from partnerships or acquisitions. I am truly excited and confident for the future of Pharma, delivering transformative medicines and sustaining our growth momentum. So now let's take a closer look at how this momentum played out in 2025, and we'll start with the full year sales. So as you heard from both Thomas and Alan, Pharma sales grew at 9% at constant exchange rates, reaching CHF 47.7 billion. All regions are delivering strong performance, led by our international region with 14% growth. And overall, our volumes were up by 13%. As Alan mentioned, COP increased by 13% versus that 9% sales increase with a COP margin of 49.2%, so a slight increase over last year.

Clearly, COP grew ahead of sales, which we mainly attribute to effective cost management, particularly in R&D, but I am going to drill down on the individual line items in a little more detail. Other revenue slightly decreased by 1% with higher profit share income from the higher sales of Venclexta in the U.S., which was offset by lower income from our out-licensing agreements.

As Alan mentioned, cost of sales increased 7% against a 13% volume growth. R&D costs declined by 3%. This was mainly driven by savings in Flatiron as well as some other operational efficiencies. But let me say that this reduction was very thoughtful and deliberate as it gives us the oxygen that we need for the upcoming CVRM Phase III trials. SG&A costs increased by 8%, and this is primarily for 2 reasons. It was driven by some investments in our growth drivers, particularly Ocrevus and Xolair, but also increased donations to multiple independent co-pay assistance foundations.

As part of our broader corporate philanthropy strategy, Genentech doubled its donations to those independent co-pay assistance foundations in 2025 versus 2024. Our corporate giving strategy is focused on supporting the patients most in need across multiple therapeutic areas, including oncology, neurology, immunology and ophthalmology. And we continually evaluate the strategy to ensure that it remains aligned to patient needs. And finally, other operating income and expenses decreased by 43%, and this is primarily due to lower gains on the disposal of products.

And so now let's look at our individual growth drivers. So as always, first, I have to comment on the graph. These are all absolute values and year-over-year growth rates are presented at constant exchange rates. At full year, our top brands, Phesgo, Xolair, Ocrevus, Hemlibra, Vabysmo and Polivy generated roughly CHF 3.6 billion in new sales at constant exchange rates. For the fourth quarter in a row, Phesgo is our #1 growth driver with 48% growth, closely followed by Xolair, driven by the continued outstanding uptake in food allergy.

You'll also notice the strong performance of Xofluza. And I just want to talk for a minute about that here, which is driven by the strong flu season that we saw in Q1 of 2025 in China. This may create a bit of a base effect for Xofluza in 2026, especially considering that we haven't seen a similarly strong flu season in China this year. So now let's dive into our TAs, starting with oncology. Oncology sales increased by 2% to CHF 15.3 billion, primarily driven by our HER2 franchise.

As I mentioned, Phesgo posted an impressive 48% growth and the global conversion rate keeps climbing. We're now at 54%, well on our way to our new goal of 60%. This, of course, also means that Perjeta conversion to Phesgo continues to impact Perjeta sales, which is to be expected. Kadcyla growth continues to be driven by uptake in adjuvant breast cancer. Looking forward to the HER2 franchise overall, and as I've said previously, we expect the HER2 franchise to peak this year at about CHF 9 billion at 2024 exchange rates, followed by a steady decline through the end of the decade with a solid tail of around CHF 4 billion.

And that CHF 4 billion is primarily Phesgo, about CHF 1 billion for Kadcyla and a bit of HMP. We do not foresee a biosimilar in the U.S. for Perjeta until the end of 2027. And let me also confirm again that we do not expect a cliff situation for the HER2 franchise. For Itovebi, we see good launch momentum in our first-line PI3 kinase HR-positive breast cancer population, and we expect 2 Phase III readouts this year with INAVO121 and 122, which could enable further indication expansion.

But of course, the highlight of Q4 was the positive Phase III lidERA result for giredestrant. I am going to cover this in more depth on the next slide, but let me give you a few quick updates on giredestrant in general. We presented the lidERA data at San Antonio Breast, and we have already filed the evERA results with the FDA that happened at the end of last year, and EU filing is expected for 2026.

Therefore, we expect evERA U.S. approval later this year, lidERA results will be filed with the U.S. and EU regulators this year. In the first half of this year, we expect the readout of persevERA in first-line ER-positive, HER2-negative metastatic breast cancer. Moving on to Tecentriq. Tecentriq exited 2025 with 3% growth and actually quite a nice Q4. For 2026, we expect low single-digit growth for Tecentriq, driven by the positive studies that we shared last year, such as IMforte in small cell, IMvigor in MIBC and ATOMIC in dMMR colon cancer.

And finally, before I move to the next slide, let me just briefly mention that we expect the first Phase III readout for our KRASG12C inhibitor, divarasib, later this year. So now let's take a closer look at the lidERA results for giredestrant. So here are the giredestrant-lidERA results in adjuvant ER-positive HER2-negative breast cancer, which we presented last December.

As you can see, giredestrant demonstrated a statistically significant and clinically meaningful improvement in invasive disease-free survival versus standard of care endocrine therapy, achieving a hazard ratio of 0.7. Let me emphasize here that this is the first oral SERD to show superior IDFS versus endocrine therapy in the adjuvant setting. And in terms of overall survival, the data was still immature, but clearly, a positive trend for giredestrant was observed with a hazard ratio of 0.79.

Giredestrant safety profile remains favorable as we've seen in previously -- previous studies. And importantly, we saw a lower discontinuation rate with giredestrant versus the comparator arm. This is a significant improvement in this setting, and it indicates the improved patient experience on giredestrant compared to standard of care. So taken together, these results further underline giredestrant's potential as a next-generation best-in-class endocrine therapy in ER-positive breast cancer.

To expand on this, let me dive a little bit deeper into the giredestrant development program. So now given the positive evERA and lidERA results we just discussed and the upcoming readouts, we're very excited for the future of giredestrant. Giredestrant has the potential to replace standard of care endocrine therapy in ER-positive breast cancer and become the new backbone of choice in this setting. As you can see in the treatment paradigm on the left, our clinical development program for giredestrant covers different lines of treatment and risk groups with the readout of persevERA in first-line expected mid-first half of this year.

Please note that the giredestrant plus CDK4/6 combination in lidERA relates to a single-arm substudy that's still being evaluated. That's in combination with [ abema ]. We have also started a sub-study in combination with [ ribo ] as well. As you would expect, we are working at speed to complete filings and collaborate with regulators to ensure that this transformational medicine gets to patients as fast as possible. And just to reiterate, we have already filed evERA in the U.S. and expect lidERA filing in Q1. Beyond giredestrant, we also have a strong pipeline of potentially best-in-class molecules in breast cancer that give us the opportunity for numerous future combinations.

For instance, our highly potent CDK4/2 inhibitor, which may overcome some of the limitations of currently available CDK4/6 inhibitors. And as I mentioned, we believe giredestrant has the possibility to become the new backbone of choice in ER-positive breast cancer. And therefore, we're exploring many combinations with some of the internal assets, as I just mentioned. Additionally, our phone keeps ringing as we are getting calls from potential partners interested in combination studies with giredestrant.

We look forward to sharing future updates on our breast cancer pipeline with you in the near future. But for now, let's move on to hematology. The hematology franchise delivered strong growth of 15% in 2025, achieving CHF 8.6 billion in sales. Hemlibra closed the year with strong growth, 12% (sic) [ 11% ] driven by increasing adoption in the non-inhibitor patient population. For 2026, we expect low single-digit growth, and that's partly driven by competitor launches, which are anticipated later in the year.

Polivy's growth momentum continued in 2025, reaching U.S. patient sales of 36% in first-line DLBCL. But in fact, we reached 2 significant milestones with Polivy last year. First, it is now the most prescribed regimen for IPI 2-5 patients in the U.S. And second, we have officially hit more than CHF 1 billion in sale in the first-line DLB setting alone. Shifting to Columvi and Lunsumio, our CD20, CD3 bispecifics. Launch performance remains on track for Columvi in third-line plus DLBCL with second-line DLBCL launches gearing up. For Lunsumio, we're happy to report that the subcutaneous formulation has been approved in both the U.S. and the EU.

And just a reminder that, that new formulation reduces administration time from hours down to actually under a minute. Additionally, we expect 2 key events for Lunsumio later this year. We expect U.S. approval for Lunsumio plus Polivy in second-line DLBCL based on the positive SUNMO results, and we expect the readout for the Phase III CELESTIMO in second-line plus follicular lymphoma.

So now let's move on to neurology. Our neurology franchise achieved CHF 9.8 billion in sales in 2025 with a strong growth of 11%. Ocrevus continues to have good momentum, delivering 9% growth globally and crossing the CHF 7 billion milestone in annual sales. We're excited to see the increasing growth momentum of our subcutaneous formulation known as Zunovo in the U.S. In Q4, more than half of global Ocrevus growth was driven by the subcut formulation. And importantly, in the U.S. and many other early launch countries, roughly 50% of Zunovo patients are naive to an Ocrevus. This represents that acceleration that we've been talking about. U.S. uptake continues to be driven primarily by community practices, which emphasizes how Zunovo is actually expanding the addressable market and can help overcome healthcare system restraints like IV capacity limitations.

Overall, we now have more than 17,500 patients on Ocrevus now globally, and that's roughly 5,000 more than we had at Q3. For 2026, we expect to hit high single-digit to low double-digit growth for Ocrevus. And as a reminder, we upgraded our peak sales expectations for Ocrevus for the Ocrevus franchise to CHF 9 billion by 2029. This includes CHF 2 billion of incremental sales from Ocrevus subcut. But of course, there's also going to be some switching from IV to subcut.

Staying with the MS franchise, you've seen the exciting news regarding the positive 3 results for fenebrutinib. We're going to cover that more on the next slide. But for now, let's take a minute on Evrysdi. The global rollout of the tablet formation continues, and we see great pickup from that as well as very positive feedback from the patient community. As Thomas mentioned, this remains the leader in SMA. Quick note that Q4 performance for Evrysdi in international was boosted by a tender-related buying pattern, but we are still expecting double-digit growth for Evrysdi next year.

Earlier this week, you saw positive data from Elevidys in DMD. We continue to believe in the positive risk-benefit profile in the ambulatory DMD population and more than 1,050 patients have already been treated globally in this setting. Furthermore, the latest 3-year data from EMBARK shows the durable efficacy and slowing of disease progression for ambulatory DMD patients treated with Elevidys. We are working with EMA continually to find a viable path forward for EU patient access here.

Quickly stopping over prasi, a quick update here where we have achieved both FPI for the Phase III study as well as we have been able to materially accelerate site activation. So we're a number of months ahead of schedule with the prasi trial, which is great news for patients. And let me close quickly by speaking a little bit about Enspryng in MOG-AD. So MOG-AD, if you are not aware, is a rare antibody-mediated autoimmune condition of the central nervous system, which causes inflammation of the brain, optic nerve and spinal cord.

The Phase III study METEOROID read out positively. And we're looking forward to presenting that data at an upcoming medical conference later this year. We expect to file these results with the U.S. and EU regulators in 2026, and this additional indication could unlock an upside of approximately CHF 500 million for Enspryng. So now as promised, let's take a little bit of a deeper look at fenebrutinib. We are very excited about the positive Phase III readouts for fene. This includes FENtrepid in PPMS and FENhance 2 in RMS with the FENhance 1 readout expected mid half of this year.

These results make fenebrutinib the only BTK inhibitor with positive Phase III results in both RMS and PPMS, and it has the potential to be both first and best-in-class in RMS and PPMS, which would also make it the first and only high-efficacy oral treatment for both relapsing and progressive multiple sclerosis. We see fenebrutinib as an opportunity to increase high efficacy treatment rates amongst MS patients and expand the footprint of our franchise. Ocrevus and now Ocrevus subcut have brought transformational impact to people living with MS, and we believe fenebrutinib has the potential to be that next transformational medicine for these patients.

Let me also briefly remind you that fenebrutinib is differentiated by design from other BTK inhibitors. It is the only noncovalent binding BTK in Phase III development for MS and has a highly optimized PK profile that allows it to reach its target, including in the brain. So stay tuned for the FENhance 1 readout in half 1. And until then, we look forward to presenting the FENtrepid results in PPMS at ACTRIMS, where we are also inviting you to attend our IR event on the 9th of February.

And so with that, let's move on to immunology. Our immunology franchise grew at 12% at constant exchange rates and reached CHF 6.7 billion in sales. Xolair's strong growth momentum continues, driven by uptake in food allergy. In 2025, we achieved 32% growth in sales of CHF 3 billion. We are also happy to celebrate a key Xolair milestone in 2025, which is more than 100,000 patients have now been treated for food allergy since launch.

Regarding the 2026 outlook for Xolair, we expect around 20% growth, and this includes the impact of an expected first biosimilar entering the market in the second half of the year. You will have seen that Xolair was selected for the latest rounds of IRA negotiations. So let me provide a little bit of extra information on this. Xolair's inclusion on this list, as you know, does not change patient access or pricing at this time. Any potential pricing impact, if applicable, would not take effect until 2028 at the earliest.

But CMS' final guidance provides that a selected drug will no longer be subject to negotiation and will cease to be a selected drug if CMS determines that a generic or biosimilar has been marketed by November 1, 2026, and we do expect that a biosimilar for Xolair will be launched before that date. Actemra sales declined by 2% in 2025. As predicted, we are now seeing increased biosimilar impact in the U.S., which resulted in a 10% decline in growth in Q4. This is aligned with all of our previous communications of an accelerating biosimilar impact in the second half of 2025, which will obviously continue into 2026.

Just like in Q3, Gazyva is one of our key highlights for the quarter. Following the FDA approval in Q3, we achieved EU approval in lupus nephritis, and we announced positive Phase III readouts in both SLE and INS. In both indications, Gazyva has first-in-class potential. SLE results have been submitted for presentation at SLE Euro in early March, and the INS results have been submitted to WCN in late March. Both indications, as I mentioned, I think, previously, will be filed in the U.S. and EU later this year.

And I'm also very happy to share that the FDA has granted breakthrough therapy designation for Gazyva in childhood onset idiopathic nephrotic syndrome, which is INS based on the positive ENSURE results. And we are not quite done with Gazyva just yet. There's one more Phase III trial, which is expected to read out in 2026, and that's MAJESTY in membranous nephropathy. And as a reminder, we see up to a CHF 2 billion opportunity for Gazyva in kidney disease.

And just finally, I'd like to mention the upcoming Phase III readout for sefaxersen in IgAN, which is expected later in the year. Now let's move on to ophthalmology. Ophthalmology grew by 10%, achieving CHF 4.2 billion in sales. Vabysmo performance, as you know, was impacted by the contraction of the U.S. branded market. It landed at 12% growth for the year, which is still quite strong. We had mentioned this contraction previously. And through 2025, we saw a decline in the branded IBT market in the U.S. of about 15%. Nevertheless, Vabysmo continues to gain market share in the branded IBT market in the U.S. and across early launch countries globally. In the U.S., we now see that more than 60% of Vabysmo patient starts are from treatment-naive patients, and this further solidifies Vabysmo's position as the standard of care.

Looking forward, we would expect the U.S. branded market to gradually recover in 2026. And taking this into account, we expect a growth acceleration in 2026, driven by the ex U.S. continued growth and U.S. recovery. In fact, as Thomas mentioned, there is a lot to look forward to in ophthalmology this year. We have 2 potential new medicines entering our ophthalmology portfolio. That's the vamikibart in UME, which is expected to be filed in both the U.S. and EU. Enspryng in thyroid eye disease will be filed in the U.S., and we are currently considering ex U.S. filings with the appropriate regulators.

Now let's jump into our CVRM pipeline. This is one slide with a whole bunch on it, but I am very happy to share with you the key developments in our pipeline as well as provide a perspective on a very newsflow-rich 2026. So earlier this week, we shared positive final Phase II top line results at week 48 for the once weekly CT-388 in people with obesity. This is study 103. For the efficacy estimand, we achieved a placebo-adjusted weight loss of 22.5%. As a reminder, the efficacy estimand includes patients who dropped out from further analysis, so the effect size measured represents the true efficacy of the medicine tested.

For the treatment regimen estimand, we achieved a placebo-adjusted weight loss of 18.3%. The treatment regimen estimand reflects a more real-world outcome, acknowledging the fact that not all patients will be able to adhere to treatment. In this case, data after treatment discontinuation, either in the treatment [ or ] placebo arm are included in the analysis. So for example, it includes data from patients who discontinued treatment early and have regained weight.

Now this is a question we received a number of times over the last week. So I'm just going to take another minute here to reiterate. Generally speaking, the difference between the efficacy and treatment regimen estimands is usually driven by treatment discontinuation, either due to patients on the active treatment arm who regained weight after discontinuing treatment or patients on placebo who go on a weight loss therapy after discontinuation. There are many ways to potentially address this phenomenon in our future Phase IIIs from a more flexible dosing regimen, which allows patients to stay on lower maintenance doses in case of tolerability issues to the incentive of a long-term extension to retain more placebo patients.

But these kinds of measures should serve to improve the discontinuation rate and eventually reduce the gap between the 2 estimands. In that context, I should also point out that in most of the recent Phase III trials in obesity, marked differences between the estimands greater than 5% have been observed. Let me also highlight 2 other key points in terms of the efficacy achieved in the study. First, we saw a clear dose-dependent relationship on weight loss. And secondly, and most importantly, we are pleased by the absence of a visible efficacy plateau at 48 weeks for the highest dose tested, which was 24 milligrams.

Taken together, this clearly indicates that further weight loss can be achieved after 48 weeks, and it gives us confidence in CT-388's potential to deliver best-in-class efficacy for obesity. In terms of safety and tolerability, CT-388 was well tolerated and the tolerability profile was generally consistent with incretin class. The majority of gastrointestinal-related events were mild to moderate and total treatment discontinuations due to AEs in all arms were low at 5.9% for CT-388 versus 1.3% for the placebo arm.

Let me also highlight here as we received this question a number of times as well, the discontinuation rate due to AEs at the highest 24-milligram dose was similar to the total discontinuation rate observed. We look forward to sharing more detail on the Phase II results with you at an upcoming medical conference later this year.

Similarly to 388, we saw positive results for CT-868 in the Phase II 004 study in type 1 diabetes. And just like for CT-388, we will share the final results at an upcoming medical conference in 2026. So speaking of the outlook for the rest of the year, let's start with our Phase II and Phase III study initiations. As a reminder, we announced for both CT-388 and CT-868 that we will move them into Phase III development in 2026.

For CT-388, we can now provide a first update. The Phase III trials for CT-388 named Enith1 and Enith2 are now scheduled to start in Q1. In addition, you can see that we plan to initiate the first Phase II studies for petrelintide in CT-996 as well as a Phase II combination study for CT-388 with petrelintide. In addition, we have a number of other CVRM readouts scheduled for 2026. There are multiple Phase II readouts to look forward to. For CT-388, we have data for patients with obesity and with type 2 diabetes, which will come later this year.

We also expect the first Phase II readout for CT-996, our oral GLP-1 and for petrelintide, ZUPREME 1 and 2 trials in obese overweight patients with and without type 2. And finally, emugrobart and tirzepatide combination data in obesity are expected towards the end of the year. So as you can see, we continue to progress our CVRM pipeline at pace, and we are excited to share updates with you throughout the year.

So last but not least, let's go to the next slide to bring us home. Here, we have the 2026 pharma key newsflow. We start the year with 4 green check marks, certainly a good omen for the year ahead. And we have discussed everything else on previous slides, so I won't go into more detail here. For any of you who are feeling the lack of the 2025 newsflow table, we have moved that to the appendix. And with that, I would say, I'll give it back to Alan but I'll be crazy and I'll give it over to Matt.

That's wild.

Wild. Wild times.

All right. Thank you, Teresa. Good morning, good afternoon, everyone. It's my pleasure to present the full year 2025 Diagnostics division financial results. So with sales, as you heard from Alan and from Thomas, sales in diagnostics were CHF 13.8 billion. We grew 2% or CHF 292 million compared with 2024 at constant exchange rates. But as you heard from both Thomas and Alan earlier, excluding the sales in China, which I would reiterate is our second largest market, was impacted by health care pricing reforms. The growth of the diagnostics business was 7%. So now let me walk you through these results by each of our customer areas. So sales in our largest customer at Core Lab were flat, again, driven by this previously mentioned health care pricing reform.

Excluding this effect, sales were plus 10%. Sales in the Molecular Lab increased 4% due to growth in our blood screening business. Now this was partially offset by reduced sales growth in the infectious disease segment, which grew at 1%. This was impacted by the USA funding stop in Q1 that caused a corresponding decrease in HIV testing, which I covered last year. Sales in our Near Patient Care customer decreased at minus 3%, mainly driven by the decline of our blood glucose monitoring business at minus 2% due to the market shift to continuous glucose monitoring as well as a decline in respiratory molecular point-of-care testing due to the late start of the 2025 respiratory season. And again, back to what you heard earlier from Thomas, we expect the CGM product to really be a driver for this customer area in the future and we continue to invest in expanding and preparing for this.

Finally, sales in the Pathology Lab grew strongly at plus 14%, mainly driven by sales of advanced staining at plus 10% and our companion diagnostics business, which grew at plus 25%. So now I'd like to show the geographic performance that's behind these results. Taking through the regional view, North America, the business grew at plus 9%, well ahead of market. You saw good growth in EMEA at plus 6%, again, ahead of market. Latin America, strong growth at plus 11%. Now Asia Pacific, again, as we discussed, minus 12%, driven by the minus 24% decline in China. Excluding the effect of China, APAC grew at plus 4%. Now as you heard earlier, our consistent ambition in the Diagnostics division is to grow our sales at mid- to high single digits. However, given that we anticipate diminished but continuing headwinds in China for 2026, we would set our ambition this year at mid-single digits for 2026. Again, our consistent ambition is to grow this business at mid- to high single digits.

Now I'd like to walk you through the P&L line by line. As previously mentioned, sales grew at plus 2%. Cost of sales, as you heard from Alan, grew at plus 7% but this was mainly driven by that unfavorable impact of the China health care price reforms, half a year impact of tariffs and the production ramp-up of our new technologies such as CGM and sequencing and the placement of a significant number of instruments in 2025. And I would highlight that we saw growth of some of our key platforms like our immunoassay at strong double-digit increase. And for example, our molecular workstation, the 5800 grew at over 40%.

So very strong placement of instruments. R&D costs decreased at minus 2%. Now this is a result of significant and focused cost containment measures across the organization in response to China impact. As mentioned previously, we are ensuring delivery on all our key priorities, especially our investment in the key new product areas such as CGM and our AXELIOS Sequencing Solution and LumiraDx. SG&A decreased by 2%, again, reflecting focused cost containment measures across the organization. This resulted in a core operating profit of approximately CHF 2 billion, declining at 4% at constant exchange rates, which reflects the cost control initiatives.

So now I would like to transition to some of the innovation that we launched in last year and really specifically focus on our cobas Mass Spec 601, which as you heard from Thomas, these are CHF 1 billion opportunities that we're very excited about and their potential to really deliver growth to the Diagnostics division. So as I mentioned before, current mass spec primarily relies on lab-developed tests and lack automation, are highly manual and require highly skilled labor. With the launch of our mass spec solution in 2024, we've introduced the first fully automated IVD platform for clinical mass spec.

So throughout 2025, we received CE mark for all of our wave 1 menu composed of 39 analytes spanning the key parameters used in mass spec testing, including therapeutic drug monitoring, steroid and hormone analysis as well as vitamin D testing. These comprise the majority of parameters used in a routine clinical mass spec lab and we are going to follow that with a second wave of additional parameters. I would add that this system, which integrates with our existing serum work area platforms, strengthens our leading position in the Core Lab.

And now I'd like to talk about some of the high medical value content that we launched last year for our serum work area, specifically our dengue antigen test, which received CE mark in October. Dengue is the most common mosquito-borne viral disease globally and represents a major global health burden. It accounts for an estimated 390 million infections per year. It has shifted from being a seasonal illness to a year-round risk with locally transmitted cases shifting from historical geographies such as South America to now in Europe and North America. Diagnosing dengue can be challenging as patients are often misdiagnosed due to overlapping conditions with other febrile illness. With our Elecsys antigen test, we will enable health care systems to diagnose dengue more reliably and efficiently by providing all 4 dengue virus serotypes differentially diagnosed with a rapid test that takes only 18 minutes. This will add one more test to our leading immunohistochemistry platform -- or excuse me, immunochemistry platform, which comprises of approximately 120 different parameters.

So now I would like to move on to a customer who're very near and dear to my heart, the Molecular Lab and switch to discussing our cobas BV/CV assay, which we received CE mark in December. Sexual health diagnostics market is valued at CHF 1.1 billion with a yearly growth rate of 11%. Vaginitis is the primary growth driver within this segment showing a yearly growth rate of 26%. With our cobas BV/CV assay, we will provide a multiplex assay designed for the direct detection of bacterial vaginosis and candida vaginitis and expand our molecular STI offering. With the addition to our STI portfolio, we will continue to enable testing the most commonly sexually transmitted infections using a single tube and a vaginal swab. In the future, we plan to continue expanding our offering in this area with home collection solutions as well as novel molecular point-of-care assays.

Transitioning to our point-of-care portfolio, I would like to discuss our recent CE mark and FDA clearance with a CLIA waiver for our liat Bordetella panel. Pertussis is a highly contagious disease that causes more than 24 million estimated yearly cases, resulting in 160,000 deaths with the majority of those in children. Diagnosing pertussis can be particularly challenging as its symptoms often overlap with those of common colds, leading to underdiagnosis. Our liat Bordetella panel offers a reliable point-of-care solution, delivering results in just 15 minutes between 3 Bordetella pathogens and again, delivers lab-like performance. This will enable health care providers to act quickly and prevent severe complications, especially in vulnerable populations such as children. As you can see from this slide, this launch, we further expand our cobas liat menu of lab equivalent point-of-care testing and we will continue to expand this in the future.

And with that, I would like to transition to our key launches in 2026 and call out a few highlights. Again, I would really want to emphasize that 2026 is the year that we will launch our AXELIOS Sequencing Solution. This is a groundbreaking high-throughput solution that will deliver high accuracy, high throughput and flexible sequencing based on our proprietary Sequencing by Expansion technology. And would also again highlight that this represents a potential blockbuster opportunity for us with sales potential and above the CHF 1 billion range. I would also like to call out the expansion of our neurology menu, including the Elecsys pTau 217, which is a blood-based diagnostic for Alzheimer's disease and Elecsys Neurofilament light chain for detection of disease activity in Multiple Sclerosis, greatly expanding our offering in neuroscience.

Additionally, I would like to -- I would really like to particularly mention our TB IGRA test, our assay to detect latent tuberculosis infection, which remains a global health care challenge and a significant commercial opportunity and I'm very convinced that we will offer a very differentiated, highly competitive solution here. And overall, this 2026 is going to be a very exciting year of launches and I look forward to keeping you updated over the course of the year.

Thank you and now I hand it to Bruno.

Thanks, Matt. And with that, we open our Q&A session. The first question goes to Sachin Jain from Bank of America.

Two, please. So firstly, on Vabysmo, I don't think you've guided to growth for this year beyond acceleration. So any color on what you're assuming within the guide? And perhaps, Teresa, you could just provide a bit more color on your funding comments. What does that doubling in '25 versus '24 mean relative to historic levels? Like where is that funding relative to sort of a 3-, 4-year average? And any color on how that flows back to patients? When we should see an impact on sales? The second question is on persevERA, if I may. It's a topic that I think has come up on prior calls but just to reiterate as we approach data. If the study hits, is any hit clinically meaningful for you, would you need to see a certain hazard ratio or absolute PFS benefit -- you used that wording in the press release. The reason for the question is, there's been speculation since your San Antonio call around passing an interim. Those are my 2 questions.

Yes. Great. So in terms of Vabysmo, I'm not going to give you specifics on the amount of money that we contributed because as you have heard me say many, many times before, our charitable giving is not in any way related to our commercial expectations for the product. So those 2 things are and have to be completely separate. I can tell you that we doubled our donations last year and that was a significant increase for us over the last couple of years as you sort of alluded to. We do believe that 2025 represented sort of a rebaselining of the branded market in the U.S. And so what we are hopeful is that 2026 will now allow the underlying growth of Vabysmo to actually be more visible. And so we would expect an acceleration in 2026.

I don't believe we've been more specific than that. In terms of persevERA, so clearly, the fact that we've now seen positive data from giredestrant in a number of important settings, both neoadjuvant, adjuvant and in a complex late-stage population, sort of underscores our belief in this molecule and that clinically, it is potent, it's active and it's combinable, it's tolerable. It's given us great confidence that we do have the opportunity to be really impactful for many different patients and to really become a new standard of care in hormone receptor-positive breast cancer. Reading through though, to different settings is complex. And so thinking about how we would read through to persevERA, happily, we don't have too long to wait to actually get the answer to that question. In terms of what would be clinically meaningful, we have designed the study to yield a clinically meaningful result. And so generally speaking, a 20% reduction would be considered clinically meaningful.

Did I answer your question?

Perfect.

Okay. Very good. Then we move on. Next one in the row would be Peter Verdult from BNP Paribas.

Pete Verdult from BNP. Two questions. Teresa, just on obesity. We understand from Zealand that the amylin data is in-house and the market seems to have set the bar at sort of being low to mid-teens weight loss. Forget the market for a second. Can we just focus on Roche? What is the minimum target profile you are looking to demonstrate for amylin in obesity? And then secondly, on BTK, you sound very confident about the approvability despite recent CRLs elsewhere in the BTK class. You know the efficacy in the first relapsing-remitting study in PPMS, which we don't. Just wanted a sense or kick the tires with you. Is your confidence based on a highly skewed benefit risk profile? Or is it more because you think the 2 cases of Hy's Law that you've seen in the data set can be attributed to other or nondrug causes?

Yes. So I'm going to start with your second question first because I think we've gotten a lot of questions over the last couple of weeks about tolebrutinib and read-throughs to fenebrutinib. And I think we have to be very, very cautious here. If you actually read that CRL, it is incredibly specific to the risk benefit that was seen with tolebrutinib. And unfortunately, they had a number of failed trials. They had a number of Hy's Law cases. So I think it is very difficult and inappropriate to actually take the language that was applied to tolebrutinib and actually put that forward on to fenebrutinib.

Let me be really clear because I think there have been some -- there's been some confusion about what we've actually seen in terms of Hy's Law cases for fenebrutinib. We had 2 cases of elevated liver enzymes with bilirubin, which was what put us on clinical hold with the FDA. Both of those cases were in FENhance 1, which currently is a study that still remains blinded. When we looked at those 2 cases, only one case was deemed by the FDA to be a Hy's Law case. The other one was confounded due to alcohol use by the patient. And so right now, in fenebrutinib, we have only one case and it is in FENhance 1. So we are sort of blinded to any more detail.

It's also really important to note that since we put liver monitoring in place in the clinical trials, we have not seen any more cases. And so I think we feel very good about the overall benefit risk profile that we have with fenebrutinib, particularly when you consider the other half of that coin, which is the benefit. When you look at the Phase II trials for fenebrutinib, you saw a significant amount of clinical benefit to patients. And the data that we've seen are sort of very consistent. And so I think when you look at that very high efficacy with a very -- what looks to be a very manageable safety profile, I think we're just in a totally different situation than what you saw with tolebrutinib. So hopefully, that kind of provides a little bit more perspective there.

So in terms of petrelintide, so as a monotherapy, we believe that petrelintide holds the potential to be a foundational therapy for weight management. We are looking forward to being able to deliver a weight loss that the vast majority of people are actually looking for, which is something more in that sort of 10% to 20-ish percent with the potential to be a much more improved tolerability profile compared to the GLP classes as well as just a better patient experience in terms of titration, quality of weight loss, et cetera. So obviously, we don't, again, happily have long to wait. We'll see that data soon. You mentioned the data being in-house. We remain blinded to that data. So we have not seen it but we do -- we expect to see it very soon.

Peter, did this answer your questions?

Yes.

And we move on then. Next one would be Simon Baker from Redburn.

Two, if I may, please. Firstly, just continuing on Pete's question about fenebrutinib. I just wonder if you could give us some idea about how we should be thinking about the relative tolerability profile of fenebrutinib versus Ocrevus ahead of the ACTRIMS data? And how do you see in light of that fenebrutinib being positioned relative to Ocrevus? And then secondly, a question for Matt. It's a little while since you unveiled to us the new sequencing offering. I just wonder if you could update us on the market feedback you've had in terms of levels of demand and where that demand is coming from, whether it's smaller scale or larger scale applications or indeed both?

You want to go first?

I would be delighted to go first.

And I would be happy to yield the floor.

Super. Wow. So yes, and I would maybe give a plug for our Dia Day in May, which will talk quite a bit more about this and Bruno will mention that at the close. Maybe I'll just say that first. But yes, we've seen a high level of demand for the sequencer, I would say in, more than we had originally anticipated ahead of launch. We're already starting commercial activities with select customers. And the feedback from our early evaluators has been extremely positive. So when you talk about applications, we're really seeing interest in a broad variety of applications from translational, such as single cell but then on to more focused clinical applications such as whole genome sequencing and germline. So what we're really seeing is the potential of an instrument with that kind of flexibility, throughput and accuracy and a dual assay format with the longest reads of Simplex as well as the very high accuracy Duplex format to have a broad applicability really across the spectrum of sequencing applications. And I think we're very confident in the potential for this technology as well as the launch.

Does that answer your question?

Simon?

Perfect.

Yes. Great. So when we think about where fenebrutinib sits, I mean, we believe that it has best-in-class potential. And together with OCREVUS, OCREVUS subcut and potentially further on -- down the line, OCREVUS high concentration, we believe, ultimately, we are going to have a range of highly efficacious and very tolerable therapies that meet every patient with MS exactly where they're at. Right now, 30% of patients are on a less efficacious oral therapy. And so that's sort of an easy place to imagine fenebrutinib starting. But I think ultimately, we believe that this is -- the combination of these 2 therapies gives us the opportunity to really sort of revolutionize the entire patient journey for MS patients. And I think we're feeling very confident about our ability to do that.

[indiscernible] Very clear.

Next questions go to Matthew Weston from UBS.

Hopefully, you can now hear me. The first one on giredestrant. Teresa, there's a lot of debate about how the commercial potential in the adjuvant setting could be impacted by the data from persevERA. Can you give us your thoughts as to whether or not you see adjuvant as independent of that frontline metastatic result? And also, there's a lot of debate about the peak sales potential of giredestrant.. So when do we -- when should we expect to hear what Roche thinks the potential of this medicine is? And then secondly, if I can just pick up on biosimilar erosion. So Q2, Q3 of last year, you made a number of comments about delays to the entry of Xolair and now similar comments about potential delays to the entry of biosimilar Perjeta. Clearly, there are multiple patents, so you can do deals with biosimilar companies. But do you think investors should get used to a more gradual erosion of some of these biosimilars at the beginning of generic entry? Or should we still continue to expect to see like a minus 40% that has been kind of the underlying trend so far when we actually see biosimilars enter the market?

Yes. So I'll take -- thanks, Matthew, for your questions. I'll take your second one first and I have a very definitive answer for you, which is that it absolutely depends. So it depends on the therapy. It depends on the part of the world. I mean, I think it -- this is one of those things where biosimilar impact is not a one size fits all. So in some parts of the world, with some therapies, you are going to see an immediate decline. With some others like Xolair, we just do expect that to be a smidge more sticky because you're dealing ultimately with a very allergic patient. And so physicians might be a little bit more tentative about switching so quickly. And so I think this is one of those areas where we are constantly monitoring the environment. We're constantly talking to treating physicians to get a sense of how they may think about the utilization of biosimilars and we give to you our best knowledge of how we believe those erosion curves will happen.

But it's very difficult to give you one answer because I think it is actually quite variable, again, by therapeutic area and by geographic area. When it comes to giredestrant, so this market is somewhere between a $20 billion and $30 billion opportunity. Adjuvant is about 2/3 of that between initiation and maintenance therapy. We do think that adjuvant and first-line is pretty separate. And we think that giredestrant has the opportunity, as we said, to really be establishing itself as a new standard of care. When are you going to get a better read-through from that? Q1 is a very data-rich -- here, so the first half is a very data-rich time for us. We're in the process of updating our own assumptions. And as soon as we have a clear read-through, we will share that with you.

Maybe just to answer your question on Perjeta as well because you had this question. So we don't expect the biosimilar for Perjeta until '28 in the U.S. and '27 in the EU.

Correct.

Just to clarify that.

Matthew, did we answer your questions?

That's perfect.

Then we move on. Next one would be James Gordon from Barclays.

James Gordon from Barclays. Two questions, please. One would be on giredestrant, actually 2 subparts. One would be, the slight delay in persevERA readout timing and I think it's now more likely to be Q2. Is that because the event rate is a little bit slower? Or could you be getting a few more events in and could that actually help the [ powering ], which has been a concern some people have had? And also on giredestrant, the lidERA Study, the [ side ] study of about 100 patients, I think they're getting on top of the CDK. How will you communicate that? And it sounds like you're filing ahead of that because you're filing the data in Q1. So is that something that then gets added to the filing package and you hope to have on the initial label? Or how does that work?

And then the other one was on fenebrutinib. So you sound very confident talking about the Vabysmo upcoming launch, which is great. And there's been some talk about liver already. But in terms of other tolerability issues, I saw the comment in the original release about additional safety data that is further being evaluated. So could there be some other off-target BTK side effects you need to think about? And just on the side effect point, though, if you're comparing it to something like Ocrevus, so you've got the advantage of oral but could you have to have liver monitoring or something like that? Could that be a barrier to becoming a very big drug? How would you think about that?

Great. Okay. So we'll start with the second question first. So we intend to assess the safety of fenebrutinib when we have the -- all of the studies read out and we look at -- when we look at the pooled safety. So obviously, we only have 2 of the 3 studies. So we need to wait a little bit in order to be able to step back and look at that. The data that we've seen so far, we haven't seen anything that is different than what you would see in the background rate of the overall MS population. In terms of liver monitoring, as is typical, when you get your label, usually for things like monitoring, you get what you studied in your label. So we would anticipate that we would have the same liver monitoring in our label that we had in our clinical trial. And again, I think when you look at the efficacy and the risk-benefit profile that fenebrutinib has, I think this still -- this is going to be a meaningful medicine in MS. So more to come as we get FENhance 1 data.

For persevERA, just to be really clear, the timing on that has not changed. We have consistently been messaging, the first half, mid-first half of this year and that has not changed. So that is remaining consistent. And again, we do plan to file the lidERA data first. We get the [ abema ] data, I believe, the substudy data comes, is that also in Q1? Guys, someone is going to have to remind me of that. And then the [ ribo ] study, which is a 200-patient substudy is just kicking off, so that will come later. So those are data pieces that clearly, as soon as they are available, we will be making public. But the adjuvant filing is going in -- into Q1 as planned and it looks like end of 2026 for the substudies.

James, all questions answered?

That's great.

Yes. Then next one is Sarita Kapila from Morgan Stanley.

So you addressed the study approval risk and I guess others have touched on it. But what is underscoring the confidence in the commercial potential? What's the initial feedback from the physician community being? So we've seen orals with LiverTox launch post CD20 approval, which have struggled to reach 1 billion. So I guess why is fenebrutinib different? And how are you viewing risk from Novartis' remibrutinib in RMS data in Q2 and they've had no signs of LiverTox so far? And then the second one is just on persevERA. It's also been touched on but how confident are you that you have enough patients in the trial to hit stat sig? And how should we think about the [indiscernible] study and the potential read across to persevERA?

Great. So let's start with fenebrutinib. So obviously, the data have not yet been presented. So the PPMS data goes to ACTRIMS shortly, and then the RMS data will be packaged together when we have FENhance 1 as well. That having been said, we've obviously shared it with those physicians who are part of the trial. And I think people have been really impressed with the data that we've seen. And in particular, people were really impressed with the Phase II data that we've seen. So what we're talking about is the ability to get OCREVUS like efficacy in an oral treatment. And for many patients for many, many different reasons, that's a very attractive option. So again, I think in this market, a lot of it comes down to the overall efficacy that we're able to deliver. And based on the Phase II data, we believe we have a highly efficacious molecule on our hands.

In terms of the Novartis data, I mean, it's important to remember, we haven't really seen anything in MS from Novartis yet. This is a dose that I think is about 4x higher than the existing dose. They had sort of second mover advantage and they started their trial with liver monitoring. So I think it's very difficult to compare because we just really haven't seen anything. We have first-mover advantage here. We've had robust Phase II data. And yes, I mean, I think we're -- it's very difficult to say anything until we actually see data. It's also, I think, important to remember that fenebrutinib is a non-covalent molecule. And in a chronic indication, that noncovalency really matters because it means that even though you're taking it chronically, if you need to stop for whatever reason, it does leave your system more quickly. And I think that really in a chronic care environment is a benefit.

So in terms of read-through for persevERA, it's clear when breast cancer is dependent on the endocrine receptor for viability, giredestrant can perform very well. And we've seen that in a number of settings. So all of these patients are, by definition, dependent on ER signaling and it's worked really well here. So clearly, in the front line, the likelihood that it's successful hasn't gone down. And we should always be really cautious with cross-trial comparisons. And so I think we are -- again, as I mentioned, the benefit is, we don't really have long to wait. So we'll know really soon. And yes, persevERA is designed to show improvement over palbo plus letrozole.

Sarita, all questions answered?

Yes.

Yes. And the next one in the queue is Richard Vosser from JPMorgan.

Two questions, please. First question, just to go to diagnostics for a little bit. Margins obviously hit by ramp-up of mass spec sequencing and the machine placements. Could you give us a bit of color on how to think about the margins from here? Those placements seem likely to continue as you ramp those 2 businesses up. So how should we think about '26 and then the improvement in the margins from there? And then second question back to pharma. Just going back to Vabysmo -- thanks for the comments on the foundations. Could we go a little bit further out and think about the future competition potentially from less frequently dosed injectable products? I think ocular has one half yearly. How you think about that sort of competition? And also closer to today, the biosimilars are really starting to come. They're having some impact in Europe as far as we can see. So just what's the thoughts globally, U.S., Europe on biosimilars from Eylea on Vabysmo?

So in this case, maybe...

No. Go ahead. I can use a break.

Thank you. So maybe starting with diagnostics. So we talked about a couple of effects. There's the new technologies. There's the tariffs of which Alan said we had half the year and we'll have a full year this year. But the biggest effect on what hit us last year on the margin was really the China effect. And as you heard from Thomas, we expect to see this meaningfully diminish this year. 2027, again, we expect to decline but it will be small enough that it won't really be meaningful. And then we expect to see a recovery. In terms of specific ambition on margin this year, I think I would refer you back to the group position that Alan mentioned earlier. But I would say our consistent ambition is to grow profit faster than sales.

And that is once we really get ourselves to the headwinds this year, that is our ambition going forward. And it's also our continuous ambition to improve the margin in diagnostics. That's something that is a goal for the entire organization. What I would call out, though, in 2025 is you had our second largest market with a 25% reduction. So obviously, there was an impact but that's something that you can see with our discipline on the cost line that you can also expect to see continue again in 2026 but we expect the gradual washout of that. Anything you would add to that, Alan?

Well, I think for '26, I think, well, we expect kind of a stabilization. I think that's a little bit here. But we will give that additional information.

Absolutely. Yes. So I would maybe just refer to what Alan said. Our goal really this year is going to be that we stabilize the margin.

And I think on a group level, you have seen that our intention is to expand margin in 2026. And what I've said in the past still holds, which is that also going forward, we will at least keep margins stable also for the coming years.

Perfect. Does that answer your question?

I think so.

You go ahead.

Yes. Great. So I'd like first just to start by talking about Vabysmo. So I mean, Vabysmo is highly efficacious therapy with a very well-defined safety profile where patients and physicians do have a lot of good experience in extending doses. And so -- and it is designed to do just that. When you look at -- you asked specifically about ocular, I mean, this drug is going into Phase III with a very small safety database and really no known data on long-term safety. And when you talk about something that's going to be used intraocularly over a long period of time, I think long-term safety is incredibly important. So I think it's very difficult to think about how these -- how something like that is a threat to something that has such good efficacy, such good safety and where you do actually have the ability to extend doses.

So I think from a future competition perspective, just like in other disease areas, sort of the bar is high here to unseat Vabysmo. And you had one other question. Oh, biosimilars. So far, what we see is that the Eylea biosimilars are taking from Eylea. And so for those patients who are really benefiting from a new and novel treatment, Vabysmo, they were just less impacted. So yes, I mean I think we saw Lucentis take from Lucentis. We're seeing Eylea biosimilars take from Eylea and we're seeing high-dose Eylea take from low-dose Eylea. So there's a lot of trading within that space. But I think for new patients who are going on therapy, physicians are picking the best available therapy out there available to them and that is Vabysmo.

And on ophthalmology, I would like to add that we have an amazing pipeline in ophthalmology. So when you look at all the different validated targets, I think we're the only company that actually has all the different validated targets in-house. And so if you look at our pipeline, we have trispecifics, tetrapecifics, [indiscernible] et cetera. So if I look at our ophthalmology pipeline, I think the one that's going to succeed surpassing Vabysmo, is then hopefully us.

Very good. Richard?

Yes. Perfect, everyone.

And next one in the queue is James Quigley from Goldman Sachs.

Hopefully, you can hear me. Just a couple of quick questions from my side left over. So firstly, again, on giredestrant and revisiting lidERA. What's the KOL reaction been from your side? So some of the KOLs we spoke to have been a little bit more cautious than the presenter at your SABCS event, given the more limited follow-up versus the CDK4/6 class. So how do you think this could impact the giredestrant trajectory, of course, assuming approval? Would it be more of a step up -- stepwise ramp or more a stepwise ramp as more data comes? Or do you -- or do your feedback suggest the potential for a faster, more optimistic ramp on giredestrant? So that's the first one. Second one, more a financial question. So underlying pharma growth has been pretty strong in recent years, driving operating leverage. So how is Roche balancing R&D investment with profitability? So how long can R&D expenses stay flat? This half seem to show that Roche has a strong ability to reallocate costs in R&D. But how long can this go on to support operating leverage?

I mean I think what we're hearing from the KOL community regarding where they would use lidERA is pretty bullish. I mean I think what we hear from -- the lidERA population is 55% of the adjuvant breast cancer population. That's 10% more than what we saw on the NATALEE trials. So I think you are really seeing physicians believe that this could have very broad applicability in their practice. And so I think that would -- that's what leads us to be fairly bullish about the opportunity. There's a 74% overlap with the population in NATALEE and monarchE. And so I think we're very -- yes, I think we're very confident that we have something on our hands here that is quite a game changer based on the data that we've seen.

Yes. Let me answer the second question. But first, something to add on giredestrant. I mean you look at the hazard ratio of 0.7. You can see that, that's, I would say, highly competitive to also some of the CDK4/6 trials that you've seen. But what you have on top of that is the tolerability. If you look at CDK4/6, you have quite a high amount of patients that actually stop using it simply because they cannot take the tolerability. So I think these are all the right arguments for SERDs to be used. So I do believe that the pickup will be strong. Unfortunately, I wasn't at the Congress in San Antonio but I heard there was standing innovation from the clinicians there.

Then the second question on R&D. So we're working very hard to be a high-performing, very cost-efficient organization. There are still opportunities, in my view, to continue to work on that. AI, by the way, plays a big role in that. We're using AI throughout the entire development process where we want to, on the one hand, speed up using AI but also reduce costs doing that. Regarding R&D expenses, also for 2026, I would say it will be broadly flat. Again, really focusing very hard on making sure that we put the money to work in the best possible way for the sake of patients in our company and for our investors.

And it all goes back to the margin point that you've made before.

That I made before, which is, it's clear for '26, expand margin. And as previously always said, at least stable for the long term.

James?

Thank you very much.

Okay. Then we move on to Graham Parry from Citi.

So one on lidERA. Thanks for clarifying the filing time line as being Q1. Just wondering if you could comment on whether you expect priority review or to use a priority review voucher or not. And when exactly do you think you would expect to see the [ ribo ] combo substudy data, 200 patients, how long does that take to recruit? And could we see something by the end of this year? Is that a next year event? And then on fenebrutinib, could you just confirm how important you think confirmed disability progression is versus annualized relapse rate reduction in showing a risk-benefit profile and differentiation versus Ocrevus to the regulator, just given the -- it's a very brain-penetrant molecule and has potentially, therefore, the ability to work on disability progression where CD20 doesn't? And then the final question is, you're technically still on clinical hold with FDA. So does that have to be lifted before you can actually file or receive approval? And what are the steps to doing that?

Okay. So we expect the [ abema ] substudy by the end of the year. We would expect [ ribo ] in 2027, that is really only starting now. So we've got a little time. That's 100 patients in the [ abema ] arm. And in [ ribo ] it's 200 patients. The FDA hold will be addressed as part of the planned [indiscernible] filings, but we've obviously been in consistent conversation with the agency over time. So there -- we've been in very close contact. I won't comment on our filing strategy only to say that we plan to bring lidERA to patients as quickly as possible. In terms of confirmed disability progression, I think we are -- the annualized relapse rate is also a very good endpoint here. And we are confident that, that is giving us what we need in order to proceed.

Graham, any additional questions or -- all good?

Then we move on. And I hand over to Rajesh Kumar from HSBC.

Two questions, if I may. First, on CT-388, thanks for clarifying discontinuation rate in the highest dose was similar to the overall group. You also mentioned that you could consider a flexible dosing in Phase III trials as an option. So could you give us some color on how you're thinking about Phase III progression? Would flexible dosing or an active comparator be something you might consider? Or is it at the moment, too early to comment on that? Second, just on giredestrant, quick follow-up. You highlighted the overall TAM this class is targeting to be quite a large number. You are filing with a few -- some persevERA data is about to read.

So just in terms of the market segmentation, how much of the market you think it have been risked -- derisked to some extent and how much we still depend on the data in your assessment of the market would be very much appreciated. And because I'm an analyst and I cannot count, the third question would be just on the clarification on Vabysmo. Appreciate the working capital impact has gone up and that sort of reflects a very strong December. So should we consider the exit rate of December close, the indication of how you're thinking about growth in 2026? Or should we take an overall slower growth rate going forward on Vabysmo?

So I would just go back to my earlier comments. For Vabysmo, we expect to see an acceleration of growth in 2026. So more to come on that. In terms of the dosing for CT-388, so what we have disclosed is that CT-388, it will be administered once a week and we're aiming to develop it at 3 maintenance doses. We are not disclosing at this time the details of that dosing strategy. But just to avoid any misunderstanding, we have not indicated that we will be doing flexible dosing within the trial. So -- but right now, the details of that Phase III design, that specifically have not been disclosed. And then in terms of giredestrant and the market segmentation, so -- and how much do we feel like has been derisked? Well, 2/3 of the market is adjuvant and we have a positive adjuvant trial. So I mean, I think a significant portion of the -- we have a significant portion of the market that has been derisked.

Okay. And then next questions are from Michael Leuchten from Jefferies.

A question for Matt, please. Abbott said last week that the Chinese may be pursuing VBP for Core Lab oncology. Just wondering whether you've heard that and how that may or may not have been reflected in your outlook and the margin commentary you made earlier? And then sorry, Teresa, just going back to Vabysmo, just your comment about 2025 in the U.S. being reset. Q4 was still soft. It didn't really improve upon Q3 sequentially. So when you say you think that's now stabilized and it can grow from here, just wondering how you look at that Q4 versus Q3 dynamic in the U.S.

Okay. 3. Wow. I know what 3 is [indiscernible]. So what I would first start off by saying is, as you may know, there was VBP for Core Lab oncology reagents last year. And so that was what you see, our China effect last year significantly represented a decrease in our Core Lab oncology reagents, which were down about 50%. So some of that effect is still pulling through this year. But I can't speak for what was said on that call but we are seeing the effect of the VBP last year and the national reimbursement reduction. So we don't anticipate additional Core Lab oncology VBP this year.

So in thinking about Vabysmo, Q4 -- so 2025, we saw a big reset in the branded market in the U.S., right? With the closure of the co-pay foundations, fewer patients were put on branded drugs, more patients were put on Avastin and biosimilars. And you saw a big just sort of reset in how many new patients and continuing patients were actually going on a branded therapy and that constricted the market by about 15%. That constriction went all the way through Q4 because normally, when donations are given or grants are given, they're given 4 years' worth of therapy. What's happening right now in the oncology world is something called the blizzard. It's where every retinal specialist in the U.S. goes and reverifies the benefits for every single one of their patients. And it's at that point in time that patients actually determine what -- will they be continuing on their current medication, will they be switching, et cetera.

And so over the course of the next couple of months, I think we're going to get a real sense of what is the trajectory of the branded market going to look like in 2026. But because that underlying base effect of 2024 is now washed out, you should be able to see the actual branded growth of people going on to new therapies actually come through. So I think there's a reason why we didn't see Q4 look any different than what the rest of the year looked like. I think we had sort of hoped that we might see some early signs of recovery but I think those signs of recovery really are going to come -- become a little bit more evident as we get towards the end of Q1. So Michael, I hope that addresses your question.

Yes. And I mean, we -- the co- assistance foundations, they are separate, right, so nothing that -- in terms of influence. But what we can say is in general, that our donation was towards the end of Q4.

Yes. And then again, we don't link those 2 things. It is interesting to know though that in Q4, we did see a 4% growth. So we saw...

Quarter-over-quarter.

Yes, quarter-over-quarter growth. We did see a 4% growth. So we did see a little bit of an uptick.

Michael, any follow-on? If not, then I would hand over to Paul Kuhn from Cowen. Paul?

Thanks, Bruno. This is Paul on for Steve Scala. Two questions, please. What feedback have you heard from U.S. oncologists and how they plan to initially use giredestrant in the adjuvant setting? And secondly, how did the change in Xolair biosimilar entry from end of 2026 to before November 2026 come about? Was this a change in the settlement with generic manufacturers?

So with regards to Xolair, I think we have long said sort of second half of 2026 is when we expected the first biosimilars to come in for the U.S. So I don't actually think that that's a change.

I think mentioning November was just related to IRA. So that is really, we need to have a biosimilar place -- a payer in the market before the 1st of November. So then we cannot get negotiated.

That is correct. So my reference to the 1st of November was purely around CMS guidance that says if you have a marketed biosimilar by November 1, 2026, then you will be removed from the negotiated basket. So that's where that date comes from. But we've always said second half of 2026, we would expect to have a biosimilar in the market. And then feedback from oncologists on where they intend to use for adjuvant. I mean, again, just to continue to reiterate, 55% of the adjuvant population was covered by the lidERA trial. And so I think you see a high degree of confidence in an oncologist to use in a very significant portion of their patients. And again, what we saw here was a very efficacious, seemingly combinable and well-tolerated therapy that I think has the opportunity to really become a new standard of care in this setting. So what we're hearing from oncologists in general is that they're pretty excited to have this in their hands and we're excited to get it to them.

Paul, if we answer all your questions?

Thank you.

Then next one would be Justin Smith from Bernstein.

Two, please. Pharma, #1, NXT007, just wondered if you could share some thoughts on when the Phase III design head-to-head versus Hemlibra will hit ct.gov. Second one, diagnostics, Matt, just wondered if you could talk a little bit about CGM and when the finger prick recalibration will be removed and the impact that might have?

Wow, 4 is new territory. So I want to first thank Teresa for generosity. But -- so starting with your question on auto calibration, which is the comparison of the CGM device with a blood glucose lancet, what we are planning to do is have that launch happen this year. I won't say exactly which quarter but that is an improvement that we expect to deliver this year.

And with regards to NXT007, we would expect those trials to start in Q1, Q2 with clinicaltrials.gov entries at around that time frame. And again, these are 2 studies, one, head-to-head and one, versus [indiscernible] and one Hemlibra.

Justin, all questions are answered?

Yes. Great.

Then I would maybe read here loud 3 questions, which I got from Luisa Hector. She had to drop off and I promised her I will go through them. There is one question on Ocrevus [indiscernible]. So the split of naive versus switch patients that we are capturing and what is the target switch rate for 2026 and at peak?

What I think we have been communicating that we have 2 billion in incremental sales for Ocrevus but this is true incremental sales. And on top, basically, we would have revenues coming from switching. So we have not yet provided a detailed outlook on what the ratio, IV to subcutaneous would be at around 29%. We might do that at a later point in time. There's then a second question I found interesting on the pipeline. 66 NMEs now on the pipeline. Is this rightsized? And what we have seen now with the turnover in the fourth quarter with 4 molecules added, 5 going out -- so 5 added, 4 going out, is this now -- is there still cleansing ongoing of the pipeline? Is this now the regular run rate and the turnover? Or would we target more NMEs overall?

Yes. I mean I can answer that question. So overall, you apply the bar not only once, you apply the bar constantly based on data that you generate but also data that you get from the outside. So clearly, I think we are at a point where we'll continue to bring in additional NMEs. We have actually -- when you look at the very early stage of our research organization, we've actually doubled the amount of molecules moving ahead there. So we do believe that we'll continue to expand on the amount of NMEs that we have in our portfolio beyond the 66. But this kind of, I would say, prioritization is just something that you have to do constantly based on just availability of data.

And then the final question here would be on capital allocation. Given your positive pipeline progress, pharma deals with the U.S. administration and with competitor developments in obesity, are there any changes to your M&A objectives and R&D investment plans?

No, I can say there is no fundamental change. I think what we have really shown over the last couple of years that we've been very disciplined, very disciplined in terms of financials but also in terms of really screening the market for interesting molecules with good data. If you look at the amount of money that we spent compared to other companies and the kind of pipeline we've built doing that, I think we've been pretty efficient. And our intention is to continue to do the same and just continue to be quite disciplined on that. The good thing is, we are not in a situation where we have a huge patent lift, right?

So we are not in a situation where we have to do late-stage deals, which are very costly. I think we are in a very good position when it comes to our late-stage pipeline. But obviously, I mean, if you look at the amount of innovation that's ongoing outside, you look at what's happening in China, we need to continue to screen the market and look at everything that's out there. I mean I looked at a statistic, for about 1,000 companies that we look at we do 1 deal. And I think that's also what I expect of our organization that we know exactly what's going on outside so that we can make data-driven good decisions.

Very good. I think with that, actually, we are at the end of our Q&A session. Let me just remind you of the 2 upcoming IR events we already have flagged. I assume there might even be more. There's on February 9, our neurology call, we will cover up the PPMS data for fenebrutinib presented at ACTRIMS. And then on May 12, we again will have our Diagnostic Day as a live event in London, where we will take you through the entire portfolio and highlight this year, I think, will be SBX sequencing. As we are now in the global launch phase, I think there is the next steps to come with pricing and so on. So I think it will be an exciting event. And with that, I hand over to Thomas for the final remarks.

Thank you very much, Bruno and huge thanks also to the team. I would say, quite exciting times. I mean, if I see all the discussions that we've had as a team over the last 3 years and the progress we made, I think it's significant. And it's not only that, it's also a lot of fun because we get to talk about what we like to talk about, which is science, which is about progress for patients. And with people like Alan and myself who like math, also we can talk a lot about financials. So I think there's a lot of good things that are going on. And we have a good momentum both on financials and pipeline. So very proud of the team. And I do believe we've always done what we said that we are going to do and you will continue to see that going forward. We continue to move with focus. We continue to move with speed. And as always, you can count on us because we will deliver.

Good morning, welcome to Roche Virtual Oncology SABCS Investor Event. My name is Henrik, and I'm the technical operator for today's call. Kindly note that the webinar is being recorded. [Operator Instructions] remreeche.com/investors to download the presentation. At this time, it's my pleasure to introduce you to Bruno Ashley, Head of Investor Relations. Bruno, the stage is yours.

Thanks, Henrik. And could I have the first slide, please?

So welcome to our eighth and last IR event for 2025, focusing on the Phase III LIDERA results for giredestrant in early ER-positive breast cancer, which just got presented yesterday at the San Antonio Breast Cancer Symposium.

As an upfront remark, let me just point out also the event originally had been scheduled for 60 minutes. We have now planned for 90 minutes just to make sure that we have enough time that we can take all your questions, which I'm sure there will be plenty of them.

So let me quickly go through today's agenda. I will start today by providing a quick update on the big picture. I thought this might be needed considering the dense and impactful news flow we experienced in the fourth quarter. And following myself, the second speaker will be Maura Wila Vice President and our Global Head for Breast and Gynecologic Cancer Product Development.

Maura will provide us an update on our expanding breast cancer portfolio, highlighting future development opportunities as we have several assets already in the clinic, which could become important combination partners in future development programs. Our second speaker today will be Dr. Aditya Badia. Aditya is an MDM Ph. He is a fellow of the American Society of Clinical Oncology and the Professor of Medicine at the Geffen School of Medicine at UCLA. He was a principal investigator in the LIDERA study and was a member of the Trial Steering Committee. Dr. Badia will take us again through the Phase III LIDERA results for gulesirant in early ER-positive HER2-negative breast cancer in the adjuvant setting, which he also presented yesterday at the San Antonio Breast Cancer Symposium.

Afterwards, we will start our Q&A session. In addition to our 2 speakers, we will be joined for the Q&A by Levi Geraway, our Executive Vice President and Global Head of Product Development and Chief Medical Officer; by Ema Ella Groudi, our life cycle leader for hormone receptor-positive breast cancer; Pablo Perez Moreno, our Global Development Leader for gireesrant; and by Stefan Frings, our Deputy Chief Medical Officer.

Could I have the next slide, please? Before we get started, I just wanted to quickly summarize the Q4 news flow and provide an update on the pipeline, some peak sales expectations and the overall outlook. As summarized on this slide, we had 3 major positive readouts in Q4, which when taken together, have a significant impact on Roche's long-term investment case.

So on the left side, you see the Phase III results for Gazyva in kidney disease. As you know, we just announced the U.S. and EU approvals for Gazyva and lupus nephritis. And in addition, in Q4, we communicated positive Phase III results in SLE, which comprises a significantly larger patient population. The SLE data are now planned to be presented at an upcoming conference. So it's probably at Euro in March. And we also announced positive Phase III data in INS with data to be presented at the World Congress of Nephrology also in March of 2026.

As you can see here, there is one additional Phase III study in membranous nephropathy to come in 2026. And let me point out here, taking all these studies together, the overall peak sales opportunity for Gazyva in kidney disease has now really increased to up to CHF 2 billion. Second, in the middle, you see the Phase III program for fenebrutinib in PPNS and RMS.

In Q4, we announced a positive Phase III outcome for fenebrutinib in PPMS, which is the only study benchmarking a BTK versus a CD20 antibody. And in addition, we announced the first positive Phase III FenNHanceE-2 in RMS. The second parallel study, FNhanceE-1 is to read out in the first half of '26. Overall, we believe fenebrutinib has now potential to become a best-in-disease medicine, and we have raised our peak sales expectations from EUR 2 billion to EUR 3 billion to more than EUR 3 billion. I can also confirm already that the PPMS data will be presented at ACTRIMS in February, whereas the RMS data will be presented at another conference once we have the Fenhance data -- FENHanceE-1 data available. And just to confirm, RMS and PPMS data will be filed together in 2026.

Finally, on the right side, you see the Phase III givudesrant development program, where we have reported now 2 positive Phase III studies, positive Phase III EVERA results in the post-CDK4/6 setting, second-line plus ER-positive breast cancer, which were announced end of September and got presented already in October at ESMO. And in Q4, we now announced and presented positive Phase III LIDERA results in the adjuvant setting, which we will discuss today in more detail.

One additional comment here. U.S. filing for givudesrant based on the evERA data is now imminent and we expect the initial launch for givudesrant in the second-line setting to happen in 2026. Could I have the next slide, please? Let me also quickly provide an update on this slide, showing the 19 NMEs, which could launch by 2030. We have seen this slide before. The last time we showed it was at the Q3 results. What I want to highlight here is now we have 3 eventually even 4 NMEs, which are approaching near-term filing. These NMEs are Vamikibart in UME boxed here in red, giredestrant in ER-positive breast cancer, which I mentioned already, the U.S. filing is imminent, and then potentially fenebrutinib in RMS and PPMS if the second RMS study reads out positive in first half '26.

And finally, not shown on this slide, we are also looking into potentially filing setralizumab in TED, but the filing decision has not yet been taken. Also to call out here on this slide, you can see that we have increased peak sales estimates for fenebrutinib from previously EUR 2 billion to EUR 3 billion to now more than EUR 3 billion.

Next slide, please. This is another slide I quickly wanted to comment on. We had shown this slide last time at Pharma Day, providing a bit of a mid- to long-term outlook for the company. Two comments here from my side. This slide will need now to be updated, and we will do that in the first half of '26. We had previously communicated that our on-market portfolio, this is shown in light blue, is expected to deliver growth until '28 and thereafter to be flat, always fully compensating for expected EUR 1 billion to EUR 1.5 billion of annual generic erosion.

As you can imagine now with the Q4 trial successes, this picture has now changed fundamentally with a clearly improved visibility on the long-term growth outlook. With somewhere between 2 to 4 NMEs now approaching filing and overall raised peak sales expectations, including the higher expectations for Gazyva and kidney disease, we expect now to deliver pharma growth until 2030 and beyond with the higher value PostBar portfolio only starting to read out from 2027 onwards.

In 2027, we expect already the first major readouts from this portfolio with afinkibart in IBD, pegaoferin in MASH and potentially trontilimab at year-end 2027. Can I have the next slide, please? Another Pharma Day slide, which you know well, where I would expect now some material changes to occur. As we had mentioned previously, consensus carried very little peak sales expectations for fanubrutinib and giredestrant, largely due to disappointing readouts from in-class competitor molecules. For givvedestrant, we had pointed out that the EUR 900 million peak sales would more or less cover the peak sales expectations for evERA alone in second line plus. So that any additional success would be pure upside. And that's, of course -- this, of course, explains now, to some extent, the recent share price move, which we have seen. For fanubrutinib in MS, there were only EUR 700 million peak sales in and also the opportunity for Gazyva in kidney disease seems not yet fully captured by the consensus.

Next slide, please. This slide is just here to finish the 2025 news flow. There is one more positive Phase III trial, which we announced today, which also came in, in the fourth quarter. That's the -- that's PSI in ACS, the Phase III COMUTE-a trial, which represents a smaller opportunity of up to EUR 0.5 billion, which is -- which has been positive. And if we go to the final slide, I just wanted to provide here an updated 2026 news flow slide. You see now a lot of U.S. EU filings scheduled for next year. And in terms of pivotal study readouts, we will have one additional giredesrant readout in first line. That's [ PASivira ]. The second first-line study in -- for giredestrant ionira is to come a year later in 2027. And then also, you see here the one remaining study for Gazyva in membranous nephropathy. We also have additional line extensions reading out next year for ITOfi, our PI3 kinase inhibitor, which is developed in ER-positive and HER2-positive breast cancer as well as Vsumio in follicular lymphoma. and we have the first pivotal readouts for 2 new molecular entities for divarasib, our second-generation KRAS G12C inhibitor to be developed in non-small cell lung cancer and for Xifaxasan, our Factor B antisense oligonucleotide developed in IgAN.

And with respect to the earlier pipeline, 2026 will be the year of our obesity CVRM portfolio with all NMEs having Phase II readouts.

And with that, I will close my upfront remarks and hand over to Maura to run us through our emerging breast cancer franchise. Maura, please.

Yes. Thank you, Bruno. And next slide. And today, I'm excited to share our strategy for our breast cancer franchise at Roche. Breast cancer is one of the 4 end-to-end oncology disease areas where we will focus to accelerate innovation from discovery through launch and to expand our leadership to ER-positive breast cancer.

We have rebuilt our breast cancer pipeline with clear focus on key priority pathways and are accelerating our time lines to best position the breast franchise for success. Next slide. We would like to share with you our current pipeline in solid tumor oncology. In breast cancer, we have one of the strongest portfolios. Giredestrant has become the focal point of this program as it has the potential to serve as a new endocrine therapy backbone in ER-positive breast cancer. We are also well positioned to develop combinations with our molecules to inhibit aberrant signaling pathways that drive breast cancer progression.

Next slide. ER-positive HER2-negative breast cancer represents a significant market opportunity, accounting for about 70% of breast cancer. In addition, ER-positive HER2-positive disease constitutes another 9%, which represents more than 50% of the HER2 population. The majority of these patients are diagnosed with early-stage disease, and therefore, adjuvant endocrine therapy constitutes the largest population of patients. There remains an unmet need for more effective and tolerable therapies with improved adherence to increase the chances of care.

Next slide. Presently, we are building upon Roche's pioneering work developing transformative therapies in HER2 and also focusing on other key signaling pathways that underlie breast tumor dependencies. We have a portfolio of potentially best-in-class differentiated molecules that are targeting the estrogen receptor with giredestrant, the PI3K pathway with ITOVi, the CDK pathway with GDC-4198 and the HER2 pathway with ZN-1041. Our goal is to redefine breast cancer.

Next slide. Despite advances in treatment, a high unmet need remains in ER-positive breast cancer. Up to 1/3 of patients on adjuvant endocrine therapy eventually develop a recurrence. Existing endocrine therapies have incomplete suppression of the ER pathway and contribute to adverse events. While available and efficacious, CDK4/6 inhibitors cause significant side effects, including neutropenia, diarrhea and elevated liver enzymes, leading to discontinuations and are limited to certain patient population.

In the first-line metastatic setting, more than half of patients experienced progression within 2 years on endocrine therapy and CDK4/6 inhibitors and many patients become resistant to subsequent endocrine therapies. Next slide. Giredestrant is a next-generation oral SERD and full ER antagonist designed to drive deep inhibition of ER signaling. Preclinical data demonstrates greater potency of giredestrant compared to other SERDs, and we are presenting supportive preclinical data at this conference. Giredestrant demonstrated superior antiproliferative activity with significant reduction in p67 and ER expression in both the QUABERA and the EMPRES trials. K67 reduction predicted outcomes of adjuvant endocrine therapy trials. Importantly, giredestrant was well tolerated with no dose-limiting toxicities.

Next slide. In the evERA trial, giredestrant plus everolimus demonstrated a statistically significant and clinically meaningful improvement in progression-free survival in the ESR1 mutant and ITT populations after CDK4/6 inhibitors compared to the standard of care endocrine therapy plus everolimus. There was an overall survival trend favoring the giredestrant arm across these subgroups. The giredestrant plus everolimus doublet had a manageable safety profile, including no Potoxia. We are presently sharing these results with health authorities.

Next slide. We have initiated a broad clinical development plan for giredestrant across patient populations. Highlights of the CDP include the Phase III PERSEVERA trial in a frontline endocrine therapy sensitive population anticipated in the first half of 2026, the Phase III PIONEER trial in a frontline endocrine therapy resistant population with physician choice of CDK4/6 inhibitor and the Phase III HERADERA trial in a hormone receptor positive HER2-positive frontline maintenance population in combination with Phesgo. In early breast cancer, giredestrant plus CDK4/6 combinations are being evaluated with an Abeema safety substudy fully enrolled and a RBO substudy planned and we will be sharing the LADERA data with you soon.

Next slide. Shifting to ITOVi. ITOVi is an alpha-selective PI3K inhibitor that can be administered safely in combination, allowing additional synergistic inhibition of the estrogen receptor and CDK4/6 pathways at standard doses. ITOVi and iNAVA120 has defined a new standard of care in frontline PIK3CA mutant endocrine-resistant breast cancer, demonstrating a statistically significant improvement in progression-free and overall survival. There was a median 2-year delay to subsequent chemotherapy, which is a meaningful outcome for patients. Discontinuation rates were low due to adverse events, confirming manageable tolerability. ITOVi is now approved in the United States, EU and China, and additional launches are ongoing.

Next slide. We have initiated a broad clinical development plan for ITOVi. Highlights include the Phase III iNAVO-123 study initiated in frontline endocrine-sensitive patients, the Phase III iNAVO-121 study, which is a head-to-head versus alpelisib and the iNAVO-122 studies in HER2 maintenance setting, the Phase II neoadjuvant study of ITOVI, ribociclib and letrozole, the NEOTOVE study and an additional adjuvant study is being considered. And we have also reported promising clinical activity of ITOVi in combination with giredestrant in an arm of the MORPHEA study and a study of ITOVi and giredestrant and GDC-4198 is planned.

Next slide. GDC-4198 is a potent inhibitor of CDK4 with substantial activity against CDK2 designed to address key resistance mechanisms that develop when targeting this pathway. Preclinical data demonstrates similar in vitro potency to parenteral and resistant cell line. GDC-4198 delivers similar activity as one molecule compared to [indiscernible] and tagociclib in combination.

Next slide. Early clinical activity of GDC-4198 is promising as monotherapy in the post-CDK4/6 inhibitor setting with durable responses seen in heavily pretreated patients with manageable tolerability. Given this efficacy and tolerability, we have initiated a Phase Ib/II study of GDC-4198 and giredestrant in the post-CDK4/6 setting.

So in summary, we have built a portfolio of exceptionally -- exceptional potentially best-in-class molecules. We have launched DYTtoVi, delivered positive data on giredestrant in both the metastatic and early breast cancer setting and continue to develop best-in-class combinations with the addition of 4198.

It's really now my honor to introduce Dr. Aditi Bardia, maybe next slide. who's Professor of Medicine from UCLA, who will now present the giredestrant LidDERA data. So Aditya?

Thank you so much, Maura. It's my pleasure to present the results of the Phase III clinical trial, giredestrant versus standard of care endocrine therapy as adjuvant treatment for patients with ER-positive HER2-negative early breast cancer.

Next slide. So as mentioned previously, this was presented at SABCS yesterday. These are my disclosures. In terms of background, ER-positive breast cancer is the most common subtype of breast cancer. More than 70% of early breast cancers are ER-positive HER2-negative. And for ER-positive disease, endocrine therapy remains the mainstay of management, including early breast cancer. If you look at approvals of new endocrine therapy in early breast cancer, the last approvals came with the introduction of aromatase inhibitors in early 2000s from the ATAC and BIIB098 study, which showed that anastrozole, letrozole were superior to tamoxifen. The hazard ratio in those clinical trials was about 0.8.

Next slide. More recently, we've seen CDK4/6 inhibitors in combination with endocrine therapy that have shown increased efficacy, but have also introduced associated toxicities. Moreover, many patients discontinued treatment early because of safety, tolerability, thereby increasing risk of recurrence. So clinically, there's an unmet need to have a therapy that's more effective and better tolerated in the adjuvant setting.

Next slide. Giredestrant or GDC-9545 is a next-generation oral SERD. The drug is designed to bind to ER, induce a confirmational change in ER that leads to full antagonism as well as degradation. One point to note is that these oral SERDs like giredestrant can block both ligand-dependent as well as ligand independent ER signaling. And this is an advantage over aromatase inhibitors, which only block ligand-dependent ER signaling.

Next slide. In multiple cellular lines as well as cellular viability assays, giredestrant has been shown to be more potent than other oral SERDs, including GDC-810, it's also been shown to be more potent than fulvestrant as well as tamoxifen. In early breast cancer, giredestrant demonstrated superior antiproliferative activity in the neoadjuvant trial in the CUPERA study versus anastrozole in the EMPRES study versus tamoxifen. As was mentioned by Dr. Dickler previously, we know that suppression of Ki-67 in the neoadjuvant setting is a strong surrogate marker for improvement in invasive disease-free survival.

So based on these results, the mechanistic data, the preclinical data as well as the neoadjuvant data, LIDERA was designed to test giredestrant versus standard of care endocrine therapy in the adjuvant setting, which is the biggest setting in terms of patient population in ER-positive breast cancer.

Next slide. LIDERA is a global randomized Phase III clinical trial that enrolled patients with ER-positive HER2-negative early breast cancer, patients with Stage 1 to Stage II disease, both premenopausal as well as postmenopausal. They were randomized to giredestrant versus standard of care endocrine therapy, which could be tamoxifen, anastrozole, letrozole exemestane. The study was fundamentally designed to find what is the best endocrine therapy in early breast cancer. In this trial, patients who are premenopausal also received ovarian function suppression, and that becomes relevant as we talk about AEs related to this trial. Patients received ovarian function suppression with giredestrant AIs. It was obstal with tamoxifen, but majority of patients who received tamoxifen also received ovarian function suppression. Giredestrant is also being combined with abemaciclib in the adjuvant setting. That's the LIDERA breast cancer substudy.

Next slide. The primary endpoint of the trial was invasive disease-free survival, and this was performed in the intention-to-treat population, was analyzed using the log rank test, and it was considered positive if the IDFS analysis was statistically significant with a 2-sided p-value of less than 0.02. The prespecified efficacy interim analysis occurred when 336 IDF events were observed. And at this time, interim overall survival analysis also occurred based on the hierarchical statistical design. Assigned protocol therapy and follow-up will continue until final overall survival analysis.

Next slide. In terms of patient flow, 4,170 patients were randomized 1:1 to giredestrant versus standard of care endocrine therapy. Majority of the patients, 84% of the patients received aromatase inhibitor in the standard of care endocrine therapy arm and 16% received tamoxifen. The median follow-up at the time of data cutoff was about 32 months. If we look at treatment discontinuation at the time of the data cutoff, 347 patients had discontinued giredestrant and 520 had discontinued standard of care endocrine therapy.

Next slide. In terms of baseline demographics, they were well balanced between the 2 arms in terms of race, region, menopausal status, about 40% of patients in this trial were premenopausal. In terms of disease characteristics, about 50% of patients in the trial had Stage II disease, about 40% Stage III disease and about 10% Stage 1 disease. 70% of patients in the trial had high-risk disease and about 30% had medium risk disease. Majority of the patients received prior chemotherapy. So overall, the baseline demographics as well as the baseline disease characteristics were well balanced.

Next slide. So these are the primary results related to invasive disease-free survival. Giredestrant was associated with a statistically significant and clinically meaningful improvement in invasive disease-free survival. It reduced the risk of invasive recurrence or death by 30%. This is a hazard ratio of 0.70, and this was statistically significant with a p-value of 0.0014. If we look at the Kaplan-Meier curves, we see an early separation in favor of giredestrant and the separation is maintained over time. If we look at 3-year IDFS results, it was 92.4% with giredestrant versus 89.6% with standard of care endocrine therapy arm. So about a 3% absolute improvement in invasive disease-free survival at this time.

Next slide. If we look at the results by type of endocrine therapy, again, giredestrant was superior as compared to aromatase inhibitor with a hazard ratio of 0.73 and tamoxifen with a hazard ratio of 0.53.

Next slide. And if we look at other subgroups, again, we see that giredestrant was pretty much superior to standard of care endocrine therapy in all the subgroups by age, region, menopausal status, risk prior chemotherapy.

Next slide. And if we look at the results by stage, for patients with Stage II disease, giredestrant was superior to standard of care endocrine therapy with a hazard ratio of 0.58 and for patients with Stage III disease as well with a hazard ratio of 0.74. For patients with Stage I disease, the confidence interval crossed 1, but the number of events were few in this subgroup, which is consistent with the natural history of Stage 1 disease.

Next slide. In terms of distant recurrence-free survival, again, giredestrant was superior with a hazard ratio of 0.69. So this corresponds to a 31% reduction in the risk of developing metastatic disease was reduced with giredestrant as compared to standard of care endocrine therapy. And again, you see the curves here separate early and the separation is maintained over time in favor of giredestrant.

Next slide. In terms of interim overall survival, the overall survival results are immature at this time, but a clear positive trend was observed in favor of giredestrant with a hazard ratio of 0.79. Overall survival testing will continue at future analysis.

Next slide. And finally, in terms of safety, Overall, the incidence of AEs and serious AEs were comparable between the treatment arms. If you look at the mean dose intensity, that was close to 99% in both the arms. If you look at AEs with fatal outcomes that was lower with giredestrant, there were 6 deaths with giredestrant in the giredestrant arm as compared to 16 in the standard of care endocrine therapy arm.

Similarly, if you look at AEs leading to treatment discontinuation that was lower in the giredestrant arm as compared to the standard of care endocrine therapy arm. So fewer patients in the giredestrant arm discontinued because of AEs as compared to the standard of care endocrine therapy arm. In terms of the percentage, it was 5.3% discontinuation in the giredestrant arm versus 8.2% in the standard of care endocrine therapy arm.

Next slide. And if we look at the type of AEs, this volcano plot summarizes the specific AEs. Arthralgias were the most common AEs seen in both the arms. As a reminder, in this trial, patients received ovarian suppression in addition to the endocrine therapy for patients who were premenopausal. The other AEs that were seen include menopausal symptoms. While the rate of musculoskeletal symptoms were similar between giredestrant and standard of care endocrine therapy, discontinuation because of AEs were lower in the giredestrant arm as compared to standard of care endocrine therapy arm.

So if you look at musculoskeletal disorders, the rate of discontinuation was 4.4% in the standard of care endocrine therapy arm. And in the giredestrant arm, it was lower, 1.8%. In terms of AEs that were higher with giredestrant, that included bradycardia, majority were grade 1. As a reminder, grade 1 bradycardia is asymptomatic. It does not need any treatment old or discontinuation and no medical intervention is needed for grade 1 or asymptomatic bradycardia. This is something that's picked up on routine monitoring. There were no grade 3 or grade 4 events in terms of bradycardia seen in both the arms. And then finally, venous thromboembolic events that was lower in the giredestrant arm, Grade 3, Grade 4 as compared to the standard of care endocrine therapy arm, which also included tamoxifen that can induce venous thromboembolic events.

Next slide. So in summary, since the approval of AIs in 2000s up to almost 25 years, LIDERA is the first study in the adjuvant setting to demonstrate benefit with a novel endocrine agent in early breast cancer. Next slide. With a median follow-up of 32 months, LIDERA demonstrated a statistically significant and clinically meaningful improvement with upfront giredestrant versus standard of care endocrine therapy in ER-positive, HER2-negative Stage I to III early breast cancer. The invasive disease-free survival hazard ratio was 0.70. The 3-year invasive disease-free rate was 92.4% versus 89.6%. So that's about 3% absolute improvement in IFS. Overall survival also trended in favor of giredestrant and distant recurrence-free interval was improved with giredestrant versus standard of care endocrine therapy with a 31% reduction in the risk of developing metastatic disease.

Next slide. And in terms of safety, that was favorable and consistent with the known safety profile of giredestrant, the discontinuation rate was numerically lower with giredestrant as compared to standard of care endocrine therapy, and that's an important observation in the adjuvant setting.

So overall, the results support giredestrant as a potential new standard for patients with hormone receptor positive HER2-negative early breast cancer. I'd like to thank and acknowledge the patients, families, the Global Steering Committee, TRIO NSABP Foundation as well as all the investigators who are involved in this trial. Thank you so much.

Excellent. Thank you. So with that, we will open the Q&A session. The first questions go to Emmanuel Papadakis from Deutsche Bank.

Hopefully me okay. Maybe I'll take 2, if I may. Congratulations on the data. Perhaps a question around subgroups and a question on potential labeling and guideline implications. So the question on subgroups is just we didn't get any disclosure of benefit by nodal subgroups, which is something we've seen in prior studies in a setting such as NATLEarE. So could you just confirm that the benefit in terms of hazard ratio was consistent in the node negatives, which was around 1/5 of the population in the study with those with a nodal count of 1 or 2 to 3? That would be very helpful. And then just in terms of labeling and guideline implications, perhaps you could give us a sense, it was a relatively high-risk population enrolled in the study to give us a sense of what you're anticipating the FDA label may look like? Are you anticipating a high-risk wording in Stage II and III patients or something akin to the Kisqali wording for adjuvant breast cancer or something close to Verzenio, which is indicated for node-positive, high-risk patients, for example? And then any thoughts you could share on NCN guideline implications as well would be helpful.

Who wants to take the first question maybe on the subgroups.

I can take that. Pablo Perez, I'm the Global Development Leader for Gx. So when we look at the subgroups, as we shared in the presentation, the subgroups are generally consistent, and this includes novel status. The stage was reflective of the status.

Exactly. Well, Stage 2 includes patients with N1 -- mainly N1 disease. And as you highlighted, Stage 2, although includes some patients with no negative disease, it's also reflective of the activity in patients.

Okay. Then who is taking a second question, which is a bit more looking already further ahead in terms of labeling, how this could look like?

Maybe I'll jump in on that. Obviously, it's difficult at this stage. Actually, we generally don't comment on labeling before there is a label, but I would just underscore the point that Pablo made, which is that we -- there's consistency of the data across all of the subgroups that we looked at. We're one of the few adjuvant studies to include Stage 1. And so -- but even there, directionally, things are consistent. And certainly, as was mentioned, across Stage II and III. So the consistency of the data, we think will speak prominently here, but we can't discuss specifics of label at this point.

Maybe on NCCN guidelines.

Yes, absolutely. In terms of NCCN guidelines, the study met its primary endpoint and IDFS was analyzed in the intention to treat population. So we would expect in terms of guidelines that the drug will be used in Stage I to Stage III because that was the primary endpoint. And so in terms of guidelines, this would be in the upfront setting, in the adjuvant setting as an option for patients with Stage I to III ER-positive breast cancer.

Emmanuel, this answer.

Yes, that's very helpful. And I'll circle back in the queue if this in.

So our next questions go to Colin White from UBS.

Colin White from UBS here. It's a follow-on, I guess, to Emmanuel's question there was we've been getting a lot of questions given that CDK4/6 inhibitors are used in a medium to high-risk population, just how broadly giredestrant will be used. I was wondering if Dr. Bhadia could comment how much it would be used in the lower-risk patients from the LIDERTA criteria. Assuming it has a label, how much you think it might be used in even lower-risk patients, that 55% of patients you identified as lower risk and also the potential for use after a CDK4/6 inhibitor. Yes, that's my first question.

And then my second question is a little bit about the difference between LADERA study and other adjuvant studies, which are -- which we're expecting, where patients in LADERA had only had less than 12 weeks of endocrine therapy, whereas for some other studies like CAMBRA-1 or EMBER-4 for other oral SERDs, they've had about 2 years of prior endocrine therapy.

So just comments on how we'll be able to maybe assess that data. And if it's an advantage that Roche has shown a benefit in patients that had such little previous endocrine therapy. That's my questions.

Great question about CDK4/6 inhibitors. So I would break this down by stage. LIDERA included patients with Stage 1 disease. Those patients were not included in NATLI or MonarRchE. So if the drug is approved for patients with Stage 1 disease, this would be the only endocrine therapy option without a CDK4/6 inhibitor.

In patients with Stage 2 disease, that was also included in NATLE. So there is some patient overlap. And ultimately, it will be a discussion between the patient and the provider, talking about the efficacy, which looks very favorable. If you do cross-trial comparisons with its caveats, the hazard ratio in NATLE was about 0.75. In LIDERA, it was 0.70. These are cross-trial comparisons. And we would also talk about toxicity when we talk to a patient about a drug, so both efficacy as well as toxicity with CDK4/6 inhibitors such as ribociclib, common toxicities that are seen include increase in liver enzymes, QTC prolongation as well as myelosuppression.

So that will be a discussion in terms of the AE profile. And then in patients with Stage III disease, there was overlap with Monarch. And again, it will be a discussion between the patient and the provider talking about the efficacy as well as the toxicity of these drugs with abemaciclib common toxicities, including diarrhea, myelosuppression. So based on the efficacy and the toxicity profile, these drugs would be used. Overall, it appears that as a single agent, which is not surprising, the drug has lower toxicity as compared to combination therapy.

So for the appropriate patient, giredestrant would be a good option to consider. trials.

Second question?

Yes, I can start addressing the second question. Thanks for that question on comparing with some of the other designs. So LIDERA, as we disclosed, it's addressed -- or fundamentally addressing the question of what's the best endocrine therapy upfront. And clearly, the results show that giradesant is superior to AI or tamoxifen, which are the current standard of care. When we think about the other study designs, they allow 2 or 3 years of endocrine therapies, which although we haven't seen the results, we're significantly ahead in this field. We anticipate that, that is an option that may become available in the future. However, in thinking about how the landscape might shape, we believe that an upfront choice of a better endocrine therapy, it's a better option for patients.

And I believe the 12 weeks, which you mentioned, were just the feasibility to include patients who had already started, but they could then enroll the trial.

Exactly. That's a common design. And sometimes patients as they are finishing their chemotherapy, they go through radiation therapy are started on endocrine therapies. So it's really not designed to have the switch concept. Many patients have a few weeks of endocrine therapies while they're finishing some of their treatments and then they're allowed in the study. And there were not a lot of patients. I mean, that is information we haven't disclosed and we haven't fully analyzed, but the majority of patients have not received any endocrine therapy prior to the beginning of the data.

Maybe I'll just add one more point on this. Obviously, it will be a long time before there's any kind of comparison between doing a switch approach versus starting upfront with a particular endocrine therapy. However, in oncology, a guiding principle, especially with targeted therapies is that you usually want to lead with your best option. It's not common in oncology that you would sit on the best medicine and wait until later. In general, you want to lead with your best options, especially if you have a chance to cure people and especially if the safety -- if the tolerability profile is favorable. So obviously, we can't speak about any data here because it will be a long, long time before we compare, but that is a general guiding principle. And as Pablo pointed out, we have shown that giredestrant is superior to alternative endocrine therapies in the adjuvant setting.

And maybe last point on the early separation of the IDFS curve. So we see already the effect of giredestrant kicking in at 6 months on IDFS. But then consider in CAMbre-1, when you only start after 2 years of endocrine therapy, this benefit is already lost for patients. You have already patients relapsing. So very clean design LIDERA, very clear answer, and it's superior and in line with what CDK4/6 inhibitors are showing as well.

Colin, did this answer your questions or you have a follow-up question?

Yes. No, that's great.

Very good. Next question is go to Peter Verdult from BNP Paribas.

Pete Verdult here from BNP. Just 2 maybe for Dr. Bal, just a variation on theme. I don't think you'll be surprised. But just interested to explore a little bit more how you think the standard of care in adjuvant breast cancer will change in light of LIDERA. I mean when you speak to people like Fabrice Andre, the orroohnsono, they're convinced SERD are going to be treatment practice changing. The discussion yesterday, perhaps a little bit more reserved in the absence of sort of SERD combo data with CDK4/6. So interested in your view, Dr. Badia, were giredestrant hypothetically approved tomorrow, how widely you'd be using it in your treatment practice? And then just a question on crystal ball gazing. When the appropriate studies are done, would you predict that, that SERD will show utility in that low-risk adjuvant patient group going forward?

Yes, absolutely. In terms of use of giredestrant when it is approved, it's the best endocrine agent we have at this time. So it would essentially replace AIs. Instead of AIs or tamoxifen, I would preferentially use giredestrant as we were discussing previously, Stage 1, Stage 2 disease, even in patients with Stage III disease. In patients with high, high-risk disease, whether giredestrant as a single agent is used, whether AI plus CDK4/6 inhibitors are used or giredestrant in combination with the CDK4/6 inhibitor is used. I think that's an open question at this time. Many physicians once there is safety data available with giredestrant plus, say, abemaciclib, we will feel comfortable using that in patients with high, high-risk disease because at the end of the day, we want to use the best endocrine option. The best endocrine option at this time appears to be giredestrant over AIs. So essentially, that's what would be used.

And utility in the -- eventually in the low-risk group, do you see that being an option or potential?

Absolutely, absolutely. So essentially, that's why LIDERA was fundamentally designed to test the question what's the best endocrine therapy option. And we see that giredestrant is superior to AIs and tamoxifen. So wherever AIs and tamoxifen is being used, that will get replaced with giredestrant in early breast cancer.

Questions from your side? Then we move on. Next in the row would be Richard Vosser, JP Morgan.

A couple of questions from my side. Just on that idea of having enough safety data in combination with the CDK4/6, when could we get the abemaciclib substudy? And second question is, does Roche plan to do a combination study anyway in the adjuvant setting for giredestrant? And would that sub-study be enough to tip the hand maybe probably not approvable for a combination, but what's the thoughts there? And then secondly, just maybe for Dr. Bardia and maybe also for Roche. We've seen in the subgroup versus obviously, a strong benefit here in the adjuvant. What's the thought process now around first-line treatment in wild-type patients as we look ahead to maybe PerRSEVERA? -- thoughts of the read from LIDERA, would be useful to hear your thoughts.

Thank you. I'm going to start with the first question. The verimaciclib sub-study is fully recruited. We anticipate data potentially towards the end of next year. We're looking into how that data matures. For other combinations, we are opening a similar approach potentially with more patients on ibociclib with the same concept as a sub-study on LIDERA. And I think there was a second question on how do we anticipate -- the -- was it on the label or the use, but I can comment. As Dr. Badia said, depending on access and physicians' willingness to combine, that could be an option that could become available. I don't know if anyone else wants to comment on this.

Maybe I could just jump in about sort of other additional studies that are planned. We're really considering our options now and are not able to publicly discuss that presently, but are definitely exploring multiple options to consider.

Maybe I'll jump in on the read-through or not to PerseEVERA. I mean one thing that's clear is that giredestrant when breast cancers are dependent on the estrogen receptor for viability, giredestrant can perform extremely well. We've now seen that in the metastatic setting where the second-line metastatic setting where ESR1 mutations essentially mark the tumors that still dependent on the estrogen receptor. It performed very well there. And now we've seen it in the adjuvant setting. All of those patients are, by definition, dependent on ER signaling. It has worked very well there. So clearly, in the front line, the likelihood that it's successful certainly hasn't gone down with these results.

However, we do need to caveat it. This is now data in combination with the CDK inhibitor. So we -- certainly, the chances are reasonable, but we have to be cautious because it's a different context, a different treatment combination. So we just have to see how the data plays out. But certainly, the overall scientific hypothesis remains very strong for the PRSEVERA study.

Any additional questions, Richard?

No, that's fine. Thank you...

And we move on. Next questions come from Luisa Hector from Berenberg.

In terms of LIDERA, do you think that regulators and doctors will want to see some longer-term efficacy data for approval or for adoption? And specifically in the Stage 1 subgroup, and this is clearly an open space in terms of treatment, but is there enough data from the subgroup of Stage 1? You flagged low level of events just due to the nature of the Stage 1, but do we have enough data there?

So that's the first question. I'd love to hear from Dr. Bardia on that. And then I wanted to check in terms of recruitment into LIDERA, was everything broadly as you planned it, sort of Stage comparator arms? And then check on the U.S. subgroup because you showed 40% or so in U.S. and Western Europe. Can you just confirm you have enough U.S. patients as well within that recruitment criteria?

Yes. In terms of use based on the data, I would say, yes, and we have historical precedents. If you look at the Monarche results or the NATALE results, they were presented with a median follow-up that's very similar, and we had 3-year IDFS results, and that led to the uptake of these agents. So we have precedents that as long as the curves separate and we're seeing good separation of the curves, we know that over time, this will be maintained. This drug also has lower toxicity as compared to standard of care endocrine therapy.

So once it's approved, I think there will be a lot of enthusiasm in terms of using this drug, not only from an efficacy perspective, but also a tolerability perspective, which is critical in the adjuvant setting. And that's why I feel even for patients with Stage 1 disease, we would be enthusiastic about using this drug because of the adlent safety profile.

Yes. And for regulatory authorities, it's about benefit risk. Keep in mind, when MonarchE and NATALEE trials reported, they added a lot of tox in comparison, certainly what we see here with dara, where we have even some benefits on toxicity. So from a benefit risk perspective, IDFS is an approvable endpoint. over survival is showing clearly pointing in the right direction. There should be no impediment what we currently see from authorities, which would put a show stop in front of us.

And I think there was another question on U.S. recruitment. And to follow up on this, of course, we're going to continue to follow up on the study. We expect the results to remain consistent, and we're going to continue reporting as the data matures. U.S. recruitment was close to 10% in this study. So we believe that's a significant representation from the U.S. And as we disclosed, if we consider U.S. or North America plus Western Europe is about 40%.

Lisa, any additional questions or -- the next questions go to James Quigley from Goldman Sachs.

I hope you can hear me. So 2 questions from -- or 2.5 questions from me. So just following up, Levi, on the probability of success for Persevera and your comment on the endocrine sensitivity. So you had that predictive transcriptomic analysis that seemed to work pretty well for the adjuvant setting. But what does it look like in terms of the first-line setting endocrine sensitivity relative to adjuvant? And did you enrich persevera for those patients who were more highly endocrine sensitivity? And what was the cutoff to assess that?

Second of all, just double-clicking on the combination plan in the adjuvant setting. Obviously, you've got the sub study with Verzenio. You've just mentioned that the Kisqali substudy will be larger. But what do you need to show here? What evidence do you need in order to particularly gain reimbursement here? And what level of evidence would you need to gain reimbursement support for the combination? Similarly, on that level, does it necessarily matter? Or would it be a case of giredestrant plus GDC-4198 could be your combination strategy? Also would love to know if that could be combined into a single pill. And the half question, what are the filing plans in the U.S. and globally for the adjuvant setting?

So I'll start on the [indiscernible] So to first treision, we didn't enrich for endocrine sensitivity per se in the front line. We -- because certainly in the frontline setting, in the treatment-naive setting, you don't really -- there's no mechanism to enrich for endocrine sensitivity. What we did do, however, is there is some enrichment for patients who had progressed relatively soon after receiving adjuvant therapy because there was some evidence that given what we know about fulvestrant, for example, in kind of analogous settings and given the bioavailability of giredestrant and its obviously ability to function as a potent SERD, we felt like that might -- there might be some advantage in such patient population. So we will need to see whether or not that hypothesis plays out. But certainly, there was no priority way to enrich for sensitivity to endocrine therapy. So maybe I'll see if others want to take the follow-up questions.

The question on there was a question...

Is required in order to obtain access in combo with CDK4...

Sorry, I'm happy to take the question about the expectation for access. I mean, usually, the -- for guidelines, the efficacy endpoint is preferred. So we expect the -- both the sub LIDERA studies for WABEA and Kisqali to provide important safety information to encourage physicians who would like to use both agents together. But it's difficult to speculate at this time that this would lead to label or guidelines change.

James, I think you had a couple of additional questions.

And definitely, we're like having plans to combine giredestrant with the 4198 CDK, but we're also considering other plans to combine giredestrant with other CDKIs other than the 4198. And this is actively being discussed now, and it's too early to disclose any of that.

Yes. And maybe I can just build on that. And obviously, combining giredestrant with other molecules in our portfolio is really core to our strategy, including 4198. And as Amar said, we've initiated a Phase I/II study of giredestrant and 4198 in the second-line metastatic setting. But it's really too early now to speculate on the combinations of these 2 molecules.

James, any additional questions?

The only additional one was the filing strategy or time lines for the adjuvant setting.

Yes. Yes.

So we are planning to file for both the U.S. and Europe next year.

Okay. Then we go on. Next questions go to Sachin Jain from Bank of America.

I've got a few more on positioning, if I may. So firstly, back to Dr. Bardia, just an answer to prior question, I think it was Pete's question on replacing endocrine therapy. Was that in reference to where it's used as mono? Or does that include where it's used as combo with CDK4/6? Obviously, one of the key debates in the market overnight has been the monotherapy endocrine usage is limited in that medium high risk given CDK4/6 usage. And the second is if you could just comment to what percentage of the population you think is unique to LR. So you've commented that Stage 1 is, but I wonder if you could overlap that. But I think it's your Slide 20, which points to intermediate high risk being 50% of the population.

Just trying to get a sense of what percentage of intermediate high risk you see as Stage 1 that is unique to LIDERA? And then the third question follows on, on the combination sort of topic. You talked to Kisqali Vzenio sub-study, but I wonder whether PERSEVERA, which is the first data of the combination could actually support combination usage in the adjuvant setting. And apologies for the background noise.

Great. Three questions. So I'll start first in terms of use of endocrine therapy as monotherapy for patients with medium or high risk. At the end of the day, we want to use an agent that is most efficacious and has lowest toxicity. That's the guiding principle in terms of treating patients with cancer, including breast cancer endocrine therapy plus CDK4/6 inhibitors showed that they were superior to endocrine therapy as a single agent, but that was with aromatase inhibitors and the hazard ratio was 0.75, and that led to uptake of CDK4/6 inhibitors for, I would say, a number of patients with Stage II disease, but it was less for patients with Stage II disease given the toxicities that we also see with CDK4/6 inhibitors.

Even in the clinical trial, the rate of discontinuation of CDK4/6 inhibitors in [indiscernible] was between 15% to 20%, while in LIDERA, this was less than 5%. So it just tells you more about the tolerability of the drug in this setting. Now with LIDERA, you see giredestrant in patients with medium as well as high-risk disease showed an improvement in invasive disease-free survival with a hazard ratio of 0.70. Again, cross-trial comparisons, but it looks favorable as compared to endocrine therapy plus a CDK4/6 inhibitor. So many physicians, particularly for patients with Stage II disease or patients with Stage III disease who have, say, 1 positive node will consider giredestrant as a single agent to balance the efficacy with the toxicity profile. But in patients who have, say, 10 positive nodes, many would be concerned about using just single agent would want to combine that with a CDK4/6 inhibitor, such as abemaciclib that's shown improvement in this setting. And that's why the safety data would be very helpful.

Once there is safety data with giredestrant plus abema, -- for patients with high, high-risk disease, many oncologists will discuss with the patient, but we feel comfortable describing giredestrant plus abema because at the end of the day, again, we want to use an agent that gives the best chance of cure in the adjuvant setting.

I think there was this follow-on question from Sachin as well, whether the first-line data, the combo data would help in establishing safety data for combos, if I got your question right, Sachin?

Yes, that's spot on. So that data obviously Per comes a long time before the Verzenio and Kisqali substudies.

It's tough to speculate without seeing the results. If the results are positive, there will be further confirmation that the best endocrine agent in combination with the CDK4/6 inhibitor is superior to AI plus CDK4/6 inhibitor. But on the other hand, as well, if the results are less significant or even if this is a negative study, that would not necessarily dissuade me from using the drug in the adjuvant setting. Metastatic setting is a different setting. It includes patients who have endocrine-sensitive disease, some resistant as well. It's looking at combination therapy. So it can maybe influence some adjuvant use. But I think ultimately, in the adjuvant setting, we have LIDERA study. It's the combination of giredestrant plus abema that will further guide the use of combination therapy.

And then I think we had one question remaining, which was about the percentage difference in terms of the patients included when you compare LIDERA versus NATALIE.

So I was just trying to get a sense of of the medium high risk, what percentage is Stage 1 that is unique to LIDERA?

So all patients with Stage 1 are unique to LIDERA because that was not included in NATLI or [indiscernible]. So any patient who had Stage 1 disease, that's only the LIERA population. Overall, that was about 10% in this trial. Patients who had Stage 2 disease, that's where it overlaps with NATLIE. And then patients with Stage III disease, not all, but the high-risk ones that overlaps with [indiscernible]

.

Apologies. Maybe my question. I'm just trying to understand what percentage of medium high risk in the population not in the study with Stage 1. So in your slide, Bruno, you have sort of 50% of patients being medium and high risk. What percentage of that 50% is Stage 1 is therefore unique to you?

So patients who have Stage 1 disease, that would be included in the medium risk that would not be included in the high-risk population. So if you look at the medium risk, that was about 30% in this clinical trial. And so that predominantly is patients with Stage 2 and Stage 1 disease. And Stage 2 was about 50%, Stage 1 was 10%. So if you do the calculations, you can probably see that maybe 20%, 25% of patients in the medium risk disease and Stage 1 disease.

Does this help, Sachin?

Yes, I'll leave it at that.

Okay. All questions from your side. And next one would be Simon Baker from Redburn.

Two, if I may, please. I just wonder if we could dig down into the dose interruption, AEs leading to dose interruptions, about twice as high in the giredestrant arm. I just wonder, are there any -- what was the nature of those -- the AEs leading to those dose interruptions, the length of dose interruptions? Did it have any outcome on -- any impact on outcomes? And does it give rise to any prospective dosing adjustments that you can take going forward? And then secondly, a question for Dr. Bardy, and it's kind of asking in reverse a question that's been asked a dozen times before on this call. Given the the consistency and the magnitude of the benefit we've seen here. Are there any patients within this population where you would not use giredestrant?

So I'll take the first question. Dose intensity was high in both arms, including the giredestrant arm. And as we highlighted, those discontinuations were lower, but you're asking about those interruptions. The data started in 2021. What we knew about the safety profile of the molecule at that time was different from what we know today in 2025. There was some protocol mandated schedule of assessments and some physicians had to interrupt because of, let's say, findings that patients were interrupted for a short period of time. As we looked into dose intensity, those interruptions didn't impact dose intensity. And then what's probably considered a harder endpoint which also physicians consider either toxicity or patients having AEs where the dose discontinuations, which were low with giredestrant.

Yes. And in terms of the second question, is there a patient population where I would not consider giredestrant? Based on the trial, I don't think so. If you look at the subgroups, the benefit was consistent. If you look at AEs, bradycardia was predominantly Grade 1. If a patient has Grade 3, Grade 4 bradycardia, maybe that's a setting where we would hold giredestrant if it's Grade 4 discontinue, but that was not seen in this clinical trial. So at least based on the data we have, I don't see a patient population or patient scenario where I would not use giredestrant.

Next questions go to Paul Koh from Cowen.

This is Steve Scala from TD Cowen. I have several questions. I think you said 10% of LIDERA patients were from the U.S. Does that meet the expectation of FDA for U.S. approval? It would appear that it might not. Secondly, what does Roche's $3 billion plus peak sales guidance for giredestrant assume for competition from other SERDs in the adjuvant and metastatic settings. All things considered, the number seems quite low. So does that imply that Roche expects a high level of competition? I appreciate the number has a plus after it, but you could have said $4 billion plus or $5 billion plus 2 billion. And then lastly, has Roche performed the interim PFS analysis yet for Perseveira? And is the interim or final PFS analysis coming in early 2026?

Let me maybe, Steve, just take the question on the peak sales. I think this is just a methodological setup here. We only provide when we guide 4 buckets, which is 1 to 2 -- what was it 0.5 to 1, 1 to 2, 2 to 3 plus. So 3 plus basically can mean a lot. It can be 3, 4, 5, 6, 7, 8, 9, 10, whatever you consider appropriate. So we do not provide more precise peak sales guidance like other companies do. We would not call out 6, 7, 8 or what we think is the right number.

And if I may add to that, Bruno, I think based on how advanced our readout is versus competition. So the next adjuvant trial is expected to read 2 years from now. And we believe with the efficacy and with the safety profile of giredestrant that we will be established as the best-in-class in that space.

On the U.S. enrollment, there's not a percentage cutoff that the FDA mandates for U.S. enrollment. What they want to see is that the results and the patient population are consistent with what one would encounter in the U.S. And so obviously, with a 4,000-patient study, we have hundreds of patients from the U.S. The effect in the U.S. was directionally similar to what was seen overall. And when you roll in similar populations such as Canada and Western Europe, you have even a much higher percentage of patients from the trial and the results are consistent. So we don't anticipate a challenge around the U.S. enrollment and applicability.

I would just add that the treatment is, I mean, basically identical across the world. It's altomoxifenI, a single-agent endocrine treatment. And certainly, there's no different health care context than to be applied. We are fully in line with what U.S. physician would expect to see.

Stephen, did this help? Or do you have follow-on questions on the competitive landscape?

The [indiscernible] question is, have you performed the interim PFS analysis of PerRSEVERA? And what exactly is coming early next year? Is it interim or final?

So I can comment on that. What we have in 2026 is the primary analysis.

Okay. Did you say final analysis?

Yes, lower primer, Yes, that's right.

Okay. Then we continue. Next questions go to Graham Perry from Citi.

Great. So just going back to the substudies on LIDERA, the Vzenio and Kisqali ones. Could you just remind us what you said on Vzenio when we see the Vzenio one? And then if you're just starting the Kisqali one, just any sort of time line on data from that. Can you confirm that second one has got both intermediate and high-risk patients in it? And then do you think those combo studies could be enough to help reimbursement if, for example, they can get a compendia listing, for example? And then the second question is just how do you think about pricing analogs here. Rromatasehibit became generic quite a long time ago, they were substantially lower priced than oncology products today. And perhaps the closest comparator you'd have now would be CDK4/6, but perhaps just help us to think about how you're thinking about that at this stage.

So I'll start with the question on specific designs and time lines on the combinations with abemaciclib and ribociclib. As I mentioned before, the abemaciclib is fully recruited. We are following up these patients. We anticipate data towards the end of 2026. Ribociclib, we're starting now. What we have used with the abemaciclib study, it's used the inclusion criteria for abemaciclib mirroring monarchine. So what we're going to do with ribociclib is mirroring the inclusion criteria of NATALE.

And with regard to access, I might I like to think about this in 2 buckets. I think when it comes to ex U.S., I don't expect like those substudies with safety endpoints will support reimbursement. But in the U.S., it could be a different story. And I think with additional data generation that we are planning to use, it's going to be a discussion within the health care system if those data could support reimbursement.

Go ahead. Any additional questions?

Yes, just the pricing analog one.

Pricing analogs, yes, CDK4/6 is this a reasonable pricing analog? Normally, we say too early. So we only will announce the price once we get approval. I think we'll leave it with that. Okay. Any additional questions?

No, that's great for me.

Okay. Then I think we will close the loop, and we have the final questions going to Emmanuel Papadakis. Emmanuel, please. Second set of questions.

I'll try and be quick. It was just a follow-up perhaps for Dr. Bardia on your optimism on potential use as a monotherapy in the lower-risk population. Just trying to understand that a little better since you had 0 patients classified technically as low risk in the study. As you mentioned, it was less than 15% Stage 1. In those, you had a 0.89 hazard ratio on IDFS. So just wanted to understand your optimism around potential clinical adoption in the lower to intermediate risk setting. Is that because you're expecting the regulatory agency may ultimately endorse a label explicitly encompassing a Stage 1 population? Is it because you're thinking oncology guidelines would recommend use in those patients? Or is it just that you think physicians will be so keen to upgrade their endocrine therapy for patients, they'll do that regardless of either of those 2 things? And then maybe just a follow-on on pricing. We have had some benchmarks established within the SERD space by the recent approval in LO and of course, EU. There's been times in the past where Roche has been very willing to come in later and take a more galitarian approach to pricing to open up access for a broader population. Is that something you would consider in this circumstance? It would seem presumably likely to access a broader population in the adjuvant setting, you might need to consider that.

So I can answer the question about patients with low risk or Stage 1 disease. The optimism is because we feel that this is the best endocrine agent based on the mechanistic preclinical as well as the clinical data that we've seen. And then also, this is the agent that has a better tolerability profile in terms of discontinuation as compared to AIs. And if we look at the hazard ratio, it's trended in favor of giredestrant. It's just that the number of events are few. So we need more follow-up before we would have a statistically significant result because that Stage 1 disease, the number of events are few. But because this is an agent that we feel is the best endocrine agent and also has a safety profile that is very clean, that is the reason why we're optimistic about this agent even for patients with Stage 1 disease. In terms of when we would see the results, I think at this time, further follow-up is continuing.

I think there was another question about pricing. And if I understood the question correctly, please correct me if I'm wrong, that whether we would consider like changing or adjusting the price based on the future patient population that would be approved through the address studies. If that's the case, I would say, I think we're in a good situation that we have both Avera and LIDERA positive that will inform our pricing assumptions given that LIDERA is going to be the broadest patient population. So I wouldn't expect future changes to the price based on subsequent study readouts.

Yes. Okay. One more question coming in from Richard Vosser, JPMorgan.

Just one follow-up to Emmanuel's question. Just in that low-risk population, the proportion of patients on tamoxifen, was it different? Did you have more tamoxifen use in the very low-risk Stage 1 patients compared to letrozole or something else? Or is it similar across the whole trial? Obviously, it might have implications for the efficacy advantage in those patients.

Thank you for the question. That's an excellent question. We -- we unblinded the study 3 weeks ago, and we still don't have that answer. However, very much with your thinking, we anticipate that the use of tamoxifen is leaning towards patients who are either premenopausal or lower risk, but acknowledging that we don't have that data yet.

Very good. And there is one more question from Cowen from Stephen Scala...

And apologies, another question about PerseVERA. So there was a planned interim analysis that was to be performed when 70% of progression-free survival events occurred. And you answered the question earlier citing the final analysis. So should we assume that the interim analysis has passed and the study is continuing to final?

We don't comment on interim analysis. So at this juncture, we'll comment at the final.

Thanks for understanding, Stephen, but we cannot provide more granularity on that one. Okay. With that, I think we are at the end of today's call. I would like to thank all the participants here. I would like to thank the IR team members who prepared the slide deck, that's Rafael Pavlolowski and Lauren Kam and also Melanie Wolf for the event organization. I hope you went -- was helpful providing a timely update to our breast cancer franchise. If there are any remaining questions in the coming days or next week, please reach out to the Roche IR team. We are happy to further follow up.

And with that, I would like to wish you all, in case we have no interactions anymore, I would like to wish you a peaceful holiday season, Merry Christmas and a good start into 2026. Bye-bye.

[indiscernible] webinar 2025. My name is Henrik, and I'm the technical operator for today's call. Kindly note that the webinar is being recorded. [Operator Instructions] One last remark, if you would like to follow the presenter's slides on your end as well, please feel free to go to roche.com/investors to download the presentation. At this time, it's my pleasure to introduce you to Thomas Schinecker, CEO Roche Group. Mr. Schinecker, the stage is yours.

Thank you very much, and good morning, good afternoon, and I'm happy to share with you our Q3 2025 results. So let's start with our performance. We continue to see strong performance for the Roche Group. Year-to-date, group sales still same as half year at 7% growth, really driven by Pharma with 9%. Diagnostics grew year-to-date with 1%. So we see a return to growth. And Diagnostics has, of course, been impacted by the health care pricing reforms in China, something that we have communicated multiple times throughout the year. If we exclude China in the numbers, then the rest of Diagnostics was growing 7%, which is absolutely in line with past performance of Diagnostics. So we do see that Diagnostics continues to perform well. On the LOE impact, we've lowered this now to CHF 800 million impact, down from CHF 1 billion, which we still showed at the half year results.

In Q3, we've had several important milestones that we've achieved. On the Pharma regulatory side, the U.S. approval for Gazyva in lupus nephritis and the positive CHMP opinion in the EU. The U.S. approval for Tecentriq in first-line maintenance small cell lung cancer, U.S. filing for Susvimo in nAMD. And we had several positive Phase III readouts. For example, the positive Phase III readout of evERA giredestrant results were just presented at ESMO on Saturday, the study met the primary endpoint, both in the ESR1 mutant population, but also in the ITT population. Definitely, Teresa will talk more about that.

We had a positive Phase III without for Tecentriq, a muscle-invasive bladder cancer. We had trial results shown from Phase III in vamikibart in UME and satralizumab. This we just presented at AAO, for example. And these are some encouraging results looking at the efficacy of these medicines, and we will discuss those results with global health authorities. Also here, Teresa will talk more about that.

Very importantly, we really refilled our late-stage pipeline also in the last quarter. We took 5 additional Phase III decisions on top of the 4 previously announced at half year, so an overall significant progress in our pipeline. There is CT-388 in obesity, CT-868 in type 1 diabetes, zilebesiran in hypertension. Study has already started. Cevostamab in relapsed/refractory multiple myeloma and in HER2 tyrosine kinase inhibitor in HER2 positive breast cancer. Importantly, 4 out of the 5 came actually through partnership agreements in the last 2 years based on positive data, we are moving those forward.

On the BD front, we've entered into an agreement to acquire 89bio with the potentially best-in-disease FGF21 analog in MASH. And this deal is highly synergistic with our CRM portfolio, providing further optionalities for combination therapies. We've also entered into an agreement with Hansoh Pharma to license the CDH17 ADC, which is currently in Phase I to be developed in colorectal cancer.

On the Diagnostics side, with a series of regulatory updates. Most importantly, the first blood-based rule-out test, Elecsys pTau181, which got the CE mark, but also the FDA clearance. On the FDA clearance with the primary setting claim, which is, I think, very important to identify patients and to bring them on therapy. Elecsys Troponin-T high sensitive Generation 6, but also the AI algorithm Kidney Klinrisk with the CE mark. There's some news flow left towards the end of the year and Teresa and Matt will cover that, for example, fenebrutinib, Gazyva in SLE as well as PiaSky and several diagnostics launches still coming this year.

So again, let's look at the overall momentum and the sales numbers. You can see here Pharmaceuticals still doing extremely well with a 9% growth rate. So very happy with that. Diagnostics growing 1%. I explained the numbers, given the health care pricing reforms in China. This is an effect that severely impacted us this year. There will be some effects next year, but we will see a continuous improvement. So overall, the group is at 7%. If you look at the growth rates over the last 2.5 years, we've consistently performed especially if you look at the base business, excluding COVID with 8% growth on average. Again, this year, Q1, Q2, Q3 were impacted by the health care pricing reforms in China in Diagnostics. You see that Pharma is still growing at 9%. So we continue to have a positive momentum.

Now let me talk about some of the key drivers in the Roche portfolio, starting with the top right and going clockwise. First with oncology and hematology, the Phesgo global conversion rates are now at 51%, plus 5% versus previous quarter. We believe that we can get to at least 60% global conversion rate. Alecensa's growth number remains good, driven by the adjuvant ALK-positive non-small cell lung cancer. As mentioned, we have the positive results from evERA in Phase III, which represented at the ESMO. Polivy's continued strong uptake in first-line DLBCL, now reaching 35% patient share in the U.S. and truly has established itself as the new standard of care. Hemlibra continues to grow strongly with full year growth now expected in the -- around 10% range. New and -- is also that the NXT007 data will be presented later at ASH.

On the neurology side, and I know Teresa will talk about that as well. Ocrevus, Zunovo, we have now 50% of starts in the U.S. are new to brand, and Evrysdi is the #1 SMA treatment and keeps growing with more than 21,000 patients being on treatment. On the immunology side, Xolair continues to work to do well. We have a strong uptake now growing at 34%, and in food allergy, we now have 85 -- more than 85,000 patients on this medicine. The U.S. approval that we've received for Gazyva in lupus nephritis and we got also a positive CHMP opinion in Europe. We expect also the Phase III results for Gazyva in SLE to read out later this year.

On the ophthalmology side, we continue to have overall strong global growth in Vabysmo, driven by its differentiated MOA and profile. But we do see ongoing contraction of the branded market segment in the U.S. because of less funding available for these assistance -- copy of assistance foundations. On the Diagnostics side, as mentioned, good growth in the overall business, the Core Lab being mostly impacted by the general health care pricing reforms without that Diagnostics is growing at 7%.

Now let me give you a little bit of a pipeline update. On the left-hand side, you see how our pipeline prioritization has continued, and we constantly apply the bar to all of the molecules that are in our pipeline. So we constantly make decisions based on new data available or also external data available and keep prioritizing so we can really allocate the funds to those projects that have the highest impact and the highest potential. And with that, we can also accelerate a number of these programs.

And you can see that also how it translates into the overall portfolio value. If you look at just peak sales per pipeline project, we've increased it now by 57%. So before this prioritization, we had an average value of pipeline project of about CHF 800 million. We're now at CHF 1.3 billion. And also the total portfolio value in this time period has increased by 27%, which I think is exactly what we wanted to achieve with R&D excellence and the prioritization in our portfolio.

What's really exciting is if we look at this slide, it's how many molecules have moved into Phase III this year. This is by far a record for Roche, and we are not even done with this year. So 10 moving into Phase III. And these are all molecules that will launch by the end of this decade. So we have a lot, I think, to look forward to. And we really see it across all of our therapeutic areas. We see it in oncology with the HER2 tyrosine kinase inhibitor, cevostamab, also NXT007 in hemophilia A, received in the neurology front in Alzheimer's and Parkinson's. We see it also in the cardiovascular area. So we are progressing very well. And this really is going to impact our growth momentum at the end of this decade and into the next decade as well.

The way you can also see how we apply the bar is as we assess PTS project internally and the value in the business cases behind each of the project. You see where the projects lie for those assets that are pre bar, which are in the gray dots. And in the dark blue dots, you see all the assets that we moved into Phase III. We haven't added because this is the slide from Pharma Day. We haven't added all the new ones that we've added into Phase III since then. But I can say they're all really promising as well. We didn't do that because otherwise, you would know exactly which dot is which dot. But I just wanted to point that out that we haven't added all of this. So you see the overvalue consistently moving up and the PTS consistently moving up. And you can see all of the blue dots on the top right corner.

Now this doesn't mean that there will not be any assets that we will have on the left-hand side. Just to point out the reasons for these 2 dots. One of them, the lower one is the novel antibiotic. We know that there is currently no market for that we see it as an opportunity. Also in case there will be pandemic and there will be a pandemic in the future, but we also see that it's an ethical responsibility. And the other one is first indication in oncology, and we know that this will move into further indications. So it will move further to the right as we proceed with additional projects. You can also see on the right-hand side, the portfolio composition through different phases of development, the ones that have moved through the bar and the ones that we are introduced pre the bar. But you can be assured that every portfolio decision that we made in the transitions, they will always have to meet the bar.

Now let me talk about also some of the news flow. And here, you can see, starting with afimkibart that we've initiated at Phase II study in a new indication, which is rheumatoid arthritis. This is a TL1A, where we will see the first readout in 2027, but not in this indication, but in other indications. Afimkibart, we've talked about the Phase III readout, where you can see the efficacy of the molecule giredestrant with a positive readout for evERA, Teresa will talk more about that.

Divarasib, there's new Phase III for divarasib in adjuvant setting in non-small cell lung cancer. We've talked about the portfolio transitions of CT-388, CT-868 cevostamab, HER2 tyrosine kinase inhibitor, and giredestrant, all based on data. Some of these data, you will see as we go into next year in the ADA, especially on the obesity and cardio metabolism medicines. Pegozafermin is potentially best-in-disease FGF21 analog. Currently in 2 Phase III for MASH in the stages of F2 to F3, but also in MASH for F4. And we have announced this deal recently, and we look forward to the closing of the deal.

Now let's move on the next slide to Diagnostics, which is also very exciting. If you look at some of the key technologies that will drive future growth. Here, you have some of those clearly mass spectrometry, which is one of its kind, something that truly sets us apart from our competitors and differentiates us even beyond Mass Spec in the clinical chemistry and immunochemistry setting. Here, we are in the rollout and we will expect a full U.S. launch in 2026.

Accu-Chek SmartGuide, CGM also here, we're making good progress. We're ramping up manufacturing, and very exciting, next year, we're going to launch the Roche sequencing solution. You have seen some of the data that was just recently presented at ASHG and also the fact that this sequencing solution set a new world record in terms of speed in sequencing a whole human genome under 4 hours. So we really have a very competitive product here, and we really look forward to launching this solution next year. All of these areas are, I would say, if you look at Pharma terms, blockbuster potential so opportunities that are in the CHF 1 billion-plus size of opportunity. So really exciting, and I believe this will really cement the future growth of Roche Diagnostics.

Now let me talk about the outlook. First, the slide you have seen previously also shown in different events like the Pharma Day event. On the bottom, you see Diagnostics. And diagnostics, we do believe that we can continue the growth momentum that we've seen in the past years. And this will really be cemented by the blockbuster opportunities with Mass Spec, CGM, SPX, but I also believe very much with something like LumiraDx. So we do believe that we can continue to grow mid- to high single digit. We will have to wash out the China effect. There will still be some effects next year on that. Corp to grow ahead of sales growth.

And then you look on the Pharma side. On the Pharma side, we have very good momentum in our portfolio. You've seen a 9% growth. And we do believe that we are one of the companies with -- which have less biosimilar erosion. I mean if you look at the next couple of years, we do believe we can compensate that with the growth momentum that we have with the medicines that we have in hand, and that we will continue to grow at least until '28. And thereafter, the growth of the new -- of the existing medicines will compensate any additional biosimilar erosion. And anything we have on top in terms of pipeline readouts, positive Phase III result will then add to growth.

So we don't see a case where we will not continue to grow. In fact, any of these readouts that you see any of these 19 potential medicines that we can launch by the end of the -- at the end of this decade will continue to contribute to future growth. So we believe we have set up growth. And we also know we're not done with BD. We can continue to invest in BD and bring in more opportunities that will continue to drive growth also into the future.

Finally, let me talk about the guidance. Group sales growth mid-single digit. And just to take that upfront when we grow 7%, we do believe 7% is in the mid-single digit range. We look at Core EPS at half year, we still talked about high single-digit growth. You saw our numbers at the half year. That's why we are also increasing here the guidance from high single digit to high single digit to low double-digit core EPS growth and we also believe that we can further increase dividend in Swiss francs.

With that, I hand over to Alan.

Yes. Thanks, Thomas. Sales calls today. So just a few comments from my side on sales currency impacts and a couple of comments on the guidance for 2025. When you look at the sales bridge, I think that's the sales price in Swiss francs, you see on the left-hand side, 2024. On the right-hand side, 2025, 2% growth in total. And then the bridge splits it up between the growth in constant rates and the currency impact. So when I go through the bridge elements for the growth in constant rates 7%, then you see Pharma quite impressive growth, CHF 3.4 billion. And then you see the loss of exclusivity impact of minus CHF 474 million in total, a growth of 9% in constant rates.

As Thomas alluded to, based on the effect that we have seen so far in the loss of exclusivity section, we think that the new indication for full year should be an CHF 800 million loss due to the loss of exclusivities. When you go to Diagnostics, also this Thomas has mentioned, Diagnostics has grown, excluding China, with plus 7% in constant rates, and Matt will lead you through this. And then you see really the China health care pricing reform impact of a minus CHF 517 million, minus 6% -- minus 6 percentage points, which then in total provides Diagnostics with a growth of 1% in constant rates. You see the currency impact, which accounts for minus 5 percentage points leads you to the 2 percentage points growth in Swiss francs.

Well, where does it come from? I think first, certainly, I think the Swiss franc has strengthened again, against all major currencies. And you see there on the left-hand side, you see the growth in constant rates, now 6.7%, that's the 7% rounded that we have mentioned all the time. And on the right-hand side, you see the growth in Swiss francs was plus 2%. Distance here is minus 4.7 percentage points, that's the minus 5 percentage points that I've mentioned before. And you see very clearly the major driver here is the U.S. dollar.

Yes. Let me get to the currency impact. And yes, we started the year quite well. Well, but now I think you see -- when you look to the left-hand side, the U.S. dollar has further weakened against the Swiss franc. And when you go to the euro, the euro has shown some stability as you can take from the slide here. But as I said, I think certainly dominated by the U.S. dollar. So when you go to the right-hand side and our -- if you like, our modeling and our forecasting for year-end, and this really is assuming that September 30 exchange rates remain stable until the end of 2025, which is certainly very unlikely. But when you look really at full year, you see really a minus 5 percentage points impact on sales, minus 8 percentage points impact on core operating profit and minus 8% -- percentage point impact on core EPS.

You might remember, when you looked at this table at half year, we had for core EPS, minus 6 percentage points, and I think that's important. So that has worsened by 2 percentage points to minus 8 percentage points. And that has certainly an impact on the consensus. And I will talk about this in the context of the guidance here. So as you've seen in Q3, we have seen an increase in the expected currency impact on core EPS for the full year 2025 by minus 2 percentage points from minus 6 percentage points which we have forecasted at the end of June to minus 8 percentage points that we are forecasting now for year-end.

When looking at the post half year 2025 consensus, the core EPS growth stood at around plus 5.5%. We believe that this estimate would need to be adjusted for the guided additional currency deterioration of minus 2 percentage points which we have experienced in Q3 for the full year 2025 projection. Looking at latest consensus estimates, this adjustment seems not yet to have happened in most of them. With this adjustment, we feel okay with the consensus overall.

Good. And I think with that, I hand over to Teresa.

Great. Thank you very much, Alan. So let's start by taking a look at the overall performance for Pharma. So as Thomas mentioned, pharma sales grew by 9% at constant exchange rates, reaching CHF 35.6 billion. All regions are delivering strong growth, led by our international region, which grew at 13%. Overall, Pharma volumes were up by 13%.

As always, let's start with a reminder that on this slide, all absolute values and year-over-year growth rates are presented in Swiss francs at constant exchange rates. At our top brands, Phesgo, Xolair, Hemlibra, Vabysmo, Ocrevus and Polivy, generated roughly CHF 2.7 billion in new sales. These growth drivers represent a diversified mix of therapeutic areas and contribution from all of our geographies to growth. Phesgo continues to be our #1 growth driver with Xolair our close second as the outstanding launch in food allergy continues.

Now let's start our TA deep dives with oncology. Oncology sales increased 2% to CHF 11.6 billion, primarily driven by our HER2 franchise. Phesgo posted an impressive 54% growth year-to-date. We're very happy to see that, as Thomas mentioned, the global conversion rate climbed 51% this quarter and as we have already achieved our goal of 50% conversion. We are now increasing our ambition to hit a global conversion rate of more than 60% before the advent of biosimilars. Looking at Perjeta, the conversion to Phesgo continues and Kadcyla growth is -- continues to be driven by uptake in adjuvant breast cancer.

I do want to acknowledge that in line with expectations, the data that was recently shared from competitors, in particular, DESTINY-05, will have a negative impact to Kadcyla of roughly CHF 700 million to CHF 1 billion over time. This is fully in line with the HER2 franchise outlook that we have previously provided. As a reminder, we expect the HER2 franchise to peak at around CHF 9 billion in 2026 followed by a steady decline through the end of the decade with a solid tale of around CHF 4 billion, and that's primarily Phesgo, around CHF 1 billion for Kadcyla in a bit of H&P. Let me also confirm again that we do not foresee any cliff situation.

Staying with our breast cancer franchise, let's briefly talk about our HER2 TKI. We did make the decision to start a Phase II/III study in HER2 breast cancer next year. Our HER2 TKI is highly selective with strong blood-brain barrier permeability and CNS retention. These characteristics could be key to overall disease control, prevention and treatment of brain metastases. About 50% of patients with metastatic HER2-positive breast cancer develop brain mets, and we believe our HER2 TKI has the potential to address a key unmet need in this population. And last but certainly not least in our breast cancer franchise. Itovebi's launch in first-line PI3-kinase mutated hormone receptor positive breast cancer is ongoing and progressing as expected.

Moving on to Tecentriq. Our sales gain overall stable for the -- at Q3 and for the full year, we expect the same. There were multiple positive events for Tecentriq in Q3 IMforte and first-line maintenance small cell lung cancer achieved U.S. approval. IMvigor011 in muscle invasive bladder cancer reported positive Phase III results and those data were presented at ESMO. And ATOMIC in adjuvant dMMR colon cancer was just included in the latest NCCN guideline update. Taken together, we see several hundred million in incremental sales opportunity from these additional indications. On the back of this and since Tecentriq has returned to growth in the U.S. in Q3, we now expect to see flat to low single-digit growth for Tecentriq in the coming years.

Also in Q3, we initiated the Phase III KRASCENDO-2 study of divarasib in first-line non-small cell, and that enrollment is actually progressing ahead of plan. Looking ahead, as mentioned at Pharma Day, the persevERA readout of giredestrant is expected in 2026. And speaking of giredestrant, we wanted to take a closer look at the positive evERA results on the next slide.

Over the weekend, we presented the evERA data at ESMO in Berlin and here are some of the highlights. As a quick reminder, 100% of patients in evERA have received a prior endocrine therapy and a prior CDK4/6 inhibitor. The results clearly show that giredestrant significantly improved PFS in patients with ER-positive metastatic breast cancer. Both for patients with ESR1 mutations and for the ITT population, displaying strong HR values of 0.38 and 0.56, respectively. Benefits in PFS, ORR and DoR were observed across key subgroups irrespective of ESR mutation status. Although overall survival remains immature at this time, we saw a positive trend in both the ESR mutant and ITT populations. Furthermore, an exploratory analysis in patients without ESR1 mutation detected -- also showed favorable trends for PFS and OS with hazard ratios of 0.84 and 0.79, respectively. Taken together, evERA results reinforce our confidence in the giredestrant clinical development program and show the potential for giredestrant to become the next-generation best-in-class endocrine therapy.

Let's also take a quick look at the safety results from evERA. EvERA was well tolerated and the safety profile was consistent with previous studies. Of note, there were no new safety signals observed, including no cases of photopsia. We find the fact that giredestrant can be combined at the full dose of the CDK4/6 inhibitor without showing any photopsia, very encouraging as other CNF inhibitors in development have shown rates of up to 20%. Again, this strengthens our belief in the potential for giredestrant to become a next-generation backbone therapy, and we would expect in this initial location something around CHF 500 million plus in sales.

Now let's take a look at our hematology franchise. The hematology franchise had strong growth of 15%, delivering CHF 6.4 billion in sales, Hemlibra continues to deliver strong growth of 12% year-to-date, driven by increasing adoption in non-inhibitor patients. While we did see some softness in the U.S. for Q3, this is primarily due to buying patterns given that we had some strong buy-ins in Q2. And based on the strong global performance, we are upgrading our full year outlook for Hemlibra to around 10% growth, and that's up from mid-single digit. A brief note on NXT007, our next-generation bispecific in Phase III development for hemophilia A. We look forward to sharing additional Phase I data with you in Q4.

Next up, let's take a closer look at our malignant hematology portfolio. Polivy and first-line DLBCL continues to drive strong growth. And U.S. first-line DLBCL patient share is climbing further and now stands at 35%, that's up 2 percentage points versus Q2. We anticipate that Polivy will surpass CHF 1 billion in sales in the first-line DLB setting for the first time this year.

Shifting to Columvi and Lunsumio, our CDK CD3 specifics, [ launch ] performance remains on track for Columvi in third-line DLBCL and Lunsumio third-line plus follicular lymphoma. These initial later line indications are expected to deliver combined peak sales of several hundred million Swiss francs. We also had additional news flow for both of these molecules this quarter. For Columvi, SKYGLO was accepted as the new confirmatory study for the post-marketing requirement and for Lunsumio, we received a positive EU CHMP opinion for the subcutaneous formulation. And as a reminder, the U.S. approval for the subcu formulation is expected later this year. As Thomas mentioned, we took the decision to move cevostamab into Phase III development for relapsed/refractory multiple myeloma, and that trial is also expected to start next year.

Now let's go to the next slide where we will cover the neurology franchise. Our neurology franchise achieved CHF 7.3 billion in sales and a strong growth of 9% year-to-date. Ocrevus continues to have good momentum, delivering 7% growth globally. We continue to be encouraged by the ongoing launch of our subcutaneous formulation known as Zunovo in the U.S. As a reminder, the permanent J code for Zunovo was granted on April 1, and we do see some acceleration in Zunovo uptake as a consequence of that. It is important to remember that for many practices switching to Zunovo is associated with logistical challenges. We are hoping to manage those, and we expect Zunovo update to accelerate further in coming quarters.

Let's take a quick look at some of the underlying data that gives us confidence in the trajectory of the subcut. In the U.S., roughly 50% of Zunovo patients are naive to Ocrevus and the same is true for many of our other early launch countries such as Germany, for instance.

We're encouraged by the fact that about 60% of U.S. subcut volume is coming from community practices with the prescriber breadth climbing to roughly 800 HCPs. This demonstrates that Ocrevus subcut is expanding the addressable market and can help overcome health care system constraints like IV capacity limitations. Overall, we now have more than 12,500 patients on Ocrevus Zunovo globally. We also presented the 2-year data for Ocrevus subcut at ECTRIMS, which shows that the benefit risk profile is maintained. And for 2025, we continue to expect high single-digit global sales growth for Ocrevus. It is worth noting here that the U.S. Q3 performance was impacted by a base effect due to a one-off buying pattern that occurred in Q3 of last year. As we previously shared at Pharma Day, we have upgraded our peak sales expectations for the Ocrevus franchise to CHF 9 billion by 2029, and this includes CHF 2 billion incremental sales from Ocrevus subcut. Staying with our MS franchise, we are eagerly awaiting the fenebrutinib results in PPMS with the Phase III FENtrepid readout expected in Q4.

Moving on to Evrysdi. We now have 21,000 patients on Evrysdi globally, and this includes patients using our new tablet formulation, which has now launched in both the U.S. and EU. A quick note here that Q3 performance in the international region was impacted by tender-related buying patterns with a larger buy-in expected in Q4. Moving on to Evrysdi and DMD, we continue to believe in the risk-benefit profile for Evrysdi in the ambulatory DMD population, and we have approximately 840 patients have been treated globally in the setting. We are continuing to work with the EMA to find a viable path forward for EU patient access following the negative approval decision.

And lastly, Trontinemab, and AD, the final Phase I/II data from the 1.8 and 3.6 milligram cohorts were presented at AAIC. We have already achieved FPI for our Phase III TRONTIER 1/2 study in early AD. And currently, there are roughly 4,000 patients in the pre-screener study traveler. So we're optimistic that enrollment there will also continue at pace. Also at AAIC, we shared our plan to initiate a Phase III in preclinical AD and more details on that to come at a later stage.

On the next slide, let's look at the immunology franchise. Our immunology franchise grew at 15% at constant exchange rate and reached CHF 5 billion in sales. Xolair continues to be the primary growth driver, thanks to a very strong launch in food allergy. Year-to-date, we see growth of 34% for Xolair. And based on this strong performance, we have also updated the full year outlook for Xolair. We now expect growth greater than 30% in 2025, up from the around 20% estimate that we shared at half year. As a reminder, we expect a first biosimilar launch for Xolair at the end of 2026.

Actemra sales grew 2% year-to-date with gradually increasing biosimilar erosion impacting performance. In Q3, we saw a decline of 6% in the U.S. due to biosimilar competition. This is aligned with our expectation that biosimilar impact would accelerate in the second half of 2025. We are anticipating that Actemra will see a single-digit sales decline in 2025, which has already been included in our LOE outlook of CHF 800 million for full year 2025.

A highlight of this quarter in immunology is certainly Gazyva. We achieved FDA approval in lupus nephritis with a strong label and are looking forward to bringing this medicine to patients. The FDA label includes all active lupus nephritis patients receiving standard therapy without limitation to specific background therapies as well as renal-related events and death data, both of which a substantial reduction. It also included simultaneous approval of short infusion and inclusion of biomarker and subgroup data that shows consistent Gazyva benefit across subgroups and rapid effect on biomarkers of disease activity. At the same time, we also received a positive EU CHMP opinion and then expect EU approval to follow later this year. We are also expecting to see the readout of the Phase III ALLEGORY trial for Gazyva in SLE in Q4 of this year.

Turning to Xofluza. Last week, we announced our direct-to-patient program for Xofluza for patients in the U.S. This program will provide Xofluza at $50 through selected pharmacies, a 70% discount versus list price. Via this program, we aim to improve affordability and access for U.S. patients and address the significant health impact that the flu continues to have on society. With the flu season in the Northern Hemisphere, expected to start soon, we believe it now is the right time to launch our direct-to-patient program.

So now let's move on to ophthalmology. Ophthalmology grew by 11%, achieving CHF 3.2 billion in sales year-to-date. The Vabysmo performance remains impacted by the contraction of the U.S. branded market with 13% sales growth year-to-date. We have all been watching the U.S. market closely. And by now, we see a decline of the branded IVT market of roughly 15% so far this year. In the EU, Q3 performance was also impacted by mandatory price cuts. However, we believe that the ongoing rollout of the PFS formulation in the EU will fuel further growth.

Considering these dynamics, we have decided to update our full year growth outlook for Vabysmo. We now expect around 15% growth at constant exchange rates, down from the roughly 20 that we had shared earlier this year. Nevertheless, Vabysmo continues to gain market share in the branded IVT market in the U.S. and across early launch countries globally. In the U.S., we now see that more than 60% of Vabysmo starts are naive patients, further solidifying Vabysmo's position as the standard of care. This is also supported by our recent survey amongst ophthalmologists conducted by ASRS. The survey shows that more than half of ophthalmologists see Vabysmo as the IVT treatment option with the best anatomic outcomes and disease control in nAMD, DME and RVO. China also continues to impress with strong update following the NRDL listing and rapidly expanding market shares.

Moving on to the pipeline. As Thomas mentioned, we had 3 Phase III readouts for vamikibart in UME -- I'm sorry, we had 2 Phase III readouts for vamikibart in UME and satralizumab and thyroid eye disease with data presented over the last few days. So let's take a closer look at both. Starting with the vamikibart results in UME, which we presented earlier this week at AAO. The Phase III results from MEERKAT and SANDCAT show rapid improvement in vision and reduction in macular edema as well as a well-tolerated safety profile. You can clearly see the vision improvement and reduction in macular edema in the BCVA and CST curves on the right side of the slide. The primary endpoint was the proportion of patients with greater than 15-letter BCVA gain. MEERKAT met this endpoint, but SANDCAT did not.

However, let me emphasize 2 points. In both studies, a numerically higher portion of patients treated with vamikibart gained vision. The miss for SANDKAT was primarily driven by 1 patient in the 1 milligram vamikibart cohort. It is our belief that the totality of the SANDKAT and MEERKAT data support the use of vamikibart in UME. Vamikibart may offer a novel targeted IVT treatment strategy for UME patients that does not involve steroids and therefore, could address a significant unmet need in this underserved patient population. As mentioned at our IR event earlier this week, we will discuss this data with health authorities to determine next steps.

Moving on to satralizumab in thyroid eye disease. We presented the Phase III SatraGO-1/2 at ASOPRS a few days ago. The Phase III results show that satra and TED have a differentiated risk/benefit profile from currently available treatment options. The primary endpoint was proptosis response at week 24 in active thyroid eye disease. For both SatraGO-1 and 2, a higher portion of participants treated with satra showed a proptosis response, however, only SatraGO-2 achieved statistical significance, SatraGO-1 did not.

Importantly, both studies did show a clinically meaningful improvement across a range of key efficacy endpoints including diplopia CAS and proptosis in the active and inactive TED population. Furthermore, the safety profile for satralizumab was highly favorable. Currently available treatments for TED are associated with sometimes severe adverse effects and none of these effects were reported for satra. As with vamikibart, we believe that the totality of evidence support its use as a treatment for TED patients, and we are planning to discuss the results with health authorities.

And now moving on to our CVRM portfolio. In terms of developments in our CVRM pipeline, the recently announced 89Bio acquisition is currently -- is certainly a highlight, and I will cover this in more detail on the next slide, but suffice it to say that we are very excited to add a potential best-in-disease enemy and MASH into our portfolio. Most of the other updates you've already seen at Pharma Day, so just very quickly, as you know, we are moving CT-388 and zilebesiran into Phase III trials. Zilebesiran Phase III ZENITH has already achieved FPI. Additionally, CT-996 achieved FPI for a Phase II study and announced today, we will be moving CT-868 into Phase III for type 1 diabetes patients into late-stage trials. And these data are currently in-house and we will be sharing at ADA next year. In fact, most of the Phase II trials here will be shared at ADA next year.

So as you can see, we are progressing our CVRM pipeline at pace. But let's take a deeper look at pegozafermin, our new Phase III asset in MASH. For those following the field, it is important to understand that not all FGF21 analogs are the same. Pegozafermin is uniquely engineered for balanced efficacy and extended dosing. It also shows good potential for combination development, including a well-tolerated safety profile. Therefore, we truly believe that pegozafermin has the potential to become a best-in-disease treatment in MASH as well as adding combination optionality to our CVRM portfolio.

On the right, you can see the Phase II data, which we also presented at Pharma Day. And in short, pegozafermin demonstrated clear treatment benefits in F2 to F4 patients. And this potential to bring a treatment benefit to F4 patients, those with the most advanced form of MASH is a big part of the reason why we are so excited for pegozafermin. There are 2 Phase III trials currently ongoing, in MASH(F2-F3) and one in MASH(F4). We expect the top line data for F2, F3 in the first half of 2027 and for F4, we expect top line data in 2028. This deal is subject to customary closing conditions, and we expect the integration of pegozafermin will further drive the strong development momentum in our CVRM pipeline.

And now let's move on to my last slide for today. Here is our key news flow with the latest updates. All of these updates, we have already covered on previous slides, so I won't repeat them, but I will just reiterate that we do have some key readouts that remain for Q4, including fenebrutinib in PPMS.

And with that, I will turn it over to Matt to walk us through Diagnostic slides.

Okay. Thanks, Teresa. Good morning. Good afternoon, everyone. It is my pleasure to present the year-to-date September 2025 Diagnostics Division results. So as you heard from Thomas, Diagnostics is back to growth with sales of CHF 10.3 billion. Sales increased by 1% or CHF 0.1 billion compared with year-to-date September 2024 at a constant exchange rate. This was driven again by the health care pricing reform in China, which include the implementation of Diagnostic-related groups, which affects volume and volume risk procurement, which affects price. As you heard earlier from Alan, excluding China, the growth of the business was plus 7%.

Now let me walk you through the sales driven by each of the customer areas. So first, sales in our core lab decreased by 1%, again driven by the aforementioned austerity in China. Excluding this effect, the core lab business grew by plus 10%. Sales in the molecular lab increased at plus 4%, and this is due to strong growth in our blood screening business at plus 5%. However, this was partially offset by flat sales in the infectious disease segment due to the USA pause in Q1. Sales in the near patient care division decreased by minus 4%. Sales in Pathology Lab grew strongly at plus 13%, mainly driven by advanced staining growth of plus 11 and companion diagnostics growth of plus 21%.

Now shifting to our regional view. I'll take you through the regional business performance. North America, great growth of plus 7%. In EMEA, the business grew at plus 6% and LatAm growth of plus 14%. Now again, in APAC, the business declined at minus 15%. As we previously mentioned, sales growth was impacted by the health care pricing reform in China. As a result of this, China sales have declined by minus 27%. This impact is expected to continue over the remainder of 2025. For 2025, I'd like to confirm our ambition of low single-digit growth. Now looking forward, our ambition, and you heard this from Thomas for the Diagnostics division is always to grow mid- to high single digits. However, given that we anticipate diminished but continuing headwinds in China 2026, we would set our ambition for 2026 as mid-single digits.

So now I'd like to transition to some of the exciting news coming out of our pipeline and really start with some of the exciting updates that reflect our continued investment in innovation, starting with our AXELIOS sequencing solution, which, as you heard, is set to launch next year. So first, I'd like to highlight the results of a recently published correspondence in the New England Journal of Medicine. Here, team is at the Broad Clinical Labs as well as Boston Children's Hospital conducted an analysis using reference and clinical samples from a research cohort to investigate the feasibility of utilizing the sequencing by expansion or SBX technology in urgent critical care settings.

For one of the reference sample, HG002, the team sequence and analyze the whole human genome in less than 4 hours, achieving the world -- the Guinness world record for sequencing. I would like to add that all runs were completed in less than 5 hours from DNA to variant calling. And for the prospective samples, the results were clinically optionable for all infants. These results demonstrate that fast, reproducible and robust genomic results can be achieved in a single day and underscores the potential of our SBX technology in critical care settings where rapid genome sequencing is crucial for decision-making.

And so now I'd like to continue with updates on our AXELIOS sequencing solution. I'd like to highlight some data that was presented last week at the American Society for Human Genetics Conference and focus on 2 specific examples. So as we've said, since we unveiled the AXELIOS technology earlier in the year that we plan to demonstrate its applicability for all key applications. So here, I'll focus on a couple of them.

The first on the left illustrates the suitability of SBX for methylation. So this is an internal workflow we developed utilizing taps, which generates direct conversion of 5-methylcytosine to thymine basis paired with our SBX-Duplex sequencing approach. Now this allows high throughput and simultaneous detection of DNA and methylation variants from the same library. This data presented shows that the performance of SBX for methylation is directly comparable to the standard SBX-Duplex workflow with F1 scores for single nucleotide variants at 99.6 or higher. So this really shows that we're able to do methylation maintaining that high level of accuracy.

The second example on the right is RNA sequencing with our longer reads workflow, where collaborators at the Welcome Sanger Institute analyzed a 90 sample cohort melioidosis highly heterogeneous infectious disease, and they demonstrated that a higher throughput of 4 billion reads per hour and a 400 base pair read length led to better isoform detection compared to traditional short-read workflows. Now when you take these together, these examples reflect our consistent development progress and the potential of SBX technology for both research as well as clinical applications. We're very much looking forward to the launch of this next year.

And so now I'd like to turn to another topic, which is quite exciting, and that's our advancement of our neurology portfolio. Specifically starting with our for Elecsys pTau181 for which we received CE Mark in July and FDA clearance in October. So Alzheimer's disease as well as other dimensions remains a key global burden on health, affecting more than 55 million individuals annually and projected to reach over 80 million people by 2030. Now despite this, 2/3 of people experience cognitive symptoms remain undiagnosed. Diagnosis takes nearly 3 years on average after symptom onset and relies on subjective cognitive tests as well as expensive imaging analysis. Our pTau181 will offer a minimally invasive blood-based solution to roll out Alzheimer's disease. And we'll call out, as Thomas mentioned earlier, our FDA label has an intended use suitable for primary care settings, and we're the only test cleared for this intended use.

So continuing with our neurology portfolio, I'd like to highlight the second example, which is our blood-based -- second blood-based biomarker for Alzheimer's disease, Elecsys pTau217, which is set to launch next year. Now here, I'd like to provide you with an update on some key data that were presented earlier this year at the Alzheimer's Association International Conference, which is the leading conference on Alzheimer's disease. At this event, we showcased both the robustness as well as the accuracy of our tests that's under development.

Now 2 key studies and the main findings on the left, you'll see the robustness of our method, which is a key aspect for a test that's going to be used broadly in a routine clinical setting. On the left-hand side, you can see the results of our own clinical study where we evaluated the impact of preclinical robustness on immunoassay performance by comparing our pTau217 to an internally developed PTL217-AB42 ratio assay. Now this study included patients from 5 cohorts, representing patients across the Alzheimer's disease continuum. And as you can see, the pTau217 isoform by itself demonstrates superior robustness to the ratio test, which is a commonly used approach in other competitive assays. And when you think about implementation in a routine setting, robustness especially preclinical robustness is going to be very important.

Now the second example on the right highlights the accuracy of our test against other available methods as compared to the gold standard of amyloid PET. Here, you see the results generated as part of a leading U.S. integrated health care network. This study encompass nearly 2,000 individuals and the interim results, which you see here, showed the area under the curve for the Elecsys pTau217 achieved the highest accuracy amongst the multiple pTau217 methodologies that they evaluated. So when you take these in aggregate, these study results support broad implementation of Elecsys pTau217 in routine clinical practice, and we look forward to keeping you in the loop as we demonstrate additional data with this exciting biomarker.

So last, I'd like to report progress on our key launch list for the Diagnostics division of our 14 launches that -- which you see here, we've achieved 7 by Q3 2025. We're making good progress on the other launches, and we look forward to updating you at the end of the year. So with that, I'll pass it over to Bruno.

The Q&A session. So the first question goes to Richard Vosser from JPMorgan.

A couple of questions. First of all, on Ocrevus, Perhaps, Teresa, could give us a little bit more color in terms of the logistical challenges, what they are, how easy they are to resolve and when we think subcutaneous will start enhancing the growth of the brand? It looks as though we may be plateauing a little bit in the U.S. in Ocrevus, just what's going on there, please?

Second question, just on the HER2 TKI. We've seen quite a few of these in the past, and I think they always had tolerability, challenges maybe diarrhea in the past. So just what's the tolerability profile and what you're thinking about there? And then I could go further, but I think I'll stop there because -- we'll stop there.

Great. So I think when you think about the logistical challenges for Zunovo, they're not that dissimilar than from what we saw with Phesgo when we were moving Phesgo into the oncology setting. It's about making sure that the workflow in the infusion suite is set up to just handle a different kind of -- a different technology. And so we are well versed in what we need to do because we have done it before, and we are steadily working with our community clinics and our academic centers to actually get them in to the flow of how to work with Zunovo. It is interesting that when you hit that tipping point, and we saw this with Phesgo as well, when you hit that tipping point, clinics go big because they get immediate good feedback from the pharmacist. They get immediately good feedback from the infusion nurses and they get great feedback from the patient, but it does take them a minute sometimes to think about how it is actually going to work in their particular clinic.

Over the course of the last couple of months, we've had now the opportunity to talk to a couple of physicians, who started slow and are now up to 100-plus patients on Zunovo. So I do think we are going to start to see this ball rolling and really picking up steam. I think we are already seeing now. We are already seeing that we are expanding our prescriber base pretty significantly. We're starting to see some competitive indications that would indicate that subcut is starting to get some traction, particularly in accounts that are in the more rural areas, where maybe Briumvi or Kesimpta would have been initially a choice. Now they're really migrating towards beginning to migrate towards Zunovo. So I think this is an area that we have a lot of confidence that we're going to start again to see this ball really picking up steam.

In terms of the HER2 TKI, thus far, it has been very well tolerated without any dose-limiting toxicities at doses of up to 100 mg monotherapy. And it has a promising GI safety profile compared with other HER2 TKIs, which I think is one of the reasons why this was so attractive to us in addition to the fact that it potentially has best-in-class brain penetrant profile. So this is a molecule that we're actually super interested in getting into clinic and seeing how it does.

Yes, it's up to 1,000 milligrams.

Sorry, up to 1,000 milligrams. Sorry, Thomas. Thank you.

Richard, does this answer your questions or any additional questions?

No, perfect.

Okay. Then we move on and next questions go to Simon Baker from Redburn.

Two, if I may, and just continue on Rich's question on Ocrevus. If I look at the trends in the U.S. and I look at the CHF 9 billion of peak sales guidance that you've given and where consensus is, there's a little bit of a disconnect here. And I was just wondering, Teresa, is the disconnect our perceptions of the ex U.S. opportunity here. If one looks at the IV subcut conversions you've seen elsewhere, it feels like we may be underestimating the ex U.S. opportunities. So any thoughts on that would be helpful.

And then just on news flow more generally, it's obviously not unusual this time of the year to see a few '25 events shift into 2026. You've shifted 6 on the latest slide. I just wondered, is there any underlying theme there? Because a lot of these don't seem to be event-driven studies. It's not immediately obvious why they might be delayed. And I was particularly thinking about CT-868 and CT-996. So any thoughts on that would be very helpful.

Sure. So I think with regard to the news flow, some of it is just when data will be presented and the most appropriate mechanism or vehicle for where that should be presented is 2026. I think -- that's the case for 996, 868. I mean, where all of those trials will be placed into ADA. And so that's the reason for the shift.

And then the event-driven studies are really determined -- are really what's determining the shift for giredestrant, which I think you know [indiscernible] shifted for competitive reasons. So there are multiple sort of series of reasons why things have shifted there. Within Ocrevus, if I may, I think we're underestimating the opportunity in both places. So what we have to remember is that Ocrevus IV was very heavily concentrated in academic and urban centers where IV capacity is very prevalent. One of the biggest opportunities that we have in the U.S. is the expanding into community neurology settings and opening up those -- those more rural opportunities, which is where you see a lot of the subcut and oral activity in the MS market.

And that is a huge part of the U.S. market that Zunovo has the opportunity to tap into. And I think that is something that maybe we're not always quantifying in the right way, that there's a big opportunity for us to penetrate in those places. The profile of Ocrevus Zunovo to patients is extremely compelling. To be able to deal with your MS twice a year, 10 minutes, that's a pretty compelling -- that's a pretty compelling value proposition. And so we really think as physicians and practices get more comfortable in the U.S. with how to use Zunovo that we will steadily see additional utilization in those areas.

And then I do also think there is a lot of opportunity in the ex U.S. market as well. Again, for the specific reason that you mentioned is that IV capacity can be more limited in some of those health care systems and for something like Zunovo where you have the opportunity, again, in a very short period of time to help patients treat MS, you will open up a large number of centers that might previously have not been as able to accommodate Ocrevus as an IV therapy. And again, there are some early signals within the competitive environment that, that is actually beginning to happen. And so I'm very encouraged with what we'll see with Ocrevus next year. And a big part of our confidence in that is why we raised the guidance to CHF 9 billion at Pharma Day.

Then we move on. Next one is Michael Leuchten from Jefferies.

Two questions, please, one for Thomas. I just wondered if you could give us an update on what's happening in Washington, we've seen a U.S. company and the U.K. company reach an agreement with Trump. I guess everybody is waiting to see what happens next, just your perspective would be very helpful. And then a question for Teresa. On the HER2 franchise. It looks like in the U.S. this quarter, the combination of the HER2 revenues took a dip. So Phesgo didn't quite pick up the revenue loss seen with Perjeta and Kadcyla, just wondering if you could go into that a little bit more. I appreciate your comments around the tail and the dynamics, but the quarter seemed a little bit soft in the U.S.

Thank you very much, Michael, for the question. On the U.S. discussions with the government there, I can say that we've been in discussions with the U.S. government for most part of this year. So it's not something that has only happened in the last couple of weeks. For example, also the Xofluza direct to patient access topic we discussed also with the U.S. government. So we are in constant exchange with the U.S. government. That's as much as I can say at the moment.

And as far as the HER2 franchise in the U.S., there isn't anything in the underlying dynamics of that business that is particularly -- that stands out in Q3. It's not something that we're worried about at this juncture.

I think Michael, maybe to add from my side. I think we had a very -- if you look at the entire HER2 franchise year-to-date, we had a strong growth. So I think this is more like order patterns quarter-over-quarter. Nothing and no trend change here. All your questions answered?

Yes.

Yes, then let's move on. And next question go to Luisa Hector from Berenberg.

So just on the oral side for Teresa, just any comment on potential positioning of evERA given the changes coming in the first-line setting? And then maybe to push further on the first-line trials and perhaps the adjuvant setting. Could you tell us how your level of confidence has changed ahead of persevERA now that you have evERA in-house? And perhaps remind us a little bit more about the combination of giredestrant with CDK4/6, which kind of plays into both settings, like what data you have that supports that combination being efficacious?

Great. Thanks, Luisa. So in terms of where we would be expecting to position giredestrant based on the evERA data, I mean, I think we would really be positioning it as standard of care in this patient population. The data certainly support that. And I think we have always had confidence that giredestrant was a targeted potent combinable tolerable SERD that actually had the opportunity to potentially redefine a new backbone therapy in this area. And I think what evERA has done is sort of reinforce our confidence in the molecule to say that this is a highly active molecule that really does have the potential to help a really significant number of patients.

When you think about combination with everolimus, I mean it could really address the resistance to endocrine therapies because it targets different signaling pathways. While it minimizes the impact of treatment from patients because you don't have to have as many injections. It's sort of an all-oral combination. I think it's also worth pointing out that the safety profile of giredestrant is really playing out incredibly well, particularly given that there's -- we didn't see any ocular talks.

So I had the privilege to be at ECTRIMS -- I'm sorry, ESMO to spend some time there over the weekend and really talk to physicians. And I would say that sort of everybody that we talked to really felt like the evERA data was going to be practice changing for them in that setting. Now what does that mean for the other trials? I mean I think we do need to be careful about cross-trial read-throughs. We know that each one of those trials in the first line and in the adjuvant setting are all designed to answer a slightly different scientific question. But I think what we believe to be true is that anywhere where ER signaling remains important to a patient's disease, we have the ability to bring efficacy on top.

And so I think we're overall, more encouraged. These are still relatively high-risk trials because of the -- just the nature of the disease. But I think we're very encouraged by what we've seen with evERA. We've always had a lot of confidence in this molecule as a stand-alone molecule. And now we get to wait to see how the trials play out.

Luisa, does this answer your questions?

Yes. Thank you.

Then we go on. James Quigley from Goldman Sachs.

I've got 2, please. So first of all, on the 2025 revenue guidance, again picking up on a comment you made, Thomas, that 7% is included in mid-single digits and there's a little bit of a concern this morning that mid-single digit means 4, 5, 6, which would suggest that growth in the fourth quarter was 0% to 4%, a good step down from what we've seen for the previous 3 quarters this year. So can you clarify that in terms of what we should be expecting on revenue growth for the fourth quarter, asking for 2 reasons. First of all, if -- with the operating profit growth -- sorry, the core EPS guidance upgrade, if revenue falls off in the fourth quarter, then maybe it's more cost savings related. And then also second reason is the exit rate into 2026.

And then second one on Vabysmo. How much visibility do you have in terms of level of funding for the foundations? Is it getting back towards pre-drop levels this year? And therefore, what is your expectation for the rebound in growth as we see in 2026? I'm asking because, obviously, last quarter, we had the guidance for 20% growth. We stepped back down to 15% growth this year. So again, it seems like there's a bit of variability in the visibility?

Thank you very much for the question. So I'm not at all concerned about Q4 growth. When I said that 7% is included in mid-single-digit range. That's exactly what I meant with the 7%. Regarding core EPS, you saw core EPS growth at half year. This is, as you see that we are doing quite well on the sales growth. At the same time, we've been very good in terms of cost control. We've always committed to keeping R&D flat. And with that, yes, we decided to increase our guidance on core EPS. So it's not additional cost savings. It's really what we're seeing at the moment in our P&L that we believe that will translate into the end of the year.

Great. And when it comes to the CAF's, I think, obviously, we don't have clear visibility into what will happen with the co-pay foundation funds because I think as we've always that needs to happen at arm's length to the business. That having been said, as we head into 2026, we would expect a gradual normalization of CAF funding in the future. And likely what we would expect is that multiple CAF's will step in to fill the gap that's been created here. That may provide some logistical challenges for offices that they just need to get used to sort of working with multiple CAFs and not just one. But I think ultimately, we will see funding return into this disease area and that will help normalize and sort of steady the market in this particular place. I think as we think about next year, we do expect recovery in the U.S. and a return to strong growth for Vabysmo.

James, did this clarify the outlook?

Yes, it does.

Okay. Then next question go to Sachin Jain from Bank of America.

Just a follow-on question to some of the topics already touched on. So firstly, on the Vabysmo lower guide. If you could just clarify what didn't happen in 2H that you thought was going to happen because it seems like the dynamics are unchanged. So just trying to clarify what's not played through? And to clarify the answer to the last question, is that an acceleration of the growth trends in '26 relative to '25 that we should expect?

The second question is just if you would touch on key pushes and pulls into 2026. Slide 8, Thomas, that you sort of talked about 7% to 8% sales growth for the last few years, is that sustainable? And I guess the question really homes back in on beyond the Vabysmo, Ocrevus and Hemlibra have seen sequential declines in the U.S. which is what's driving the concern for investors today?

And then last clarification on the U.S. administration question was already asked. Just I wonder if you were able to comment what's your ability to do a deal similar to those that have been announced, is there any prohibition from your business mix that stops you from doing that?

Thanks for the question. So I'll take your questions in the opposite order. So I'll first take your question on the discussions with the U.S. administration. I really cannot comment more, except that we've always been in exchange with the U.S. government. And for example, Xofluza DTP was something that we also discussed with the U.S. government. So I really wouldn't want to go into additional details here.

Regarding 2026, I think we gave you a bit of an outlook now into Q4 a bit more concretely than we'd normally do it because there was a concrete question on that. When it comes to 2026, we will update you at full year. What we can say is we will have a continued good momentum. And so we feel comfortable that we will also be showing good growth next year.

Okay. And so Sachin, I think your first question was it on granularity of what's going to happen with Ocrevus in 2026? Or would it -- what happened and what did we think was going to happen with Vabysmo in Q3...

Apologies, It was Vabysmo lowered guide. What hasn't happened that you thought was going to happen?

Got it. Okay. I'm sorry. You broke up a little bit, and I just didn't catch the question. So I think, ultimately, we had hoped that we would see a return of the branded market, and that just didn't happen. So it was -- I think it was more just the underlying dynamics that we saw in the first half continued into the second half, and that just didn't normalize. So hence, our revised outlook.

Because of the dramatic change that we saw to the branded market in 2025, I don't think that market shares in 2025 and 2026 are going to be directly comparable. But I think as we start to see a resetting of that baseline, you will see growth rates sort of continue to bounce back to sort of more of what we -- more of what we would have expected. The other thing I think it's important to point out is that in Vabysmo outside of the U.S., we are now fully launching our prefilled syringe. And so we expect to see a nice uptake ex U.S. of the prefilled syringe.

So I think there will be 2 dynamics at play in 2026 for Vabysmo. One is a normalization and sort of a resetting of the baseline in the U.S. market, which will allow some of those underlying growth dynamics to play out and be more visible, but then also the -- you'll start to see the benefit of the uptick in the prefilled syringe launch in ex U.S.

Sachin, If I understood you right, you also had a bit of question about the other 2 growth drivers, Ocrevus and Hemlibra looking forward into next year?

Yes, if you wanted to touch on it, just given the sequential declines in the U.S. seen in the third quarter?

Yes. So I mean we think we just touched on Vabysmo. I think as far as Ocrevus goes, it is going to be about breaking into new areas for Ocrevus Zunovo. I think you've always heard us talk about when it comes to Ocrevus, there are sort of 2 things that we were interested in doing, obviously, converting some of our IB business, but we were never going to reach the goals that we have for Ocrevus if all we were doing is converting.

So yes, it's important to convert, but it's also extremely important that we open up new areas for Ocrevus particularly in the U.S. in those community settings. And again, when you look at the underlying dynamics the number of health care providers that are prescribing Zunovo is increasing. We're seeing new prescribers come in, and there is that tipping point that once you do it 2, 3, 4, 5x, all of a sudden, you really get a tipping point into those practices. So I think it's both -- it's both the conversion and then the opening up of new spaces that are going to be important for Ocrevus next year.

I mean it's definitely in the plan that we should see a pickup into next year as we also have -- have it in the plan for Vabysmo.

Yes, absolutely.

Okay. Very good. Sachin, did this answer your questions?

Yes, perfect.

Yes. Then we move on. Next one would be Matthew Weston from UBS.

Apologies. It seems that technology is extremely slow. The unmute box just popped up. Two questions from me, please. The first on Hemlibra. First of all, the sharp slowdown in Q3 over Q2. I think that was stocking last quarter that then looks like it's unwound. But can you please confirm that? And also, it is a product that's meaningfully outperformed consensus expectations over the course of '25. So what's a realistic expectation for this franchise continue to grow next year? Should we just see a continuation or was there for some reason, a bolus of demand in 2025?

And then the second question is around vamikibart. You're obviously excited about the data you've made that clear today. You held an investor event that was solely focused on it or 50% focused on it, but it seems to be in your pack as a CHF 500 million to CHF 1 billion peak sales estimate, which is a very modest product, quite frankly, for Roche total sales. So is there something differentiated about vamikibart over time that may mean that it can get bigger than that? Or it's just an asset that you want to flag because of the innovation?

Great. Great questions. Thank you. So for Hemlibra, you hit the nail right on the head. In Q2, we did have a big buy-in that hit in Q2 versus Q3. And so that is really the reason that you see that disparity in quarter-over-quarter, it's purely a buying pattern balancing effect. And you're right, we have seen very good growth for Hemlibra in 2025, and that really has been due to increased penetration into the non-inhibitor population globally, something we've always said we wanted to do moving into that more moderate patient base. We are also seeing patients return from [ ALTUVIIIO ], which is fantastic to see. But we are sort of tempering our expectations for next year given the significant growth that we have seen in 2025. And for 2026, we're really looking at low single-digit growth for Hemlibra is sort of what our outlook would be there.

In terms of vamikibart, I think the reason we're excited about this is that UME is -- while it is a relatively small patient population, these are younger patients, who are getting very high-dose steroids, which is just not ideal from a safety perspective over time. And so the idea of a therapy that can actually meaningfully improve that patient outcome is very compelling from a clinical perspective. And it also fits very neatly in with the commercialization of our other products. So it's not like it's a big cost to commercialize this molecule. So I think for us, we're very excited about the science and the innovation that it provides and really, frankly, what it can do for patients from a safety perspective.

We also think that adoption here is going to be a relatively fast thing to drive. Is it just a simple -- it's just a simple injection and it prevents vision loss. And again, a much younger patient population. So very significant unmet need fits very neatly into our commercialization, great science. And yes, I mean we're sort of projecting at around CHF 500 million. But oftentimes, when you have something like this, you don't totally know what you have until you actually get it into the market and then maybe we'll be pleasantly surprised.

Maybe wanted to add on here. I think, in general, we were excited about IL-6 in ophthalmology. I think we also share that there's an opportunity in DME where we have a bispecific which will move ahead. So I think it's just the first 2 cases, where we clearly have proven that IL-6 is key to a couple of diseases.

Yes. Great add, Bruno. Thank you.

Okay. Then we go on. Next one is Peter Verdult from BNP Paribas.

Pete Verdult with BNP. Maybe give Teresa, a couple of minutes rest. Just a quick one to Matt, on China Diagnostics. I think we're all aware of the pricing dynamics, but it feels like volumes have also come into the equation in terms of volume decline. I mean in terms of the visibility you have, Matt, in terms of that stabilizing return to growth, anything you can share with us would be helpful.

And then Teresa back to SERD, the market has been quick to write off the chances for the class in first-line adjuvant despite the data you shared. That was not a view shared by the community at ESMO over the weekend, which we attended. So if that holds, the only thing to discuss really is the GI tox profile because that's something that's been put to us that puts giredestrant at a disadvantage to other competitors. So anything you can talk about that as it relates to the evERA data or your comfort with the GI tox profile of giredestrant?

All right. Sure. So thanks for the question. And this is something that I've addressed since the beginning. So we saw a couple of effects going on in China for the last year. One of them was obviously the volume-based procurement and the reimbursement cuts. And the other was the implementation of diagnostic-related group audits, where they look at appropriate use of testing. The DRG effect is really on volume and the other on price. And we've seen both of those pull through the system over the last year, and we expect, again, the peak is going to be 2025, continued negative impact sales in 2026, but to a lesser extent than this year.

And as I said, for this year, our ambition in Diagnostics, low single digit. For next year, we expect to grow -- our ambition is mid-single digit. And I want to be clear on that. It's a very dynamic situation. I don't think I can be more precise than that.

So Peter, thanks for the break. Appreciate it. So in terms of SERD, I think we are very comfortable with the safety profile that we have seen over multiple studies now with giredestrant. And I think in talking with physicians, they feel like -- what they've seen is also very manageable and in line with what they might expect. The first-line setting is going to be an interesting one. I think you're right, the investment community has been pretty pessimistic on the earlier line studies. I think that we will need to wait to see what the actual trials read out as. We're very confident in giredestrant as a molecule. We're very confident that anywhere we see the year pathway remaining important to people's disease that we can add efficacy on top and we'll just need to see how the trials read out.

One thing that I might sort of invite people to consider as you're looking at your models, is that right now, most people have about-ish [ CHF 800 million ] in for giredestrant with what we think we are likely to get with evERA, you could argue that we're most of the way there with giredestrant in the model. And that anything that comes on top, whether it's first-line or adjuvant is sort of all gravy. And so I think this is a space to watch. Again, I think we have a lot of confidence in giredestrant as a molecule.

The trials for all the reasons we've mentioned sort of remained a little bit of coin toss. But if we do hit, this is going to be an incredibly significant drug and one that is going to help many, many patients. And again, based on the data so far, we're really not seeing any dose-limiting gastrotoxicities, and the profile has seemed to be very well tolerated.

And I just want to underline that, that's basically, what you have in the models, that's the current results from evERA. So anything now is going to be on top. So I think you have more upside than downside when you look at giredestrant. But again, we don't know what the results are going to be. We'll see. But definitely, I would say, from a model perspective, it's definitely upside.

Peter, you are done?

Yes.

Then next questions come from Sarita Kapila from Morgan Stanley.

Sarita from Morgan Stanley. Just a follow-up on Xolair, please. I know we touched on '26, but how should we think about it in the context of remibrutinib competition in CSU or there's been quite a positive reception from the KOL community. And then the second one on evERA and read to persevERA, given the limited effect size on PFS and wild-type patients, are you confident that you've enrolled enough patients into persevERA to hit stat significant? I believe 990 patients and your competitor has roughly 400 more patients. So does that put them at a greater chance of success?

Great. So while a nonfood allergy question on Xolair, nice job. So I think with CSU, we do continue to see relatively good utilization with CSU. Certainly, as with asthma CSU is becoming an increasingly competitive space, but the data on Xolair are very good. People are comfortable with it. It's a fairly entrenched option. And so I think we're not expecting to see a massive erosion of that quickly for what it's worth. And then for evERA and persevERA, I mean I think we are -- we have enriched for the ESR mutations for about 40% in that trial.

I'm sorry, that's pioneer you're talking about persevERA. PersevERA, I am hopeful that we have the right patient population in there. I think we feel like we've designed that trial well. We've designed it based on what we believe we will need from a patient size in order to see -- in order to see what we -- we've powered it appropriately to be able to see what we believe the effect size is. This is an all-comers population. In this setting, as our mutations make up less than 10% of the population. So we'll see. This is -- these are the trials where, honestly, it just gets a little more difficult for everybody.

Maybe just to add here, Sarita, we have 2 studies in first line. One is basically the endocrine therapy sensitive study, which is coming out first, and this is all-comer study. This, of course, has less ESR1 mutant patients in and therefore, tests the all-comer hypothesis, how broad can we go? And then we have the 40% of the first line, which is defined as endocrine-resistant patients. And there, of course, you have like an accumulation of 25%, 30% of ESR1 mutant patients, and we are further enriched. This is to come in '27 only. So -- but we have split the first line into these 2 buckets, and we are basically testing -- and ESR1, we know it works, but the big question is will we see the benefit of the all-comers?

Exactly. And I think the clinical trial program has actually been designed quite robustly to make sure that we can really answer the very specific patient population questions.

Sarita, did this answer your questions or any additional questions?

Yes. Cool.

Okay. The next question come from Stephen Scala from Cowen.

I have one observation and then 2 clarification questions. But given the passage of time and the lack of clarity, I guess, we have to consider a possibility where no pricing deal with Trump is signed. Thomas, is it acceptable that we have that thought in mind. The clarification question, just to be clear, the LOE exposure is expected to be more severe in Q4 versus the cadence through 9 months, what products are causing that acceleration? And then lastly, to be crystal clear, will Roche file evERA for all-comers?

So maybe, Thomas, I'll do you and answer the last one first. So I won't comment on what our filing strategy is for evERA. I'll leave that for you guys to ponder. And then you are correct that the pace picks up in Q4, and that is largely due to Actemra. We haven't seen much biosimilar impact in the beginning of the year. We are starting to see the kick in the U.S. I think you've heard me mention that we had about 6% in the U.S. -- in Q3 alone, and we would expect that to accelerate as we go through the end of the year.

Yes. On the U.S. topic, I mean, I think I said everything that I could. We are in active discussion with the U.S. government, and we've done that even prior to the last couple of weeks. So -- and one of the things we've discussed with the U.S. government, for example, is the DTP program with Xofluza. So again, we're in discussion with the U.S. government.

Steve? Then we move on. Next question go to Yihan Li from Barclays.

Yihan Li from Barclays. I guess, like 2 questions from my end. The first one on Vabysmo. Thank you very much for the clarification and also the commentary. And also, it seems like we are going to see some reverse likely from the beginning of next year. But I just like wanted to further clarify on your expectation for the fourth quarter because it seems like you now upgraded the guidance for 15% year-over-year growth at CER, which indicates the U.S. growth will likely at high single-digit range. So just curious like what underpins your confidence for this Vabysmo growth in the U.S. for the fourth quarter?

And then the second question is actually on your [indiscernible] GYM 329. So we noticed you have like 2 Phase II readouts SMA and also the FSHD readouts pushed into 2026. So just curious like what is the underlying reason? And also, we know you have already had your Phase I dose finding data obesity in-house for some time. So just curious, like, did you observe any similar profile that your competitors showed at ADA? Yes, like any commentary there. And also in the Phase II data, are we going to expect it at ADA next year as well?

Yes. So, emugrobart, I think I mentioned, is shifting into 2026 for competitive reasons, that data will be shared next year, including data at ADA and obesity. And for Vabysmo, in terms of Q4, I mean, I think we are expecting to see it continue to perform in the way that a new standard of care does. I mean we're -- the growth rates are still in mid-single -- mid double digits. We're still -- it's planning to grow at 15%. We still do here from retinal specialists around the world that is the new standard of care. It is their go-to drug for new patients. And so I think we continue to believe very strongly in the profile of Vabysmo, what Vabysmo can do for patients. And as the U.S. market corrects itself, we would expect strong growth next year.

Okay. I think with that, we are at the end of our Q&A session. Thomas, over to you.

Thank you very much, Bruno, and thanks for everyone attending today's call. As you've seen, our sales continue to be strong at 7%. Pharma continues to perform well at 9% and Diagnostics is starting to recover. So clearly, I would say we're delivering on the sales front, but not only in the sales front, what we've also said is that we have good cost control so that we are raising our guidance when it comes to earnings expectations. I think everyone is very interested in the readouts that are going to come shortly on fenebrutinib and giredestrant. All I would say, good opportunities. We're clearly -- if you look at the models, there's probably more upside than downside at this point in time. But we will have to see what the data ultimately says.

But really looking in the pipeline more general, we have now moved 10 molecules into Phase III. So we really rebuilt and shaped our late-stage pipeline. We've up to [ 19 ] medicines that can launch by the end of the decade. So I think we've done really tremendous step forward in terms of our pipeline. So I'm quite confident in the long-term outlook for our company, but also you can see that our on-market portfolio is performing.

On the Diagnostics side, we have seen the major launches that we've had each a blockbuster on its own. If you take Pharma in terms of revenue expectations. So -- and I do believe that we can continue to take a good momentum into next year and that we will continue to deliver. Thank you very much.

My name is Henrik, and I'm the technical operator for today's call. Kindly note that the webinar is being recorded. [Operator Instructions] One last remark. If you would like to follow the presented slides on your end as well, please feel free to go to roche.com/investors to download the presentation.

At this time, it's my pleasure to introduce you to Bruno Eschli, Head of Investor Relations. Bruno, the stage is yours.

Thanks, Henrik. Could I have the first slide, please? Thank you. So welcome to our seventh IR event for this year, where we will focus on the latest clinical results from our ophthalmology portfolio, which got presented last week at ASOPRS and AAO.

And let me quickly take you through today's agenda. The event will go for 60 minutes, I assume around 40 minutes for the presentation, and then we have 20 minutes for Q&A. If we go top down, then our first presentation will be given by Nilesh Mehta, our franchise head, ophthalmology and global product strategy. Nilesh will provide a quick introduction to our IL-6 development programs in ophthalmology.

Thereafter, our second speaker then will be Dr. Nisha Acharya. Dr. Nisha Acharya is an Elizabeth C. Proctor Distinguished Professor of Ophthalmology, Epidemiology & Biostatistics at the University of California San Francisco and Director of the Uveitis and Ocular Inflammatory Disease Service at the Proctor Foundation.

She's also Associate Director of the Proctor Foundation and Vice Chair for Faculty Development and Mentorship of the Department of Ophthalmology. She is a clinician scientist and an international leader in ocular inflammatory diseases for over 20 years. She has led an NIH funded research program conducting global clinical trials and outcomes research. Dr. Acharya will take us today through the Phase III SANDCAT/MEERKAT results for vamikibart [UME], which represented last Friday at AAO.

And then our third speaker for today will be Cesar Briceno, who is an Assistant Professor for Ophthalmology, Dermatology and Surgery at the University of Pennsylvania and Chief of the Division of Oculoplastic Surgery. He's also the Co-Director of the thyroid eye disease program at the Scheie Eye Institute and was site principal investigator for the SatraGO-1 and 2 trials which we'll discuss today.

In addition, he has served as Assistant Dean for Student Social and Professional Networks at the Paramount School of Medicine at the University of Pennsylvania. Dr. Briceno will take us today through the Phase III SatraGO-1 and 2 results for satralizumab and TED, which he presented at ASOPRS, I think last Thursday.

In addition to our speakers, we will be joined then for the Q&A by Chris Brittain, our Senior Vice President and Global Head of Product Development and Ophthalmology, which many of you had the opportunity to meet at our Pharma Day in London last September. Could I have the next slide, please?

Before we get started, just quickly, 2 slides from my side to update on the upcoming news flow for the remainder of the year, and then also providing an outlook into the news flow for '26. As you can see here on our updated 2025 key news flow slide, we have 3 more readouts to come. That's Gazyva in SLE, [indiscernible], the fenebrutinib data in PPMS. For several Phase II assets, we will have the data in-house. However, it has been decided in the meantime, primarily for competitive reasons, to share these results only at conferences in H1 '26. So these studies include the [emugrobart] data in SMA and FSDH and the CT-868 data in type 1 diabetes and the CT-996 data in type 2 diabetes.

For a summary of the 2026 news flow, let's quickly have a look at the next slide. That's the updated '26 news flow. Overall, we expect now, 12 pivotal Phase III readouts in 2026 and several important Phase II readouts, which will inform the next development steps of our CVRM portfolio. The current planning is that most of these Phase II results in CVRM will get presented at ADA, which is in June next year.

So with that, let me hand over then to Nilesh for a quick introduction to our 6 development approaches in ophthalmology. Nilesh, please.

Thank you, everybody. I hope you can hear me. It's a great pleasure to be with you all. And from my side, I'd like to just set a little bit of the context in which you can hear the other 2 data sets that will be presented by our experts today.

So moving to the next slide. Just grounding us on some bigger picture and vision for ophthalmology, with the pun intended there. These data sets are relevant for our ambition. And on the left, you can see that we aim to establish ourselves in a leading part in ophthalmology by clearly building a beachhead in retina. And I think we've made significant progress there.

To do that, we need a really differentiated pipeline of solutions for ones that are innovative enough to be transformational. And then we'd like to expand into areas where unmet need and science boundaries are intersecting. Today's focus, though, is on 2 areas in UME and thyroid eye disease, where we know there's a significant patient need, and you'll hear the experts share the data in depth. On the bottom of the slide, we talked a little bit about how we would approach that. And you see 2 elements in the middle there.

Clearly, leveraging our commercial portfolio today, which includes Vabysmo and Susvimo, but then 2 areas where we want to go beyond VEGF and sustained outcomes in retinal vascular diseases, but also beyond complement and geographic atrophy in this next period. We'd like to pioneer innovation in vision restoration. And then obviously, underpinning a lot of that is build leading capabilities that sustain that innovation.

On the next slide, a little bit conceptually what we're looking at. This ischemia shows how retinal conditions progress. On the left, we know there are a lot of anatomic changes that are suggestive of disease progression in early stages of disease. But many times, they're asymptomatic until they advance to a stage to edema, most normally and therefore, subsequent central vision defects. Today, that's typically where we -- where the current solutions where we intervene, we play a part in improving visual acuity, but we know patients still progress to permanent vision loss. And actually, we want to be able to address both directions here, going earlier to prevent disease and preserve vision for longer as well as being able to restore vision in later stages of disease.

I showed on the next slide, a little bit of the pipeline that you're quite familiar with and that Roche has really built over the last few years. And of particular interest today are the Phase III data from uveitic macular edema with vamikibart and with satralizumab in thyroid eye disease. To note also, and I'll talk a little bit about this in a subsequent slide, is the VEGFxIL6 DutaFab which is a bispecific antibody that we declared recently.

If we move to the next slide. IL-6 targeting vamikibart as a molecule targeting this pathway, which is significantly involved in inflammation. And we have data in-house in both Phase II in diabetic macular edema, both combined with VEGF as well as in monotherapy. And today, you'll hear a deeper dive into the Phase III data from uveitic macular edema, an indication which is strongly associated with inflammation.

If we move to the next slide. As I mentioned, our next-generation bispecific antibody targeting both VEGF and IL-6 simultaneously, one injection as part of the DutaFab platform. This makes a potential option for also the Port Delivery system. We've seen encouraging data from the Phase II BARDENA study on functional improvements that go beyond VEGF alone. And that gives us a lot of conviction to progress this bispecific into diabetic macular edema. We have an already running Phase I study ongoing there.

And finally, on the next slide, satralizumab. This is subcutaneous anti-IL-6. Again, we know inflammation is a hallmark of this disease, particularly suitable for [4W home] administration in an area of high unmet need. There's a wealth of evidence, both preclinically and as shown today, suggesting the role of IL-6 in the pathology of thyroid eye disease, and you learn more from the data from our other experts. So we're particularly encouraged as we build further innovation in the ophthalmology space.

And with that, I'll hand over to Nisha.

Great. Good morning, everyone. It is my pleasure to be here and share with you the Phase III results of 2 trials for vamikibart with uveitic macular edema. As was mentioned, I am a uveitis specialist, so I'm an ophthalmologist who takes care of patients with uveitis and other ocular inflammatory disease conditions. I've been doing this for a while now, a couple of decades. And so I think I have a very -- I see these patients all the time, and I think I'm able to kind of put that into context when I see the results. We can go on to the next slide.

All right. So as mentioned, there are 2 Phase III trials, MEERKAT and SANDCAT, and we'll be presenting the results of that. So just as a bit of background, uveitis is inflammation of the uveal tract of the eye, which is the pigmented vascular layer of the eye, the iris, the choroid, the ciliary body. It's the middle layer of the eye. And when that blood ocular barrier is lost, you basically have inflammation in the eye. Normally, the eye does not have its sort of an immune privilege site. But when you have uveitis, that immune tolerance or immune privilege is lost and there's inflammation in the eye and many different structures in the eye can get inflamed.

So this is an immune-mediated process. Most cases are noninfectious. It's not known exactly what the causes are. It's probably just like any other immune-mediated disease, there's many different pathways, but it's autoimmune and pathogenesis. Unlike a lot of other conditions affecting the eye, uveitis affects patients of all age ranges and primarily the working age population, so -- and can be chronic and recurrent. And so there can be significant implications in terms of that population's trajectory and their life experience and their working ability. The prevalence of uveitis, there's different estimates, but the best we know from various studies is about 121 per 100,000 individuals, adults. We do think it accounts for about 10% to 20% of blindness in the U.S., that may be an underestimate and higher in the U.S. and in the European Union, but probably higher in the rest of the world.

All the photos on the bottom are my patients. These are photos of their eye showing what we see in the eye when we have different types of uveitis. These are all non-infectious uveitis. You can see this inflammation from the front of the eye all the way to the back of the eye in the retina. And on the very far right, blood vessels and even macular edema, you can see on the fluorescent angiogram in the center there, there's leakage.

We can go on to the next slide. Uveitic macular edema is the leading cause of vision loss in patients with uveitis. You see an OCT image there showing what macular edema looks like when we're able to get this noninvasive, basically cross-sectional really view of what's happening in the retina, you can see that the anatomy is really distorted there. So what happens in uveitic macular edema is you get thickening because of fluid accumulation in the macula, which is, of course, the center part of the retina where you have the highest density of photoreceptors responsible for your central vision. So very important part of the retina.

We -- estimates are about 1/3 or maybe even a little bit higher percentage of uveitis patients will develop uveitic macular edema at some point in their uveitis course. Currently, corticosteroids or steroids of some form are the mainstay treatment for us to treat patients who have this. But unfortunately, steroids have very predictable and very common side effects, and those include cataract formation. And then also, I would say, even more worrisome is intraocular pressure elevation and glaucoma because that's not reversible when you have damage. So glaucoma is where you have high pressure in the eye, which steroids can cause and then it basically causes nerve fiber layer damage and optic nerve damage, permanent damage, which results in loss of the visual field and then eventually, it can result in full blindness. So uveitic macular edema is actually one of the more difficult situations that we face. It's common in uveitis, but it tends to be persistent or recurrent or chronic with our current therapies.

And we've had multiple studies now, NIH-funded trials and other studies that have shown that when you're on systemic immunomodulatory or anti-inflammatory therapy, which is what we use to often control uveitis, that uveitic macular edema is often persistent in about 40% of eyes despite using other therapies for uveitis.

We can go to the next slide. So currently, as far as treatments, I just -- as I said, it's steroid. It's steroid of some form. We do have steroid drops. They don't penetrate that well into the back of the eye, so limited efficacy there. Then, we have intravitreal steroid injections, implants and also steroid injections around the eye or in the suprachoroidal region of the eye. So these are just basically administering steroids to the eye. We don't have any FDA-approved non-steroid injections or implants. We do have -- there are anti-VEGF agents or injections, which have been used off-label for uveitic macular edema. We recently had an NIH-funded trial comparing steroids to anti-VEGF strategies and it had very limited efficacy. And the thought is that in uveitic macular edema, VEGF maybe plays a little bit role in permeability, but does not address the inflammatory basis of uveitic macular edema. So unfortunately, it's not been that helpful.

So our current strategy is we do use steroids. Generally, it's going to be injections or implants. Those are usually relatively short acting. They do have a lot of side effects. And we do try to use other agents. We use immunosuppressive therapy, both conventional and biologic therapy. But with this approach, the majority of patients still have recurrent or persistent macular edema and need additional local therapy. So there is a big unmet need for an effective and safe nonsteroid intravitreal therapy for uveitic macular edema. We can go on.

We do know from multiple studies now in uveitis, multiple different labs, different studies that IL-6, interleukin-6 is very important in the pathogenesis of uveitic macular edema. It's a pro-inflammatory cytokine. And when it's dysregulated, there's a number of different downstream inflammatory pathways that are activated. You can see on the slide that this results in things like vascular leakage, chronic inflammation, autoimmunity because there's a -- the T cells are kind of targeting more towards Th17. And you also have just increased vascular permeability because of blood retinal barrier disruption and you have leukocyte infiltration in the eye.

So a number of studies have shown that IL-6 is elevated in the aqueous humor of patients who have uveitic macular edema. And thus, there is strong scientific rationale to target IL-6. Go on to the next slide. So vamikibart is the first and only anti-IL-6 monoclonal antibody that's engineered for intravitreal therapy delivery. It's a recombinant humanized IgG2 monoclonal antibody. The Fc segment of the antibody has been modified. So there's very fast systemic clearance, which is good. And the half-life after intravitreal dosing is about 7.5 days.

We do know from prior studies that with administration of vamikibart, there's a very rapid and sustained suppression of aqueous humor IL-6. You can see on the graph on the right, and this is from an earlier DOVETAIL study. So the Phase I and earlier studies that there's injections at the beginning here 0.4 in 8 weeks, and you have a very marked suppression of IL-6 that then looks to be sustained with 2 different doses that were used of vamikibart.

Next slide. So MEERKAT and SANDCAT are 2 identical Phase III trials for uveitic macular edema. Patients had to be at least 18 years of age. They had to have non-infectious uveitis and uveitic macular edema. We define that with the central subfield thickness of 325 microns on an OCT scan, and they had to have vision about 19 to 73 letters. So not perfect vision and not vision that was already blind basically. This is a good range of about 20/40 to 20/200. And the trial was designed with 3 different treatment groups. You have the vamikibart higher dose, 1 milligram, vamikibart 0.25 milligram and then the sham treatment group. Patients got 4 injections 4 weeks apart. You can see there in the schematic. And then the primary endpoint was assessed at 16 weeks.

And the primary endpoint was the proportion of patients who gained at least 15 letters, so 15 letters or more of best corrected visual acuity from their baseline vision. The trial does is continuing after that, and there's as-needed dosing of treatment and vamikibart after the primary endpoint. And so the end of the study is 52 weeks. Just other things to note is that rescue treatment was allowed from 4 weeks. And patients who have uveitis are often on systemic immunosuppressive therapy or they've had prior treatments. There were criteria to allow patients who've had these to be included in these trials with washout periods or -- and if they had stable immunosuppressive therapy, that was allowed as long as it was stable and then continued in the trial.

We can go on to the next slide. Here are the baseline demographics for the treatment groups for the 2 Phase III trials. You can see that they are well balanced overall. The age was about -- it's in the 50s, maybe a little bit higher age in the SANDCAT trials, but in the 50s is the main age. The majority of patients were female. We tend to see that in uveitis in various immune-mediated conditions in the body, including uveitis. Patients were predominantly white and then also Asian. And then patients were enrolled in Asia, North America and the rest of the world. So a good generalizable population here.

Go to the next slide. So these are ocular characteristics at baseline, and they were very balanced across treatment groups. Vision was about the same in all of the treatment groups. What I want to point out is that the central subfield thickness is very high here. It's in the 500, so 500/20s, 500/30s. That's -- we just remember that 325 was the necessary central subfield thickness, but keep this in mind when you look at the results. So this is very typical that patients have pretty substantial macular edema in uveitis. We did have about close to half, a little bit less were pseudophakic, meaning they've had cataract surgery. So it's very common in uveitis that 50-year-olds would have already had cataract surgery because of inflammation and all the prior treatments they've had. And about 20% to 30% had a problem with pressure in the past, too.

Most patients had uveitis that was affecting the back of the eye and most patients had controlled uveitis. Remember, they could be on other therapy, but about 30% to 40% still had active inflammation and about 30% to 40% were on systemic immunosuppression. So here is the primary endpoint at 16 weeks for both trials, and we'll start with the MEERKAT trial. So here, you see the proportion of patients who gained at least 15 letters, 15 letters of best corrected vision. It's about 3 lines on an eye chart and between the different treatment groups. So 6.3% of the patients who were in the sham group achieved that versus 26.1% in the vamikibart 0.25 milligram group and 43% in the 1-milligram group.

And then statistically, when you look at the difference compared to sham, each of those was highly statistically significant. When you look at the treatment difference, it was about 20% and about 37% difference. Those are very statistically significant differences. In the SANDCAT trial, 13 patients or 13.3% of patients did gain at least 15 letters. In the 0.25 milligram group, about 34 -- or exactly 34% of patients achieved that endpoint. And in the 1-milligram group, 24.2% of patients achieved that endpoint. So when you look at the treatment differences versus sham, this results in about a 20.7% treatment difference when you compare the low-dose group compared to sham and about 10.9% when you compare the 1-milligram group to sham. So the p-values here, it was highly -- so the way the trial was designed was that the primary analysis was prespecified to be the 1 milligram against sham. You can see the p-value here was 0.069. And so that did not reach statistical significance over sham.

And so because of that, you can't really have a definitive call on the lower dose, but you can say that there's a nominal significant p-value of 0.0019 when you look at the 0.25 milligram group compared to sham. We can go on to the next group or the next slide, okay. So this is showing the actual change in visual acuity, best corrected visual acuity from baseline for all 3 treatment groups in both trials. And for both trials, there was a very comparable, very robust improvement in visual acuity. In the MEERKAT trial, sham was 3.5 letters and then both treatment groups, you have 9.6 letters in the lower dose and 12.8 letters in the higher dose. In the SANDCAT trial, they sort of reversed a little bit, but these are very -- pretty similar when it comes to letters, 9 and almost 12 letter gain compared to 5 letters in the sham group.

So pretty consistent results across trials in terms of vision. Can go on to the next slide. Here, we see the reduction in central subfield thickness, that's CST measured in the -- by OCT. And this is really a marker of anatomical of what's happening to the anatomy of what's actually happening to the macular edema. Remember that we started very high in the 500 range. And you can see for both trials, sham was in the 40 -- well, for MEERKAT, there was a reduction of 58 microns. And by the way, about 50 microns is within the range of 25 to 50. So really not much reduction there with sham. But in both treatment groups, very similar and pretty rapid and comparable reductions close to 200 microns, so about 190 to 196 micron reduction, and each of those was statistically significant compared to sham. And in SANDCAT, very similarly, the reductions are about 200 micron reduction in each of the different drug doses.

And then we come to safety. Overall, there was a very low incidence of treatment-related ocular adverse events through week 16 in both trials. So the numbers are quite low actually in terms of what it is. You can just recall that these patients have uveitis, and so they already have inflammation. And so it's natural in the course of uveitis that you may sometimes see a little bit of inflammation happen and investigators could report that as an AE. But there was very low numbers of AEs leading to treatment discontinuation across all treatment groups and very, very low SAEs as well. We can go on. And more I think this is probably of most interest in terms of significant adverse events that we really do care about are intraocular inflammation, which, again, is what uveitis is, but really the more devastating complication can be retinal occlusive vasculitis, which is where the blood vessels in the retina become inflamed and you can have ischemia and infarction of the retina that can really have a risk of vision loss.

So these patients have uveitis, so it's natural that the uveitis can go up and down a little bit. Even with that, there's very low levels of intraocular inflammation, just a few patients in each of these treatment groups. So overall, that's pretty low, and that's expected. But there were 0 cases of retinal occlusive vasculitis across the entire study in both Phase III trials. No endophthalmitis and it's a 16-week study. So it's a short time point, but very low rates of cataract and pressure. Again, these things can happen in patients with uveitis, but they were very low across all study arms. We can go on.

So in summary, vamikibart offers -- this is the first non-steroid intravitreal therapy that is an option for patients with uveitic macular edema. In terms of efficacy, it really -- it did lead to rapid improvements in vision and macular edema. We did show that there was that -- that the primary endpoint of vision did meet or did miss the primary endpoint in the high-dose group in the SANDCAT trial, but the results are very consistent really across both studies in terms of vision, anatomical outcomes in terms of reduction of central subfield thickness.

So from my standpoint, I'm seeing that there's a clear signal that this anti-IL-6 strategy is leading to improvement. And in terms of safety, overall, very well tolerated and really the key no events of retinal occlusive vasculitis. Given that we currently really struggle to take care of patients with this and that all we have are steroid injections, which have very known side effects, that are pretty -- that are very significant. There is a huge unmet need to have other therapies available. This could be a very novel nonsteroid option that could, I think, change the treatment paradigm for patients with uveitic macular edema. I think -- and that's what we have for now, and I look forward to the discussion. Thank you.

Hi, everyone. This is Cesar Briceno from University of Pennsylvania, and I'll be sharing with you the results of the Phase III SatraGo-1 and SatraGo-2 satralizumab trials in thyroid eye disease.

Next slide, please. So thyroid eye disease is a complex inflammatory autoimmune disease that at this point has limited durable treatment options. Currently, there are treatment options targeting the IGF-1 pathway, namely trontinemab, and that can be associated with some serious adverse events as well as a not infrequent relapse of disease activity.

As a result, there is an unmet need for other targeted therapies that modify disease progression and that minimize long-term complications and have more favorable side effect profiles. IL-6, as has been previously discussed, is known to drive inflammation and that has been shown in thyroid eye disease as well. Basically, in steroid-resistant patients, it's been studied as a modality of treatment by way of tocilizumab and shown to reduce clinical activity score in a significant way versus placebo. Next slide, please. The SatraGo-1 and SatraGo-2 trials were identically designed in Phase III randomized, placebo-controlled studies.

And in this cartoon, we can see the crossover design. So patients were randomized 1:1 to receive satralizumab every 4 weeks versus placebo every 4 weeks. And after 24 weeks, based on their proptosis response, then they were rerandomized to receive either placebo or satralizumab for an additional 24 weeks. This particular presentation is of the 24-week data where the primary endpoint was the proportion of active participants with thyroid eye disease achieving a greater than 2-millimeter reduction in proptosis from baseline at week 24. I should note that these trials also included inactive participants in order to look at how this molecule behaved in inactive thyroid eye disease in addition to active thyroid eye disease.

Next slide, please. The baseline demographics were largely unremarkable and similar between the groups with 2 notable exceptions. First, there was a slight preponderance of female participants in the satralizumab arms. And in SatraGo-2, there was a slightly higher proportion of Asian patients.

Next slide. Looking now at the efficacy, the proptosis response was observed in a higher proportion of participants receiving satralizumab. On the left side of the slide, you can see the active thyroid eye disease primary endpoint. In SatraGo-1, there was a 49% response rate versus 31% for placebo. And in SatraGo-2, there was a 53% versus 23% with a larger treatment effect. As a result, SatraGo-2 was able to reach statistical significance here where SatraGo-1 was not.

If you note, the primary difference is in the rate of efficacy in placebo rather than the rate of efficacy in the active arm. On the right side, we have inclusion of the inactive patients in addition to the active patients. The inactive arms were capped at 25% of enrollment. And in these, the primary endpoint was met in both SatraGo-1 as well as SatraGo-2.

Next slide, please. Here, we look at this in a little more detail. We can see the proptosis response over time over the 24 weeks in SatraGo-1 on the left and in SatraGo-2 on the right, and this was for the active TED population. As you can see, there's a steady increase in dark blue in the proptosis response for the treatment arm. The placebo arm in SatraGo-1 was quite variable. In SatraGo-2, you can see a similar pattern with a slight increase over time in the proptosis response, but the placebo arm was more steady in SatraGo-2.

Next slide, please. Here, we have similar histograms looking at the results when we include the inactive patients in addition to the active patients, and they're rather similar. In SatraGo-1, you have a steady increase in proptosis response over the 24-week period versus some variability in the placebo. And in SatraGo-2, steady increase over the 24-week period with a more relatively steady placebo arm.

Next slide, please. We also looked at diplopia and how that was reduced by treatment with satralizumab. In the active TED population, the diplopia reduction scores were 44 versus 34%, which did not reach statistical significance. And in SatraGo-2, it was 61% versus 26%, which did reach statistical significance. This was, by the way, measured as a reduction of 1 point by Gorman score. On the right side, we have the active plus inactive TED population. And here, again, in SatraGo-1 at 47% versus 28% did not reach statistical significant. But in SatraGo-2, it was 59% versus 28%, which did reach statistical significance.

Next slide, please. We also looked at clinical activity score reduction as one of the secondary endpoints in this trial. Here, we're looking at a greater than 2-point reduction in clinical activity score. On the left side of histogram, we see that there was a significant reduction in CAS in both SatraGo-1 and SatraGo-2. And the achievement of a clinical activity score of 0 or 1 at week 24 was also assessed, and this turned out to be significant in both SatraGo-1 and SatraGo-2.

Next slide, please. We also looked at overall response rate, which was defined as a greater than 2-point reduction in clinical activity score and a greater than 2-point reduction in proptosis. When looking at our active TED population in SatraGo-1, this was achieved in 44% versus 18% of patients, which was statistically significant and in SatraGo-2, in 50% versus 20%, which was also statistically significant.

Next slide, please. One of the things that's most impressive about this trial is the low incidence of adverse events and treatment discontinuations with satralizumab. There were no serious infections noted. And as you can see, the adverse events in the treatment versus placebo arms were essentially equal between -- in both SatraGo-1 as well as SatraGo-2.

Next slide, please. Most of the common adverse events were also listed here. You can notice a similar pattern. There were very, very few in the way of adverse events in the treatment arm, and they were equal statistically to what was observed in the placebo arm. And this included measurements such as neutropenia and also things like nasopharyngitis. And there's a notable absence of the typical adverse events that we worry about with IGF-1R blockade, such as hearing loss, spikes in diabetic sugar measurements as well as exacerbations of irritable bowel syndrome and muscle spasms.

Next slide, please. So in summary, satralizumab was efficacious and extremely well tolerated in participants with TED. We looked at the pooled data between SatraGo-1 and SatraGo-2 since these were 2 identical Phase III trials. And when pooling the data, looking at satralizumab versus placebo, statistical significance was reached across all the endpoints that are listed in this slide, including the proptosis response, CAS reduction, achieving CAS-01, overall response and diplopia reduction.

Next slide, please. So in summary, satralizumab showed clinically meaningful improvements across key efficacy endpoints, including those that we listed in the prior slide, proptosis, diplopia and CAS in those patients with active thyroid eye disease. The safety profile is consistent with what we've seen with satralizumab in NMO spectrum disorders, which is to say that it is highly favorable with no reports of the adverse events commonly associated with other treatments for TED. Satralizumab, thus, is the first IL-6R inhibitor, combining both a positive benefit to risk profile for active and inactive TED with a convenient subcutaneous dosing and a wonderful adverse effect profile. And it could, as a result, represent a new treatment paradigm and certainly a new very useful option for our patients with TED.

Next slide. I think we're done.

Okay. I think with that, we can start the Q&A session. I'm currently having a technical issue here. Let me maybe quickly jump to a couple of the questions we had coming in, in the Q&A section, do them first. So there were a couple of questions from [Gena Wang] from Barclays. For the SANDCAT and MEERKAT studies, what was the reason to use the percent of patients to achieve a 15-plus letter gain as a primary endpoint? And is there any drawback compared to single BCVA changes versus the control arm?

Maybe I'll take that to start with. Essentially, the 15-letter primary endpoint is a regulatory required primary endpoint because that's recognized by the FDA as the letters required to demonstrate a clinically meaningful improvement or lack of deterioration in visual acuity. So that's why we use 15 letters. However, I think as you see, as Nisha shared very nicely, the drawback is that -- because it's a dichotomous endpoint, you either hit 15 letters or you don't, there's increased risk of not kind of capturing the full benefit of the visual acuity gains, which we see.

And that's why the data which showed the improvements in visual acuity on average, the mean average changes from 9 to 12 or 13 letters with the 2 different doses are very, very robust gains. But Nisha, anything to add from you?

Yes. I mean I will tell you that we actually -- in the clinical world, we actually consider even a line gain. This is a very hard condition to treat. And we did recently completed a lot of NIH trials for uveitic macular edema. And even with steroids with patients who had uveitic macular edema, we were seeing -- we were seeing 5, 6 letters, and we were happy about that, right? And so 15 letter is a high bar. I understand that, that's been the norm in other kind of retina trials. But uveitis can also damage vision, and it can be difficult to always have vision correlate with the benefit of a drug, right, because you have uveitis and you also have macular edema.

So I would say that clinically, seeing the net change in vision is very important for us and seeing gains of 10 to 12, 13 letters is really, really very meaningful.

Okay. That's very good. Then we take the first questions here from the queue. It's Michael Leuchten from Jefferies.

Maybe just going back to that exact question on the spot estimates versus percentage of change. Just wondering about the regulatory discussions around filing. Can you pull the data or do the studies have to be submitted independently, MEERKAT and SANDCAT, that is?

And then a question on CST versus the visual acuity. In terms of dose response, it seems to be very little between the 0.25 and 1 milligram on CST. Does that matter in the way 1 thinks about the dosing and the patient response?

Sure. Maybe I'll take a stab to start with. In terms of the regulatory pathway, so we -- as you've seen on the press release, we are discussing with the regulators the path forward. Obviously, it was unfortunate that the 1 milligram failed the primary endpoint. And because we have hierarchical testing, the 1 milligram went first and the 0.25 milligram went second. So therefore, the 0.25 milligram p-value becomes nominally significant rather -- and that's why the study is regarded as technically a failed study. Nonetheless, we are -- as I say, we will be discussing with the regulators, and we'll be looking forward to kind of sharing next steps as soon as we're able to on that side. In terms of the dose response, I think Nisha shared quite nicely the -- as a reminder, the fact that the vamikibart at both doses suppresses intravitreal IL-6 all the way out to the 16-week time point. And therefore, it was no major surprise that there was no dose response bearing in mind that we were dosing these both doses at every month.

But maybe, I mean, in terms of kind of the clinical difference on those 2, Nisha, any comments on CST.

In terms of the -- do you mean in terms of the CST, in terms of both treatment groups, they're both very robust and looks very comparable. So I think that really fits. We've seen that IL-6 was suppressed in both groups. From prior trials, we know that there's rapid IL-6. So it could be that you're getting an effective and sufficient suppression, right, to be able to get your maximal benefit from this with both doses. So I think that actually explains that in my opinion.

Michael, all questions have been answered?

Yes.

Then, the next one would be Peter Verdult from BNP Paribas.

Yes. Thanks, Bruno. Peter Verdult here from BNP. And thanks to the doctors for their time. Just a couple of questions, please. On vamikibart, you talked about steroids being the current plan of care. Just to help us understand the pace of adoption should vamikibart get approved? And also, could you just parse out the competitive landscape? I apologize, I'm not -- I'm all over this disease area. So if you could just put it into context. What else is out there that's sort of grabbing your attention?

And then on Satra, forgive me for calling a spade to spade, you've got 1 failed study, 1 positive. If you were a half, you'd say the absolute level of efficacy versus TEPEZZA is much lower. I take your point on placebo effect, but when you do it on efficacy, it's lower. They are going to have subcuts so the convenience at home goes out of the window. So what is the proposition here? Is it about safety? Or where does this really fit in the treatment paradigm?

And to you, Chris, is there any scope to pull the data? Or is there a risk here that the FDA will say we want 2 positive studies before considering approving the drug? Sorry, a lot of questions there, but over to you.

Should I take the first question was about where this will fall clinically, right? And how would the uptake be and all of that? Yes, I mean, we have a lot of patients with macular edema because we said that it's 30% to 40% of patients have macular edema at some point in their uveitis life, right? So we see a lot of patients with this. We have to treat it. I mean it's not an option not to treat it, and we only have steroid injections. We have a lot of them, not a lot. We have a handful of different versions of steroid injection and implant. But the bottom line is they're all steroids. You can't -- we have to treat because you lose vision.

And the problem with macular edema is once you lose the photoreceptors or you have damage, even if you then treat late, you may not regain your vision. And so we have to treat. So we basically just are forced to kind of use what we have and we deal with the side effects. And we just -- patients are suffering from kind of having repeated injection of steroid and they have cataracts and they have high pressure.

So I would say that if we have something come out that is -- that's shown to be effective and safe in Phase III trials, there would be very -- there is very strong interest in the uveitis community for that. There's been -- we haven't had that, and we need it. And it would be almost, I think, an immediate uptake. It's an injection. It's a simple injection to give, so everyone is comfortable with it. Retina doctors also can give this. It's not something that's hard to administer, right? It's just an intravitreal injection.

So I think in terms of uptake, I think there would be a very quick uptake to try substituting this instead of steroids. There's lots of patients where steroids are a big problem because they have glaucoma, there's other issues where they have like very -- you don't want to give steroids. But even in all patients, there's risk with steroids. And so I really see this as being applied as really as an alternative to steroid, which is preferable. I don't think we mentioned this in the talk, but if you look at the CST reductions, which is what we presented here and you look at what they were in recent trials of like the OZURDEX implant, which is the go-to steroid injection implant for macular edema, they were really comparable.

We saw in this -- in MEERKAT/SANDCAT, it was like 38% or so -- it was a 38% combined CST reduction, right? That's almost 40% reduction. If you look at the dexamethasone in OZURDEX implant in the MERIT trials that were done, it was 35%. So you're seeing really comparable kind of efficacy without the side effects. So I see it as being a replacement for steroids.

Yes. And I think we can move on to the Satra question. And Cesar, if you want to comment.

I'd be happy to. Yes. So I think your questions are spot on. I mean 1 of the key differences between satralizumab and the known data on IGF-1R blockade is that the efficacy is lower. I think there's no secret about that. I mean the efficacy, if you look at the Phase II and III trials for teprotumumab was in the 80s. If you look and include the Phase IV data, it drops a bit to maybe around 70 or so, and we're talking about around 50 for this trial.

I think that the primary value proposition is surrounding safety. There are 2 groups of people that primarily get thyroid eye disease. It's a bimodal distribution. And the first group is a young and relatively healthy group of people. The second bump is an older and relatively less healthy group of people. And in that second bump, which is a lot of what I see at the university, there are a lot of both relative as well as near absolute contraindications for treatment with teprotumumab. None of them are absolute. But in my mind, I wouldn't give it to certain patients. Those are brittle diabetics. Those are people who already have sensor renewal hearing loss, and there are people who have trouble with IBS or some sort of other GI inflammatory issue. Given that we live in the U.S., this is not a small number of people.

And then on the flip side, there's also the people who recur after treatment with teprotumumab. Even if the efficacy is quite remarkable, within 6 months, we see about 1/4 of those people then really have a reintroduction of their proptosis close to levels that they were before treatment or even sometimes beyond or to the same level. And so there is an unmet need for another class of drug that has a label for thyroid eye disease.

So as a clinician, I hunger for this, and I know that a lot of my colleagues do as well for treatment of those 2 specific groups that I just described, the ones in whom we're too hesitant to give the IGF-1R blockade because the side effect profiles are unacceptable. And then the second group of people who recur and then you need to have some sort of other option in order to treat them because, otherwise, you're right back to where you started before you expose them to 6 months of another biologic.

And so there is a need because there's only one FDA-approved drug for this out there for class variability so that we can have a better opportunity to treat our patients in a more tailored way. So it's not just hammer and nail. It really becomes more tailored and more thoughtful as far as how we treat our patients.

And next question goes to James Quigley from Goldman Sachs.

Hopefully, you can hear me. I've got, well, 1.5 on the SANDCAT and MEERKAT trials, and then maybe another 1 on satralizumab. So within MEERKAT, MEERKAT looks like exactly what you'd expect from the distribution, but is there anything that you still noticed about SANDCAT as to why the higher dose led to a lower response? Is there anything in sort of how the trial was conducted. Anything in baseline characteristics that might give you clues as to what happened there?

And the half question for SANDCAT is between weeks 12 and 16, there's almost doubling in the placebo response. So sort of what happened there? It seems unfortunate that's when you measure the primary endpoint that something happened in placebo for the BCVA to almost double. So any sort of color you can give there would be brilliant. And then similarly on the SatraGo-1 trial, any thoughts on why the placebo group was so volatile relative to what we saw in SatraGo-2?

Do you want to try to take that, Chris? I mean -- or you can comment first. I don't know.

Maybe I'll comment first and you chip in a little, then do the same for Satra. Essentially, really great questions. And we've only just recently gotten the data in-house, and we're still exploring. I think that's the basics. The fact is that uveitis is a very complex condition, whereby patients' disease can rebound at any given time, and there are lots of confounders, including systemic medications.

So the simple answer is we're still actively looking through the data, but we've not found anything. But maybe, Nisha, do you want to comment on that.

I mean I don't see anything in the baseline characteristics that is obvious as a reason why. I think that uveitis is a heterogeneous. It's not one disease, right? It's like I mean it is an inflammatory condition, but it's probably 40 or 50 different types of entities. So you have -- everybody has different levels of uveitis, different involvement with uveitis. And so if you look, I mean, I think in SANDCAT, right, it was so close on the 15-letter mark. I don't see that as being like a -- that there was something wrong with it. I just think that it was 1 patient or 1 or 2 patients, right, that probably if they had gotten -- or if they had gotten a little bit more vision, a letter or 2, you might have reached the endpoint. I think 15 letters is a very high bar in uveitis. I understand that was necessary that you were told you had to have that.

But I mean, I was actually -- I mean that's a very high bar to get. But when you look -- I don't think that -- I guess when I see that it didn't reach the p-value, right, for me, that doesn't mean from an efficacy or a signal standpoint that there was something wrong with that dose or that there was something that happened. It just -- I kind of look at it in totality. And I realize there may be implications for filing. But when you look at the overall signal of efficacy across all the endpoints, it's really consistent across both trials and across both doses. And that's the key point for me is that the strategy seems to be effective and safe.

But there could be -- when you look down at the specific cases, maybe you'll find that there were some really difficult uveitis cases, right, that got randomized into that group. There's a reading center for this trial looking at every single image, looking at vascular leakage, looking at everything. So there may be insights that we'll learn that led to that. I don't think we know yet.

And then maybe a quick comment in terms of the Satra and then I'll pass to Cesar is, again, similar to vamikibart. We've only recently gotten the data in-house. We are giving a really good look to really attempt to understand exactly why the placebo response rate was so much higher than we've seen in any other study in thyroid eye disease. Of note, obviously, these are 2 global Phase III studies. So that's not been carried out to that extent by any other therapeutics. So maybe, Cesar, in terms of your feeling on the variability?

Yes. I mean I think one of the things that we're waiting for, of course, is for not only the longer-term data in both trials, but also a deeper dive into some of the nitty-gritty about recruitment and how things were done in the trial between the 2 to explain that difference.

But looking at the baseline demographics, that's the only thing that stands out to me as a potentially relevant difference. Now this is just a product of the fact that these were worldwide sites and some sites did better than others. And so as a result, if you look at SatraGo-2 versus SatraGo-1, there's a higher proportion of Asian patients in SatraGo-2 than SatraGo-1. So you may ask yourself, why does that matter? The reason is because we know thyroid eye disease to actually be quite different in Asian orbits versus others. The anatomy of the orbit is quite different. And I'm speaking specifically about East Asian patients, primarily in China, Korea and Japan.

There is just a different phenotype of how thyroid eye disease sort of manifests. And I wonder whether some of the differences that we are observing can be attributed to that. I don't know the answer to that right now. But it's the only thing that actually stands out to me upon initial review of the baseline demographics that could lead to some heterogeneity there between the 2 trials, even though they were designed to be identical. The preponderance of female patients in the active arm doesn't surprise me because that's how thyroid eye disease behaves, and that was actually quite similar between the 2 trials. So I'm less concerned about that.

And so I'll be curious to see if there's anything that really bears out from that difference in demographics between the trials.

Then we move on, and next 1 would be Steve Scala from Cowen.

A few questions. First, the first question perhaps was just answered. But was there a difference in rescue therapy use between SANDCAT and MEERKAT that might explain the difference in the outcomes?

Second question is, is there a precedent in ophthalmology to garner approval with one successful trial and one failed trial when the failed trial shows a very good trend. And then thirdly, I'd like to ask the KOLs and how many patients do you use ocular Port Delivery systems? Is it just a few? Or are there quite a few?

Okay. I think, maybe I'll start. In terms of the difference in rescue therapy, both study sites were treated. Both studies had exactly the same rescue criteria for vamikibart. In terms of precedent that we've seen with the U.S. regulator specifically, I think I would probably go through 2 specific examples. OZURDEX a number of years ago had a technically failed primary endpoint and then they reviewed it and they amended it, and those were the GENEVA studies.

And more recently, we see the technically failed primary endpoint for the OAKS and DERBY studies in geographic atrophy, which with additional data, obviously, pegcetacoplan did go on to get an approval for the first approval for patients with geographic atrophy.

So there are precedents out there. And again, going back to what our plan is, we are looking forward to discussing the data with the regulator to get their feedback on the best path forward because we do believe that patients would benefit from both of these therapies as additional treatment options. And sorry, your last question?

Port Delivery systems. I'm just wondering if they're commonly used or are they're rarely used?

Yes. I mean, I'm a uveitis specialist. So I do take care of, obviously, retinal involvement with uveitis, but we do not use the -- nobody uses the port system in uveitis. It's used, I think -- and I can't comment on how often it's used, I know it's used in AMD and DME, right, for more extended delivery of anti-VEGF. But we don't use -- we can't use -- we don't use this in uveitis because all we have approved and all we have available are the FDA approved steroid injections and implants. And those are not -- those are what they are. So there's been no application of extended port. That's not to say that, that couldn't be something in the future, but that's not there in our field for uveitis.

Yes. I should just comment. Thanks for the question. So first, the Port Delivery system, Susvimo was only approved at the moment in the U.S. for indications for neovascular AMD and recently for DME and also recently for diabetic retinopathy, and we're seeing that uptake tick up. But not at the moment, indicated for either uveitic macular edema or thyroid eye. There wouldn't be a place for thyroid eye anyway, but it's mostly for the back of eye retina additions.

We don't utilize that in [oculoplasty] at all.

Steve, any additional questions.

That's it.

Then next one would be Matthew Weston from UBS.

Thank you very much for my questions, Bruno. I'm really trying to understand the target dose and dose response on vamikibart. The FDA is increasingly commenting that it wants drug companies to focus on the lowest efficacious dose. So are we sure that the 0.25 mg isn't a reasonable target dose given the totality of the evidence that you've shown? But then the question there is why do we get more intraocular inflammation at the lower dose?

I mean, Chris, maybe you can talk about the -- yes, because I don't think you've said what dose or what dose it's going to go ahead for. But to me, yes, the 0.25 dose, they both look very, very strong and comparable. In terms of the inflammation, just keep in mind, these patients have uveitis, and there was very low intraocular inflammation. I thought it was like there's 3 patients in the 0.25 and maybe 1 patient in the 1 milligram. Those are really low numbers. These are patients who have uveitis, right? So any of them could have a fluctuation in their uveitis where it gets reported as intraocular inflammation. Those are such low numbers. I don't see that as being a meaningful difference between -- that's dose related. It's not enough in my view to say that you're having more intraocular inflammation with the lower dose. Those are very low doses across both studies, and you have uveitis patients. So it's just sort of -- you could have a patient in there who has more fluctuation.

Yes.

Obviously, it's very clear. Chris, can you comment on the target dose?

Yes. So the answer is no. But what I would reiterate is that, obviously, we've seen with the PK data that it does suppress out to 16 weeks. Both doses suppress very nicely. We see very comparable, almost identical best corrected visual acuity outcomes across both doses and central subfield thickness is very -- almost nearly identical for both doses as well. But we've not kind of made a formal decision on dose yet, and that will depending on the regulator.

And I think we have a final question coming again from Michael Leuchten, Jefferies.

Just a quick follow-up on the last one, please, Chris. So just to understand the rationale from going 0.25 to 1 milligram, if you picogram per milliliter is having such a profound effect on IL-6 already going up the dose curve by 4x, that's quite a step up. And I understand why you would test the 1 milligram first because, I guess, rationally, a higher dose should get you there and then you can test something lower. But that step up 4x is quite something. What was the biologic thinking going that high?

So as you saw, we moved directly from Phase I to Phase III. It was a large kind of a large 30-odd patient Phase I extended arm, a couple of extended arms there. But I think that gave us confidence that we would have a desired treatment effect. At the same time, we need to have a reasonable dosing difference, so you can show difference in efficacy, which was the intent of having a fourfold difference in the 2 dosing regimens.

So that's why we kind of selected 1 milligram as the highest and 0.25. And then finally, after the primary endpoint, we are actually transitioning to a kind of an as-required treatment regimen, as you saw in the trial design. So there's a potential for a higher dose to have a longer impact. Obviously, if you -- every time you double the dose of an intraocular drug, you get an extra half-life of durability. So there was that opportunity there. But that data will come at the 1-year time point. But for now, I think it's great to see the comparable efficacy of the 2 doses.

Very good. I think this was the final question. And I think we will close the event. Thanks again to all the speakers for your time and your support and commitment and dedication and also the IR team members who prepared the session today, especially [Anita Tang and Lauren Kam], who had the lead and did all the speaker preparation. Also Melanie Wolf who did the event organization.

I hope the event was helpful and a timely update of the -- to the ophthalmology franchise and the data recently presented. If there are any remaining questions, then please reach out to the IR team. We are happy to come back to you. With that, I wish you a good day, and bye-bye.

So welcome to our Pharma Day 2025 here in London. It's good to see here full house. I was told we have, I think, a record attendance here on site, and we also have quite a high number, I think, of web participants. So let me quickly take you through today's agenda and make a few upfront remarks before we kick off the event.

Regarding the agenda, we will have again 2 sessions. We have a 2 hours morning session focused on strategy and the 2 hours, 10 minutes afternoon session focused on the evolving pipelines of our 5 therapeutic areas. Both sessions include a 30 minutes Q&A at the end, where you will have the opportunity to ask all your questions here in the room and also via the web. Let's start with the morning session.

So our first speaker, as you can see here, will be Teresa Graham, our Pharma CEO. She will provide an update on the pharma strategy based on our 5 therapeutic areas, and she will also take us through our well-established on-market portfolio. In addition, and for the first time, Teresa will focus also on our emerging obesity strategy, where we have made significant progress in building a leading portfolio with 2 major deals announced this year, the Zealand deal and just last week, the bio89 (sic) [ 89bio ] deal.

The second speaker for the morning will be Levi Garraway, our CMO and Head of Global Product Development. Levi will build on last year's presentation, providing an update on our R&D excellence initiative, which was started now 2 years ago and which has had a tremendous impact on how we manage our pipeline portfolio and how we allocate our resources to achieve long-lasting success and to really make sure that we achieve the best outcomes for most patients.

Let me also mention here, and I think you have a little card on your desks as well that we will have a fire alarm going off at exactly 11 a.m. This will be for 30 seconds. And there is no need from your side to rush out of the room. And just to be very clear and to avoid any rumors, I have not been behind this fire alarm to make sure our speakers stay in time, just want to avoid any mistakes here. And following the second -- so following these initial presents and the 30 seconds fire alarm, we will have the 30 minutes Q&A session.

On stage, we then will be joined -- we'll have our speakers, but we also will be joined by Karsten Jung, our Global Head, Pharma Strategy; and by Morten Lammert, our Global Therapeutic Area Head for the CVRM franchise, who joined us from Novo. And there afterwards, we will go to lunch. Lunch will be 50 minutes from 11:30 to 12:20. At lunch, you will have the opportunity to meet all of our speakers and also our IR officers will be around. And we also will be joined by our sponsor, our CFO, Alan Hippe.

After lunch, we will continue then with the deep dive sessions on our late-stage pipelines. So we have 5 sessions, one each for our therapeutic area. The first session, as you can see here, will be held by Charlie Fuchs, our Global Head of Oncology and Hematology Product Development, who will take us through the oncology/hematology pipelines and who will comment on our news of the day. As most of you have probably seen, we had a top line release out this morning on the positive Phase III evERA results for giredestrant in second-line plus hormone receptor-positive breast cancer in the post-CDK4/2 setting.

Second speaker will be our Global Head of Neurology and Product Development, Hideki Garren. Hideki will provide us an update on some exciting pipeline projects, which have recently entered late-stage development. This is trontinemab in early and late Alzheimer's disease and prasinezumab in Parkinson's disease. He will also touch on exciting emerging early-stage assets like the gamma secretase modulator in Alzheimer's disease and the NLRP3 molecule in Parkinson's disease.

The third speaker then will be Larry Tsai, our Global Head of Immunology Product Development, who will share with us an update on different B-cell depleting approaches developed for autoimmune diseases like lupus. These are CD20, CD19, CD3 bispecifics being tested or allogeneic CAR-Ts. And he will also provide an update on our TL1A development program, where we are exploring additional indications beyond the IBD and where we have taken the bispecific p40/TL1A into clinical development. Fourth speaker then is Chris Brittain, our Global Head of Ophthalmology Product Development. Chris has also some exciting early news to share. We just got data in-house for several studies and the data are currently still analyzed.

There is data for vamikibart, Phase II/III in DME and UME. There's also the Phase III data for satralizumab in TED. These data are planned to be presented in the next few weeks at 2 conferences, ASOPRS and AAO, and we have also scheduled now an IR call for October 21, where we'll cover these data in more detail. And finally, to close with, we will have the presentation of Manu Chakravarthy, our Global Head of Cardiovascular, Renal and Metabolism Product Development. Manu will take us through the significant progress we have made over the last 12 months in building a leading obesity CVRM portfolio, creating a lot of development and combination optionality.

And as you have seen, the latest add-on to this has been the deal announcement from last Thursday, acquiring bio89 (sic) [ 89bio ]with FGF21 analog in late-stage development for MASH Stages II, III, IV. After the closing remarks from Teresa, we will have a 30-minute Q&A with all speakers from the morning and afternoon sessions on stage, including Alan. And the event will then close at 2:30, but you are invited to stay around for a buffet reception.

Let me also mention one more housekeeping item. As in previous year, we have again prepared a short 10 minutes feedback survey. The link to the survey will be shown to the participants in the webinar 50 minutes before the end of the event. Participants in the room will receive an e-mail roughly 1 hour after the event. We would very much appreciate if you could give us your feedback as we always strive to further improve. Then let me quickly go to my next slide. I just wanted to have this slide in here to remind you that Roche has consistently delivered strong top and bottom line growth over the last decade.

This slide summarizes our sales development from 2015 to '24 at constant exchange rates. And what you see here is that we have delivered a 10-year CAGR for group sales, which has been plus 5%. And then if you look at the core EPS level, even a CAGR of plus 8%. Of course, reported numbers look slightly different since the Swiss franc has throughout this 10 years, continuously strength versus all other major currencies on the world. And let me also highlight here that during this period, we had to manage several challenges, especially the CHF 21 billion patent cliff in 2018. And by successfully managing this cliff through in-house innovation, we are now left with a rather fresh portfolio relative to many of our peers in the industry with currently 17 blockbusters on the market. And this brings me to my final slide.

As you can see here in light blue, this is the segment of our current on-market Pharma portfolio, which now is expected to deliver growth until '28. So this is -- has been pushed out 1 year. Last year, we had communicated growth until '27. And after '28, actually, we expect this current on-market portfolio to be stable until the end of the decade. So this means our current growth products are expected to always compensate for any generic erosion occurring in this period, and there's no patent cliff forecasted. And Teresa will provide you some more details and an update on this on-market portfolio.

You also see here 2 layers on top. In light gray, you see the expected contribution from the current in-house pipeline and in dark gray, additional contribution from future business development. The in-house pipeline, we'll discuss later today. And let me also quickly add here that we have made new epidemiology slides for our pipeline projects available on the IR homepage, so you can download them. To close, let me also quickly comment on the Dia division. As you know, we are quite optimistic on the long-term outlook, as we have a couple of launches ongoing and upcoming, including our unique mass spectrometry solution for the hospital setting. The launch of our CGM solution and our revolutionary -- truly revolutionary SBX next-generation sequencing solution. All of these are clearly blockbuster opportunities just to use Pharma slang here and to characterize the opportunity.

And finally, one remark I have to place on group profitability. We have been continuously communicating we strive to keep the group margin at least -- defend the margin and keep the margin at least stable. And with that, I hand over to Teresa. Teresa, please.

So I have to admit I'm a little hurt that this fire alarm was about us not staying on time. So we will have to talk about that, Bruno, later. So thanks, everyone. Thank you, Bruno, and thanks to everyone who's joining us here in London live and also to everyone on the phone, and welcome to Pharma Day 2025. When we were together here last year, I shared with you for the first time our pharma strategy. And at that time, I also made a commitment to you about what you could expect to see from us at Roche going forward. And my commitment to you was that you would see rigor in the science and discipline in the business.

And I think what we have to share with you today is going to draw a double underline under that commitment. We'll show you how the pharma strategy is doing exactly what we expected it to do to provide focus and clarity across the entire enterprise. R&D excellence continues to identify efficiencies and to allow us to accelerate our pipeline. As Bruno mentioned, through the course of the day, we're going to share with you how we are consistently applying the Bar across every aspect of our R&D organization, allowing us to make rigorous scientific decisions. We will share our laser focus on the commercial success of our on-market portfolio and how we're thinking about life cycle opportunities in each of these areas.

We will show you how we are applying significant financial discipline across the entire enterprise and most importantly, how we are maintaining the culture that has always defined Roche and Genentech, allowing us to attract the best talent and develop the best science. So we have quite a morning ahead of us. As always, there's a lot to talk about, so let's go ahead and jump in. So I'm going to start with outlining a little bit of the progress that we've made since we were last here. So I'm sure all of you have printed out this slide and have it hanging above your desk in terms of what the 10-year ambitions are for Roche. We have committed to delivering 20 transformative medicines, addressing the areas of highest societal burden. We committed to increasing the value of our portfolio by 40% by ensuring that our portfolio continued to provide the best level of innovation with 80% of our pipeline having best-in-disease potential. And most importantly, because it is fundamentally here what we are here to do, we committed to treating 3x more patients with that transformational pipeline.

So how are we doing? So in 2025, we're halfway there to our transformative medicines. We've launched 10, and we have more than 10 additional NMEs with transformational potential that could launch by the end of 2029, one of which we just received some very positive data on today with giredestrant, which Charlie will talk to you more about later this afternoon. And this does not include any potential BD deals that might be coming post 89bio. In terms of the value of our portfolio, we have actually overachieved what we had hoped to do. I'm definitely not going to complain about that. We've added 55% in terms of value to the portfolio with the average peak sales per pipeline project.

In terms of innovation, we're on track. Today, 67% of our late-stage projects have best-in-disease potential, and that's an increase of 9% from when we first started talking to you about these things in 2023. Now when you look at access, you might say it feels like you're a little bit behind the curve here because we're at plus 40% versus our goal. But you have to remember that when you look at our pipeline, the majority of our large launch products that will treat significantly more patients launch more towards the latter half of the decade. And that's what gives me great confidence that we will still be able to achieve -- that we will still be achieving this goal. It's just a little bit more time shifted towards the back half of the time frame.

So you can already see that we're making tremendous progress. So what else has happened since we were last together? You saw at half year that we posted 10% growth with our in-market portfolio, 13% COP growth at half year and 1.7% COP margin growth, discipline in the business. R&D excellence, 26% total portfolio value. 55% of NMEs are now post the bar, and we're fast tracking those assets that we believe have the most potential to help the most patients. Levi will talk a lot more about this in his section.

And how are we pulling through the pharma strategy? All 5 of our TAs now have TA strategies that have been aligned to the overall pharma strategy, and we are in execution mode. Our 11 end-to-end disease areas are taking shape, and we will go through each -- we'll go through quite a number of them in the later part of the portfolio. And finally, we'll also be able to share with you a first look at our obesity strategy. So over the last 12 months, a tremendous amount has been accomplished, and we've done it with rigor, and we've done it with discipline. So now let's talk about that on-market portfolio with 17 blockbusters that we have in market today.

Let's start, as always, with solid tumors. Our well-established HER2 breast cancer franchise is expected to peak in 2026. No change here with that strong tail. And as Bruno mentioned, no cliff. We also now expect with our positive data in giredestrant to be able to expand into the HR-positive breast cancer segment, again, the largest part of breast cancer overall. And our lung cancer franchise is also stable, but with some exciting potential with divarasib, which Charlie will go in later today. For hematology, we continue to have one of the most established portfolios in NHL with the potential to further improve that standard of care, as Polivy continues to entrench and the bispecifics continue to launch and as new indications are added.

We also continue to maintain our leadership with Hemlibra in hemophilia A with a lovely life cycle extension with NXT007 entering late-stage development. And speaking of leadership, Ocrevus remains the undisputed leader in MS. The subcu launch is ongoing, and we are eagerly awaiting the fenebrutinib data end of this year, beginning of next. Evrysdi continues to lead in SMA, and we're excited about the opportunity to potentially launch into 2 very large additional diseases with Alzheimer's and Parkinson's, something Hideki will share much more with you about in his section.

In terms of immunology, Xolair food allergy continues to do very well, strong uptake in food allergy and again, no biosimilars expected here until 2026. Gazyva, PDUFA is expected in October. I think you've heard -- many of you have heard me say over the years, I'm a big fan of Gazyva in immunology, and I'm excited to actually have this first indication onto the market. And then, of course, Vabysmo. Vabysmo is their foundational asset in establishing that leading position in ophthalmology, and it is absolutely redefining the standard of care in retinal diseases.

And as you will hear from Chris later today, we have one of the most diverse pipelines in ophthalmology in terms of targets, MOAs, devices. We are clearly positioned for long-term leadership here in ophthalmology. And I always find this phrase a little funny, but even though it's not yet in the doughnut, our obesity profile. Maybe that joke plays better in America. We have an emerging obesity profile -- portfolio, and we have one of the broadest obesity profile -- portfolios out there, and we are well positioned for leadership in obesity, something we will talk about much more later today.

So let's talk a little bit more about the numbers. We are delivering this best-in-disease portfolio with an amazing amount of financial discipline. If you look at our OpEx development over the last 3 years, we have steadily controlled costs, while increasing sales double digit. And that means that our margin development has been impressive. We focus on this every day. How are we spending our money? Are we spending it in the best places? Are we getting the best return? And are we redeploying those funds in the best possible way?

So let's talk a little bit about our portfolio focus and the implementation of the bar. Roche has been and always will continue to be a company that is fundamentally grounded in following the science. But one of the most important things that the pharma strategy has allowed us to do, working in close concert with R&D excellence is to ensure that we are following the science with intention, that we are looking in those disease areas where we can help the most patients, where the science is most fruitful and where we believe we can have the biggest impact.

And what allows us to make sure that we're making those decisions in a rigorous fashion end-to-end across our R&D portfolio, it's the application of the bar. So you can see the 5 Bar criteria here. Again, Levi will talk to them in a lot more detail in his section, but the Bar fundamentally defines what makes it into our portfolio and what advances in our portfolio. And it allows us to ensure that the R&D leaders across our entire organization have a consistent way to make those decisions. So what are those -- what does the outcome of those decisions look like? So today, 55% of our portfolio is now officially post bar.

And you can see that these assets are really compelling and have significant sales potential. This is truly a rejuvenated portfolio from where we were 12 months ago and certainly from where we were 24 months ago. There are a lot of things on this list that you'll be familiar with, a couple with that you may not, cevostamab in relapsing remitting multiple myeloma. Again, Charlie will talk about this a little bit later today, but this has now been moved into Phase IIIs. And of course, you all saw this morning that CT-388 in obesity has also been advanced into Phase III, and Manu will talk about this a little bit later today. And then, of course, post-closing, pegozafermin for MASH will come into the portfolio as well.

So a very nice range of assets across our core therapeutic areas, all with the potential to help a significant number of people and all with significant value. So some of this has happened through business development, and the Bar is also a critical part of how we assess opportunities in BD. So you can see across this slide a number of deals that we've done, all of which closely align to our end-to-end disease areas. 89bio coming in, very exciting addition to our CVRM therapeutic area; Zealand as well, very exciting deal to bring in a best-in-class amylin. But you can see we cover the gambit here, and we're very pleased with the assets -- we're very pleased with these assets and their opportunity to be best-in-class.

So now let's take a little bit of a deeper look in each of our therapeutic areas and how we are defining where we want to play and how we intend to win. So as you know, the pharma strategy was introduced last year. And at that time, we defined clearly our 5 therapeutic areas. You can see them here. But I think what's really important to remember about these 5 therapeutic areas is that they constitute 60% of the total global burden of disease. They also represent 80% of where we believe growth is coming from in the pharmaceutical sector in the coming years. So we are targeted in the right place to help the most people.

Currently, we have identified 11 disease areas in those 5 therapeutic areas that we are investing in end-to-end, all the way back from discovery to commercialization. This allows us to gain the kind of experience that leads to better judgment and scale and commercialization, which will be critical as we think about moving into the future. So let's start our deep dives as always, with oncology. The focus with oncology today is to strengthen our franchise through a very focused approach. We are focused on our end-to-end disease areas, breast cancer and lung cancer in solid tumor and malignant heme and hemophilia in hematology.

And where we are intending to play or where we are intending to win is to find ways to accelerate innovation in the most effective way, depending on where -- which therapeutic area or which disease area that we're in. Breast cancer is probably our most well-developed end-to-end disease area in the portfolio. It's anchored certainly by our HER2 franchise, which, as I mentioned earlier today, is expected to remain the standard of care in the majority of early breast cancer settings. I think we've seen a lot of KOL feedback to this end. And the recent studies sort of indicate there is not going to be a one-size-fits-all solution for patients in breast cancer.

Physicians are going to make very tailored and individualized decisions based on the need of these individual patients. And our HER2 portfolio is both well understood. It's well characterized. People are familiar with it, and they're very comfortable with its efficacy and its safety profile. Phesgo will play a big part in how this -- in how the HER2 program actually plays forward. We're currently at a 46% global conversion rate, and we are set on converting as many patients as possible. But of course, that expansion into hormone receptor positive breast cancer is something that we've been eager to do, and the positive results from this morning's evERA trial as announced, will enable that further expansion.

So you can see we have a very strong on-market presence, and we have a robust breast cancer portfolio that Charlie will talk to you a little bit about later this afternoon. Moving on to malignant heme. This is another place where we are very much currently a leader. Polivy, first improvement in DLBCL in 20 years, raising the Bar on our own therapy, rituximab era. That's exactly what we love to do. We love to raise the Bar on ourselves. Polivy is establishing itself as the new standard of care. You can see these are the U.S. market shares. They continue to grow. And again, we have no doubt that over time, this is -- Polivy will be the treatment that the majority of patients with DLBCL receive.

Our 2 bispecifics, Columvi and Lunsumio are rapidly differentiating themselves in the real world. Much like we expected, their individual profiles allow them to treat and reach different kinds of patients and are uniquely beneficial to health systems in different ways. This will only become more evident as we move into earlier lines of treatment. And finally, as you'll hear more about this afternoon, we are expanding into multiple myeloma with cevostamab and our allogeneic CAR-T programs. Multiple myeloma, very significant patient population, a significant amount of continued unmet need, and we think we have several assets here that could really make a difference.

And no discussion of hematology is complete without talking about Hemlibra in hemophilia A, the undisputed standard of care around the world, more than 30,000 patients on treatment currently. We all know that it has around 80% of patients who are not experiencing bleeds without Factor VIII inhibitors, and this is supported by a tremendous amount of real-world evidence and experience. We're very excited to be moving the next generation of this molecule forward with NXT007. And we have just moved that into Phase III, including a trial that will go head-to-head against Hemlibra. We also recognize that in any chronic disease, convenience for patients is extremely important.

And so we plan to bring an auto-injector forward not only for Hemlibra, but also for NXT007 to ensure that we can help patients have as convenient a therapy as possible. Moving on now to neurology, where, again, we are already the leader. And our goal is to extend that leadership position. And where we want to actually extend it is into preventative neurology, actually helping patients prevent progression into severe neurological symptoms. We also are very keen to work closely with our Dia colleagues to ensure that we are creating an end-to-end integrated journey for these patients. And this is one of the disease areas, as you'll hear from Hideki later today, that we're probably the most far along in that partnership between pharma and dia.

We have 2 end-to-end disease areas, unsurprisingly, MS and Alzheimer's, but we're very encouraged by the data that we have seen out of prasinezumab for Parkinson's and are eager to see what that Phase III data holds. The MS franchise, clearly, the cornerstone of that is Ocrevus, firmly established as the global standard of care, more than 420,000 patients treated globally, and its reach just continues to grow. We are, of course, in the middle of the subcut launch. We now have more than 12,000 patients on treatment. 50% of those are new to brand. And we are rapidly moving into development with our new device, an on-body injector with a high concentration of Ocrevus that will hopefully allow at-home treatment twice a year every 6 months.

And again, I think we should never forget when we talk about life cycle for Ocrevus, we are always also talking about extending that convenience. Treat your MS today with subcut 10 minutes twice a year. That's a very compelling message for patients. Launch of the on-body auto-injector would be projected in 2028, and it is our intent to convert as many of our IV patients as possible to subcut and then subsequently to the on-body injector. Fenebrutinib has the potential as a very selective oral BTK to be very disruptive in the MS segment, and we're very much looking forward to that data. Switching over to immunology. Immunology is a really unique area of science. And our goal here is not only to maximize the individual indications that we have with drugs like TL1A, but to actually leverage those mechanisms across different TAs.

And so leadership for us in oncology -- I'm sorry, in immunology, not only means being successful in specific disease areas, but how those mechanisms may actually also allow us to grow into different therapeutic areas. Our end-to-end disease areas here are IBD and COPD, but I want to call out here the pan-immuno pathways, TL1A, CD20, CD19, OSMR. These are pathways that we are very early looking at not only in immunology, but how they apply in different parts of the portfolio. Immunology today, of course, Xolair is the star in this portfolio right now. You can see the growth in food allergy just continues. As I mentioned before, no biosimilars expected until 2026. Gazyva in LN showed superiority over the standard of care. Lupus nephritis is a devastating disease, extremely serious.

We believe we can help many, many patients with Gazyva in lupus nephritis. And again, that PDUFA is set for October. Moving on to ophthalmology, where Vabysmo is clearly paving the way to a leadership position. We have a broadly diversified pipeline that has significant transformational potential. We intend to be a big player in retinal vein disease for many, many years to come, but would also like to expand into other therapeutic areas, including geographic atrophy and dry eye. Of course, winning in ophthalmology is a lot about how we win commercially today with Vabysmo. And so let's look a little bit at what's happening in Vabysmo in ophthalmology today.

So clearly, there are challenges in the U.S. commercial market. The dynamics that we reported at half year are largely unchanged, but Vabysmo continues to gain market share in the branded segment in naive patients. And that is precisely what we needed to do. There has also been new clinical data that was just recently presented that continues to reinforce strong efficacy and demonstrates the ability to get out from a duration perspective to intervals up to 5 months. Chris will talk to you a lot more about the additional data that we have now in ophthalmology. But you can see, again, we have a large portfolio and one that is very diversified relative to competition.

And so we are really excited about the ability to be a player in ophthalmology very long term and to maintain that leadership position. And finally, moving on to CVRM. Our ambition here is to be a strong entrant before 2030 with competitive products. We intend to differentiate by using the unique capabilities of Roche, and we're going to talk a lot about that in the next couple of minutes. And our goal is to become a top 3 player. And I want you to know that I am serious about this goal. I believe we have the pipeline, I believe we have the commercial capabilities. And most importantly, I believe we have the will.

The end-to-end disease area here, of course, is obesity, but there are numerous places that we are looking at breakthrough innovation, including MASH. So why do I believe that we can do this because I do. A, we've done it before. MS, hemophilia, SMA, retinal vein disease, we know how to enter new markets, we know how to understand our customers. We know how to understand our patients, we know how to understand ways to drive to market that actually allow us to do that efficiently and effectively. And while the disease areas that we are headed into are larger, and we know that they will require us to do things differently and at a different scale, I am just as convinced in our ability to make it happen. So how are we going to succeed?

When we were at -- when we were in London last year, we talked about the core capabilities that were resonant within the pharma strategy, and there were 6 of them. All 6 of these are going to be necessary for us to succeed not only in obesity, but frankly, in all of our disease areas going forward. And we have made significant progress in all of these areas in just the last 12 months. I'm going to start with R&D and manufacturing. As you well know, pharma and dia have the full value chain in the U.S. It has always been our belief as a company that you should manufacture your products close to the people who need to use them.

We recently expressed our commitment to invest $50 billion into the U.S. until the end of the decade. And that includes importantly, one manufacturing site in Holly Springs, North Carolina, where we have already broken ground and a site in Boston, which is focused on R&D and AI and ML and CVRM. All of our key medicines are already produced today in the U.S. with the tech transfer for the one remaining product well underway. And as you all know, we have a high level of flexibility in our very large and diverse manufacturing chain.

And because we have invested so significantly in the productivity of our manufacturing organization over the last number of years, it means when you look at our existing capacity in the U.S., we are at less than 50% capacity, which means we have a lot of opportunity to continue to grow in the U.S. with the facilities that we have and the ones that we continue to invest in. And I call out that when we are a company that actually never left the U.S. We stayed when we acquired Genentech. So that footprint is very much there, and all of the IP for medicines invented in the U.S. is held in the U.S.

But one of the things that has been key to many of the investments that we've made in our Pharma Technical organization and part of the reasons that we've been able to see such efficiencies over time is our investment in data analytics and AI in Pharma Technical. We have spent a lot of time thinking end-to-end across Pharma Technical about ways that technology can actually help us speed delivery to patients, avoid costs and continue to make our processes more robust. We have invested significantly in these areas, and we can see those benefits today.

Our tech times have dropped. Our output yield is up by more than 10%. We can generate required reports faster than we have ever been able to do before, which frees up time to do many more value-added things. And we've been able to reduce our CapEx because we can actually do more with the facilities that we have. So this is a really practical application of data and AI in a way that has real fundamental impact on our business every day. Related to manufacturing is our investment in devices. So again, you heard me talk last year about the fact that 60% of our current portfolio is going to require a device. We need to be able to manufacture various kinds of devices for different uses at scale. In order to do that, we are moving to a device -- a drug delivery device platform, which will allow us to be able to do this much, much more quickly.

We've already identified the first 4, the auto-injector needle safety device, the ophthalmologic prefilled syringe and then our intravitreal device. There will be several more that will be added. But these are going to be critical to ensuring that when we have a product that requires a device, we can get it very quickly into trials. Part of what will allow us to do that is the investment in a device pilot facility in Basel for medical devices and drug delivery components. This is going to give us a platform where we can actually look at how we do this at scale before we actually have to move into big manufacturing facilities.

That's the old building, by the way, in case you were thinking it didn't look very impressive. That's going to be replaced with the new building. That's just the site. We are also leveraging AI in different parts of our business. Right now, in the commercial and medical organizations, we are singularly focused on making sure that we have one of the most robust customer engagement models in the industry powered by AI, allowing us to look across the globe at how our customers are reacting, what they're doing, what that means and how we can inform decisions going forward. This will remove a lot of manual work out of the system, which is great, but really what it does is it makes the people on the ground vastly more effective and efficient.

And that means we can continue to deliver more medicine, to more patients more economically. These AI solutions will also help us continue to increase our share of voice, which as you enter into more and more highly competitive markets is critically important. But of course, all of this is just so many words on a piece of paper unless you actually have the most talented people in the industry working in your organization dedicated to making change. Culture has always been a significant part of what makes Roche and Genentech, Roche and Genentech. And we are singularly committed to making sure that we continue to be an attractive employer, that we drive to be the most high-performing organization that we can be and that our ways of working continually elevate us in the industry.

Our people strategy is critically important to making that happen. And we are constantly looking to make sure that we are identifying the needs of the organization, where do we need critical experiences and competencies that we don't have today and how are we bringing them into the organization. And you'll see a wonderful example of that in a little bit when Morten joins us on stage. So I know many of you have been waiting for this, which is a look at our obesity strategy and how we intend to become a top 3 player in obesity.

We firmly believe that our capabilities strongly position us to deliver in obesity. And that there are 5 capabilities that are going to be critical in order to make that happen, that we have a best-in-disease portfolio, that we're able to leverage the synergies across our TAs, that we have an appropriate manufacturing and supply chain, that we have a highly efficient and effective global commercial footprint and that we are maximizing our partnership with dia to create a meaningful end-to-end patient journey. Now Manu is going to talk to you about the end-to-end patient journey, so I won't make you listen to that twice.

We've talked a little bit about the global commercial footprint, and we've also talked about manufacturing and supply. So in my talk here, I'm going to focus primarily on how we believe the obesity market looks today, how we think it is going to evolve and fragment and then why we believe we have a right to win in this space. So let's start with the obvious. Obesity is different from the other areas that we're in at Roche. The patient is the key decision driver. There's actually a multitrack system for how patients get their prescriptions. So you're not talking with just one type of physician, you have to talk with multiple types of physicians.

In many parts of the world, obesity isn't even established as a chronic condition, which means there is no reimbursement. But yet it is of unprecedented size and scale, something that we have not seen in the industry before. And there is a deep amount of heterogeneity and fragmentation that results from all of these components. So let's start by actually -- let's start by looking a little bit about the patient journey. In the U.S. today, 55% of patients are actually the ones showing up at the doctor's office and asking to be put on an anti-obesity medication.

And they do that for a variety of different reasons. It may be about their health and making sure that they can address things that are physically troubling. It could be emotional and about their psychological well-being or it could be that they have a lifestyle goal that they wish to achieve. But the reality is every patient shows up for a different reason with a different ask of their physician. And they show up to different kinds of physicians. 88% of the time in the U.S., they show up to their primary care physician. The rest of the time, they're showing up to an endocrinologist or some other kind of specialist. And why is this actually so important?

Well, it's so important because depending on the kind of physician you are, you're going to ask a different set of questions about why that patient is there. And you're going to care differently about the treatment that you select based on how you're evaluating that patient and any potential comorbidity that they may have. If you go to an endocrinologist, chances are first thing they're going to care about is diabetes, then cardiovascular, then potentially renal, probably true for GPs as well. Cardiologists, mostly going to care about the heart. And then they'll also think about type 2 diabetes and renal involvement. But what that means is that every patient is on a different journey and every physician is going to have a different set of questions.

And even when they align on the fact that an anti-obesity medicine is appropriate and they make the selection as to which one they're going to go on, there is actually no guarantee that, that prescription gets filled. Again, focusing on the U.S. just because this is where the most data is available, obesity is already identified as a chronic disease. Reimbursement is available, but 33% of the time, a patient who has been deemed appropriate and received a prescription will actually not get a drug reimbursed. They will have to pay out of pocket, which means there is a significant out-of-pocket component to this market, not only globally, where I think is where we relatively always talk about it, but also in the U.S.

So when you're talking about a cash market, that is also a different thing that you have to think about commercially. Now we also mentioned the size and the scale of obesity, which I think everybody understands, but I think this chart does a great job of identifying how many people in the world today are actually eligible for treatment. 51% of the global population will either be overweight or obese by 2025. That is truly an epidemic. That is a massive amount of people who are at higher risk for significant comorbidities and mortality. And yet today, with the treatments that we have, we see a relatively small amount of people with obesity are actually being treated, about 10% or 15% and the mean BMI of those patients being treated is 39.

So you have a relatively small percentage of your most heavyweight patients being treated, leaving a tremendous amount of potential opportunity for patients who could benefit, but for whatever reason, are not receiving therapy today. This leads to a market which is exceptionally fragmented. And there are many different ways in which that fragmentation could actually present itself in the commercial setting. Many different drivers of why a patient or a physician would potentially choose any given therapy. But I want to start by talking about comorbidity management because from a human health perspective, this is a very significant one. And we know that patients with higher BMIs have many more comorbidities.

Over 70% of patients with just one -- 70% of patients living with obesity today have already one comorbidity, and that number dramatically increases as BMI increases. And so what does that mean? It means that obesity is a major risk factor for over 220 different diseases. All of these diseases have significant impact on human health. And it should be noted that we are one of the only companies that has treatments in a lot of these diseases. So not only do we have the ability to help patients with overweight or obesity, but we also have the opportunity to look across our therapeutic areas and see how we could potentially combine different therapies together to actually address some of the most significant health needs that patients have. That is something that is materially different about our portfolio relative to the rest of the industry.

And that patient-centric approach is going to be really critical moving forward. So these are just 3 different types of potential patients who could show up at a doctor's office. There's the patient who is overweight and eligible for an anti-obesity medication, but is probably not morbidly obese, but still eligible. There are people who have an early sort of diagnosis of obesity, Class I with some comorbidity. And then there are Class II to III that have high-risk obesity. What these patients are looking for and what their physicians need to treat them with is very different. And if you -- just to orient you to this slide, if something is bright green, that means it is of highest interest to that patient.

So unsurprisingly, if you think of depth of weight loss, that's actually most important to people who are at highest risk, for the people who are the heaviest who need the most weight loss. But they're not as interested in how quickly they lose it because they're on this journey for a longer period of time. For those people who are potentially looking to lose less weight and have fewer comorbidities, you can see that speed to weight loss is much more important. Tolerability is much more important. Convenience is much more important because they have a very different set of treatment goals than the people on the other side of the spectrum.

And what's exciting for me is that our portfolio allows us to address all of these patients. Now it's undeniable that incretins have unlocked a new era in the treatment of obesity. If you look in the past decade, this was a relatively smallish market at less -- at about $6 billion. But we hit an inflection point with the GLPs coming into the market in force. And we're now looking at a significantly larger market a significantly larger market and significantly more options for patients -- or more options for patients. But the reality is that in the real world, there is still a lot of unmet need. Treatment persistence, as you guys know, is a big problem.

The share of patients who are able to get to that higher dose -- highest dose and be able to experience all the weight loss they want to see, not able to be done so much. And actual weight loss is often below target because of those 2 things. So whatever comes next is going to have to address that remaining unmet need. And we've identified 6 things that we think are going to be really critical to whatever comes next to addressing that unmet need, tolerability.

We all know that nausea and vomiting are major drivers of discontinuation, particularly as you think about those patients who are in the lower -- have the lower desire for weight loss spectrum. The ceiling effect on weight loss. Most patients plateau after 12 to 18 months. They never get to their target weight. Weight maintenance, even if you lose the weight, once you stop treatment, the weight comes right back. And because it's not -- and because many patients don't find existing treatments tolerable, it's harder for them to stay on longer term. We've talked extensively about the comorbidities and the need to treat those as well as to treat the underlying weight problem.

Lean muscle mass loss, again, you guys know this, 40% of weight loss comes from muscle loss. That's less optimal, clearly. We want to find ways to preserve that lean muscle. And then less talked about, there's about 20% of patients who actually have no or suboptimal response to these treatments. And so the existing treatments actually aren't working for them and that's where our portfolio comes in. Our differentiation potential relies on having that breadth of options to address all of the different patient needs. From petrelintide to CT-388 and monotherapy to all of the different kinds of combination therapies, we can cover the breadth of this unmet need completely. And that's not even talking about the opportunities to actually look forward into ways to combine to address comorbidities.

So to sum it up, our capabilities strongly position us to deliver here from multiple pipeline assets with best-in-class and best-in-disease potential to the ability to leverage the synergies across the portfolio, our global robust manufacturing network, our presence in more than 150 countries around the world with our global commercial footprint and of course, our very unique ability to leverage our pharma and dia connection, we believe that those capabilities in total, executed correctly will allow us to become a top 3 player in obesity, and we are committed to making that happen.

Now to close out on future growth opportunities, these are the new 8 NMEs that are new to Phase III in 2025. I think we've talked about all of these, again, just to call out the addition of pegozafermin in MASH, our newest addition after closing and cevostamab in relapsing remitting as well as CT-388 moving into Phase III. You've all seen the consensus slide before, but just to reorient you to it, after half year, we went and took a look at what the consensus was saying. In 2024, this is sort of where people believe we are in terms of sales -- or where we were in terms of sales. In 2029, if you look at consensus, it looks like we have a gap of about CHF 6 billion related to loss of exclusivity. This is primarily Xolair, Actemra, Perjeta and Kadcyla.

But if you then look at where consensus says we have growth opportunity in that same time period, consensus says we can grow by CHF 12.9 billion. There is a lot of opportunity in our pipeline. And by the way, there is a significant number of things in our pipeline, including many of the things that I've talked to you about today that aren't even included in consensus in a meaningful way today. Our cardiovascular and metabolism assets in general are not yet included in models. There are a number of things in the oncology, neurology, immunology and ophthalmology franchises as well that are currently not represented.

And so we believe the true value of our pipeline is actually quite underestimated. And why do I say that? Because by 2030, we have the opportunity to launch up to 18 new NMEs. And that includes 15 NMEs with blockbuster potential. We are well set up within our 5 therapeutic areas to continue to diversify and add leadership in the areas that we know will matter most to patients. And importantly, we have the skills and capabilities and resources in order to ensure that we will make those launches successful and that we will reach the number of patients across the world, who deserve to be reached.

Before I turn it over to Levi, I just want to reiterate what I started with. As an organization, we are committed to bringing transformational medicines to patients. And we are committed to doing that in a way that is uniquely Roche. I am committed to continuing to deliver rigor in the science and discipline in the business to ensure that when we're standing here next year and we're able to share the advancements that we've made, we're continuing to make progress at an even faster clip than we have today. So I hope you'll agree, a lot has happened in 12 months. We're very pleased with where we are, and we're also very excited about the opportunities ahead.

And with that, I will turn it over to my colleague, Levi Garraway, to talk to you more about R&D excellence. Thank you.

Well, good morning, everyone. It's always a privilege to have the opportunity to describe our pipeline progress to the investor community. So as Teresa mentioned, I'll provide an update on R&D excellence, which actually we first rolled that out here at Pharma Day a couple of years ago. And then after launch, of course, our 5 therapeutic area heads will provide deeper dives into specific programs. Now by way of reminder, our overarching objectives for R&D excellence are twofold. So first, we want to deliver many more transformative medicines to the world. And second, we want to become one of the most productive R&D engines in the industry as measured and evidenced by top quartile performance.

And here are the solutions that we've been implementing over the past 2 years to achieve these objectives. And 3 solutions, which are shown at the top of the slide, are focused on expanding the transformative potential of our pipeline. And I'll delve fairly deeply into each of these. Two other solutions are about improving the efficiency of our R&D engine, and then one is focused on incentivizing the organization to perform in this way. Now in addition, streamlining our system landscape and our data foundation, of course, that's also necessary to make sure everything works well. And that effort has essentially become a seventh R&D excellence solution. So I'll touch on that briefly at the end.

Now a key premise for R&D excellence is that if you want to deliver transformative medicines, you need a pipeline that consists exclusively of assets that are capable of becoming such medicines. And of course, that's easier said than done because most therapeutic candidates fail. But on the other hand, pharma R&D pipelines do obey the power law. So that means that although most fail completely, some become medicines that are so impactful that they truly transform the lives of patients. And so we must intentionally identify candidates with that type of potential wherever they might be found along the R&D continuum. And that's where the Bar comes in.

So the Bar is an evaluative framework that we developed a couple of years ago to describe the essential characteristics of transformative medicines. And we've now applied the Bar end-to-end across our R&D pipeline. And so that has allowed us to identify and prioritize assets that do have transformative potential, but also to attrit and deprioritize those that do not. And the Bar consists of 5 criteria, which I'll come back to actually several of them over the course of this presentation. And none of them are particularly surprising in and of themselves, but each of them must be met if a medicine is going to have transformative impact.

So first of all, the drug must answer a clear unmet need, and that need would have, of course, become addressable through advances in disease biology and understanding. And second, it must engage a foundational target. So that means a protein or pathway that drives disease pathophysiology or the salient clinical manifestations. The third is that the asset must possess worthy pharmacologic characteristics compared to other contenders that are out there. And fourth, it must be able to achieve meaningful therapeutic differentiation. And finally, the medicine must unlock a path to value.

And first and foremost, that's value for the patients being treated, but also value for the company that delivers the medicine to the world. And so adoption of the Bar end-to-end has now become business as usual across Roche Pharma R&D. Now one way to visualize the impact of the Bar is through our pipeline evolution at the NME level. And 2 years ago, our pipeline consisted of 81 publicly disclosed NMEs. So you can see that on the left. Today, we have 20% fewer NMEs, 65 instead of 81. But we've been quite active in sculpting our pipeline during that interval. So on the one hand, the initial application of the Bar in 2024 brought a fair bit of pipeline attrition, and we expected that. However, we've also achieved many pipeline additions either through our internal pipeline or from business development deals.

And those pipeline additions are starting to mature nicely, especially in our late-stage pipeline. As you just heard from Teresa, we're on track to progress quite a large number of NMEs into our Phase III portfolio in 2025. We've actually done 8 thus far, and there might be still other opportunities before the calendar year ends. So having applied the bar, although we have fewer overall NMEs today than we did 2 years ago, the quality of our NMEs seems to be increasing as we had hoped. And 2 examples specifically illustrate this point. First, the percentage of projects that we believe could have best-in-disease potential has increased to 67% of pipeline projects, up from 58% just 18 months earlier.

And also the projected average peak sales per pipeline asset has increased by more than 60% to CHF 1.3 billion. And in keeping with average peak sales estimates, overall portfolio value estimates have also increased significantly. Now one of the most important questions that we can ask is whether application of the Bar is improving our success rates. And of course, it will be several years before we know for sure, but one possible leading indicator is the impact on team assessed probability of technical success for our Phase III trials. And so here, we're looking at a plot of pipeline projects for which the decision to initiate pivotal trials happened prior to application of the bar.

And team assessed PTS is shown directionally on the Y-axis, adjusted net present value is shown also directionally on the X-axis. Now for obvious reasons, we're not disclosing the actual names and values. But the graph shows that several of these projects had either a team assessed PTS or an adjusted NPV that was lower than we would like to see. And in some cases, actually both of those things were true. So that -- those are the dots in the lower left-hand quadrant. All right. Now let's overlay the projects that underwent pivotal trial go decisions after introduction of the Bar. These are the dark blue dots. And you can see that a much larger proportion of these Phase III projects has a team assessed PTS within our target range, so above the center horizontal line compared to what we were seeing before application of the Bar.

And most of these projects have a much more robust adjusted NPV than what we were seeing before. And actually, even for the couple of cases where the PTS still seems maybe kind of border line, they're actually -- there are sometimes additional upcoming data cuts that could push the final assessment upward. So it's still early days, but we can already draw 2 conclusions from the results that I've just described. First, adoption of the Bar has been successful. And second, it appears to be having the desired effect, both on pipeline quality and the potential for future Phase III trial success.

All right. Another important R&D excellence solution was to reconfigure our portfolio management and governance so that it would be centered entirely around executing the Bar effectively. And implementing this solution has required multiple changes, one of which was to establish cross-cutting review boards. So these are panels of functional experts who provide critical assessments at key points along the molecule journey. So we have development boards. They focus on asset strategies and clinical development plans. And then we have business boards, which assess key commercial drivers and risks. And all of these review boards are jointly chaired by leaders from early-stage and late-stage R&D. So that ensures the end-to-end cohesiveness that we're looking for.

And these boards have been highly successful. So teams are benefiting from rigorous reviews, but they're also preserving the speed necessary to ensure that programs are competitively positioned. Now another change, which is summarized along the bottom left, involves the creation of what we call one asset teams. So once a molecule demonstrates proof of concept, so in Phase Ib, for example, we form a single team that will guide that asset through the remainder of its development and eventually its life cycle. And the team has membership from both early and late development and the membership evolves over time with the needs of the asset.

Now we've also created an enterprise-level pharma portfolio team that includes several members of the Corporate Executive Committee, including several people who are here today. And this group meets regularly to make holistic assessments of our end-to-end portfolio health from research all the way through late-stage development. And this can include composite assessments of the Bar. You can talk about volume at various stages, risks, strategic fit, et cetera. And these frameworks inform broader strategic considerations to shape the near-term or the longer-term shaping of the pipeline. Now this is the changes to portfolio management and to governance, this is all well and good, but how do we know that these are evaluating the quality of our decision-making.

And again, the ultimate proof will be in the delivery of transformative medicines. However, we can give you at least a flavor of the thought process around some recent pivotal go decisions. Now one thing I want to emphasize first off is a desired outcome of reconfigured end-to-end governance is to ensure that we generate early clinical data sets that are sufficiently robust to inform optimal Phase III decisions. So in the past, sometimes competitive pressures to move quickly made it challenging to fully derisk certain programs early in development. And so that meant that our Phase III pipeline was carrying extra risk, and we unfortunately saw that play out with more disappointments than we would like at times.

But going forward, one goal here is to generate enough data early on to learn not only whether to do a Phase III trial, but also how to design one that is maximally likely to be positive. So zilebesiran on this slide is a good example. Our partner, Alnylam, completed 3 Phase II trials of zilebesiran in hypertension. Now Manu is going to discuss the trials this afternoon, but it's actually unusual to have that much Phase II trial data available, but we were actually quite glad to have it. Now when the first trial, KARDIA-1 read out, we gained confidence that zilebesiran engaged its foundational target, angiotensinogen very effectively. So that's obviously a Bar criteria.

And then the second trial, KARDIA-2, taught us that zilebesiran could lower blood pressure meaningfully when combined actually with 3 different types of blood pressure medicines, but its effect was particularly pronounced in combination with a diuretic. So KARDIA-1 and KARDIA-2 taught us that zilebesiran had worthy pharmacologic characteristics, another Bar criteria. But there were still important unanswered questions, including the right dose, whether zilebesiran would work safely in combination with multiple medicines and the exact population of patients to study in a cardiovascular outcomes trial. So we ran KARDIA-3 to help resolve those questions. And in KARDIA-3, the zilebesiran effect on blood pressure was greatest once again in the subset of patients, who are on a diuretic.

But very importantly, we also learned that zilebesiran works best in people whose hypertension is clearly uncontrolled at the time of enrollment. And this actually makes sense because if your blood pressure is already adequately controlled, you don't really benefit from adding another medicine. And of course, if people whose blood pressure is already adequately controlled, make their way into a clinical trial, the risk of a type 2 error increases. So that means you falsely conclude that a blood pressure medicine is less effective than it actually is. So the big picture is because of the full KARDIA Phase II data set, we can feel confident in the ability of zilebesiran to achieve meaningful therapeutic differentiation when studied with the right treatment combination and in the right patient population.

In other words, the robust early clinical data here has taught us how to design a Phase III trial that has a good chance of working. Now prasinezumab is another example of the importance of having a robust Phase II data set to guide both decision-making and design of a confirmatory Phase III trial. And for context, there are no disease-modifying treatments for Parkinson's disease and really only a few options for symptomatic relief. So the unmet need is very high. Now the first Phase II trial we conducted, which is the PASADENA trial, was suggestive of possible benefit, but it was hard to be sure that those results by themselves weren't simply due to chance.

So we did another much larger Phase II trial, the PADOVA trial to further explore the signal that was seen in PASADENA, but also to assess how prasinezumab might perform in combination with medicines such as L-DOPA, which are commonly used for symptom relief in Parkinson's. Now Hideki is going to describe these data this afternoon, but the results were certainly supportive of the hypothesis that prasinezumab could bring a meaningful therapeutic benefit and the effect was even more pronounced in combination with L-DOPA. So when you take the entire Phase II data set together, prasinezumab shows a consistent efficacy signal. It looks safe, and we learned that combining prasinezumab with standard of care L-DOPA has a reasonable chance to show clinical benefit in a well-powered Phase III trial.

Now of course, early clinical data sets don't always need to be that large to be highly convincing, and trontinemab provides evidence for that. Pathologic beta amyloid, we all know, clearly, it's a foundational target. And yet the existing treatment options have still barely made a dent in the unmet need surrounding Alzheimer's disease.

Hideki, again, will describe the data for trontinemab later, but already even in Phase Ib, it became evident that trontinemab possessed worthy pharmacologic characteristics that offer the potential for meaningful therapeutic differentiation because we're seeing rapid and deep clearance of amyloid plaques, more than 90% of patients have become PET negative after several months. And the safety profile is encouraging, less than 5% incidence of ARIA-E after 28 weeks.

So -- based on the Phase Ib/IIa trial results, we already felt confident to ungate Phase III trials, both in early Alzheimer's disease, but also in the preclinical Alzheimer's disease setting, where we'll see if we can prevent or delay progression in patients who have beta-amyloid plaque deposition, but do not yet show symptoms of cognitive decline. So hopefully, these 3 examples make it clear that configuring our decision-making around the bar and ensuring robust clinical data are enabling a data-driven approach to design confirmatory Phase III trials that have a good likelihood of being successful.

All right. The next R&D excellence solution, I'll touch on briefly, involves accessing the best external innovation. And when the concept of the bar emerged 2 years ago, a vision that we had was that we would invest in candidate medicines that cleared the bar wherever they might be found, whether from our internal pipeline or the external biopharma landscape. Now for this to be feasible in practice, we needed to make several adjustments in terms of how we support pharma partnering across our organizations and vice versa.

And although we're certainly not done, it's fair to say that the business development results thus far speak for themselves. The clearest example, of course, is in our cardiovascular, renal and metabolism therapeutic area. And as you've already heard, this progress was, of course, anchored by partnerships with Alnylam and with Zealand and through acquisitions of Carmot and now 89bio pending successful closure.

But equally, our other therapeutic areas have also been enhanced as evidenced, for example, by the Zion and the Rigor -- Rigor and Poseida deals in oncology. And Poseida, by the way, also opens a potential door for allogeneic cell therapy in multiple sclerosis. And then we have acquisition of Telavant, which brought afimkibart into our immunology therapeutic area. And then there is the AntlerA deal, which boosted our ophthalmology pipeline.

And so overall, we're leveraging internal and external innovation to architect a pipeline that is enriched with candidates that can become transformative medicines. Now in my remaining time, I'll highlight our progress on the second overarching R&D excellence objective, which is to achieve top-tier productivity. And here, one of our solutions involves embracing ambitious productivity objectives. Now just to remind you, there are 5 key components of R&D productivity.

You have volume, value and success rates on the one hand, which defines the effectiveness of an R&D engine. And you have cost and cycle time, which describes the efficiency of the engine. And the bottom line here is that we remain on track towards our 2030 ambitions for each of these components. For example, we've added 29 assets to our pipeline since we launched R&D excellence, including 12 thus far in 2025.

I already described the increase in average peak sales for pipeline asset. On the efficiency side, we've reallocated approximately CHF 1.2 billion towards programs that meet the bar and also to other productivity initiatives. And we've done that while keeping our overall R&D investment flat since the end of 2023. And we're on track to achieve a significant acceleration in our end-to-end cycle times. Of course, we can't yet quantify Phase III trial success rates, but I've shown you that the team assessed probabilities of technical success have already increased substantially.

So based on progress across these 5 components, our R&D productivity seems to be moving in the right direction. Now the ultimate proof for increased R&D productivity will require a meaningful reduction in R&D spend per NME launch, and it will take years to show that conclusively. However, we do see some preliminary evidence that this might be starting to happen. In the graph on the left of this slide, we're looking at R&D spend per NME launch for Roche together with 11 peer companies, and we're looking at this across 2 adjacent partially overlapping 5-year windows.

And overall, across the peer group, the average spend per NME launch has increased from the 2018 to 2022 window to the 2020 to 2024 window. But Roche, shown in the middle of the graph, is pushing against that trend. If anything, our total R&D spend per NME has decreased slightly from one window to the next. And remember, R&D excellence didn't kick in until the last year of the second window. So this suggests that even a single year of productivity improvement has already had an impact.

Now the graph on the right of the slide illustrates that our internal reallocation has produced a markedly increased proportional investment in high-value R&D programs. Now I should point out that many of the so-called lower-value programs are actually line extensions, and we certainly need to be committed to line extensions to realize the full value of NMEs. But overall, we are starting to see qualitative and quantitative productivity gains from R&D excellence.

Now last year, we created an internal fast-tracking mechanism for a few programs with exceptionally high potential. And this, by the way, is not to be confused with the FDA, Fast Track Designation. But this approach has been quite effective for us. And you could see that the reductions in projected time to filing range from 6 to 9 months to as much as 21 months compared to original expectations. And there are a wide range of interventions across these programs that have contributed to the acceleration.

And I won't read through them all here, but they're listed on the slide. But what's really nice about this is that the learnings from this program are becoming best practice across our pipeline regardless of whether the program receives a special designation. All right. There are several additional ways in which we are evolving our underlying R&D engine to make it more efficient and more cost effective.

For example, resource optimization. So a moment ago, I mentioned the extensive reallocation toward higher-yield investments over the past couple of years, which amounts to more than CHF 1 billion since the beginning of last year. Now we've invested over CHF 600 million of that savings to fund programs that we brought in through business development. And then an additional CHF 70 million was deployed to ensure pull-through of the Fast Track programs that I just described.

But furthermore, CHF 370 million has been deployed towards future capabilities, some of which I'll describe further in a moment. But all of this sums up how we have aggressively funded programs that clear the bar, all while keeping our overall R&D spend flat during the same interval. Now a year ago, we described to you how we were radically simplifying our outsourcing model to help make our clinical trial infrastructure more efficient.

And 12 months later, that effort is starting to bear fruit as measured by accelerations in trial start-up times or improving the site experience and annual cost reductions, which include CHF 100 million of savings since 2024. And we expect in the fullness of time that this will deliver up to CHF 300 million in annual savings by 2030, and that will provide additional resources that can be reallocated to support future transformative programs.

Now lastly, I'll say just a few words about how we're assembling a systems and data foundation to enable future R&D productivity enhancements. And -- as you might imagine, we expect generative AI, Agentic AI and other automated platforms and analytic platforms to have a significant impact really across most elements of drug discovery and development. And one specific use case from the development side involves content generation at scale. By now, it's self-evident how this kind of work can be enhanced markedly by generative AI.

And so what you're seeing on the slide are multiple content generation elements where AI is actively being deployed to produce efficiency gains. And of course, one can step back at a meta level and orchestrate all of this with -- through agentification, for example. But now we're moving far beyond content generation. We're mapping out many diverse workflows across our drug development continuum that might benefit in the future from Agentic AI applications.

So in the fullness of time, these types of efforts could bring many efficiencies, and they also should augment the capabilities of our broader workforce. Now we have many AI and machine learning initiatives ongoing on the drug discovery side as well. And one of these, which is called lab-in-the-loop has been championed by Aviv Regev, who leads gRED. And the key concept behind the loop is the iterative learning that becomes possible when, for example, screening data or perturbation-based data at scale is leveraged recursively to train models.

And in turn, the incorporation of such models can enhance target discovery, target enablement, lead identification, lead optimization and other downstream efforts. So overall, these applications should lay the groundwork for a data foundation that can help drive improved productivity across the R&D continuum. All right. So in summary, 2 years into our R&D excellence initiative, the majority of solutions seem well along the path to being embedded in a new business as usual. Now we still have a lot more work to do, and some of these are still closer to the beginning than to the final outcome.

But we believe this initiative could push us to top quartile R&D productivity by 2030. But most importantly, R&D excellence is really fortifying our pipeline with many candidate medicines that truly have transformative potential.

So with that, I'll stop and either we'll have the fire alarm or we'll have some Q&A first.

So I would ask -- as we have now 40 minutes for Q&A. So the speakers did an excellent job in keeping the time. We will be interrupted by the fire alarm at 11. Please stick to questions which are -- were covered in the morning sessions. You still have the opportunity then -- for drill down on details when it comes to pipeline or to data questions. Start here before Richard, please.

Richard Vosser from JPMorgan. Maybe one question on the bar on zilebesiran and then one other. So on zilebesiran, you gave an example of all the criteria at the bar, but I'm wondering about competition. When I look at KARDIA-3, the blood pressure lowering is a little bit lower than some of the other agents, maybe the aldosterone inhibitors.

So just how do you factor in competition when you're looking at that to go into Phase III and push forward? And then one on R&D spend. Clearly, very good reallocation over the last few years, but we're looking at a significant step-up in Phase III development going forward. So just thoughts on how R&D spend can develop from here? Are we looking more at low single digit or growth in line with sales from now on?

So maybe I'll start. And of course, we'll have a lot more in-depth discussion on zilebesiran this afternoon, but a couple of high-level points on the competition. So the first point, and actually, I'm glad you brought up because it's worth mentioning the entire premise of zilebesiran is that you can dose it twice per year -- and because it's an RNAi, you get 24/7 coverage of blood pressure for that twice per year dosing.

So including nocturnal control, which is something that orals -- so the whole issue is with oral. So one thing is, of course, the -- you don't necessarily get nocturnal control, you don't get that 24/7. But the other issue is the adherence. A major reason why many patients are not controlled is because it's hard for them to take 2, 3, 4 orals multiple times a day. So we think that, that will set zilebesiran apart. So -- but we'll talk more about that this afternoon.

In terms of big picture R&D investment, certainly, as you've seen, I think one important lesson for R&D excellence is that we have now really worked the discipline around making the trade-offs that are needed, but then also maintaining overall fiscal discipline. So that will continue to be a major effort here -- to a first approximation, the idea is to make the trade-offs that we need within R&D or within other elements of the P&L, but to maintain tight margin discipline. And I'm sure Alan will have more to say about that this afternoon.

Yes, Sachin Jain, Bank of America. One question on obesity and then one on the pre and post Bar. So on obesity, since you did the Carmot acquisition, the commercial landscape seems to have changed a fair bit as we observed it. So just two questions. One, the price point at which you expect to launch at, has that materially changed given the changing landscape? And secondly, how do you think about the split of the eventual market cash versus reimbursed given that is also changing?

And then on your blue dot sort of bar, pre and post Bar, I'll chance my on this one pre-Bar asset, which is in the upper right quadrant, both large peak sales and high probability, any sense of which that is? And then the one I'm more interested in actually is the post Bar, you have one in the bottom left quadrant, so small and low probability, why would you do that?

Yes. Thank you for that question. The first part being the evolution of the market since Carmot acquisition. I think we all agree that this is a very volatile and evolving market, but we are way within the bookmarks that was part of the initial acquisition. So the price point and the volumes is within what we have anticipated even though it's evolving faster. To the split between cash and reimbursement, I think that will have significant differences across the different geographies.

And you will see all three ways of funding AOMs continue to grow with a growing market. So we are building in optionalities in our go-to-market plans, ensuring that we can cater for different future outlooks and make sure that we only make the decisions when really needed to ensure that we can meet the patients where they are and tap into whatever health care system evolution that we will see over the coming years.

Great. So on the pre and post Bar -- so I'll just -- so first of all, for obvious reasons, we're showing directional to give a flavor of how the decision-making has changed. So we're not going to comment on exactly which dot was where. But I will say that one of the programs that we moved into Phase III is an antibiotic. And we do that because of our commitment to society. It doesn't necessarily fit the path to value in the same way that the others do. So I'll leave that with you as you will.

Matthew?

It's Matthew Weston from UBS. Two questions, please. Obviously, Teresa, one thing you highlighted was the excitement around the giredestrant press release this morning with evERA. And in the PR, you highlight that you saw efficacy in the ESR1 subgroup and in the total population. Can you give us confidence that there is actually some efficacy in the non-ESR1 patients or whether or not it's just those mutant patients that are driving the efficacy?

And I ask because we've increasingly seen the FDA focusing on subsets of patients and only giving a label for where the efficacy is. And then the second one also for you, Teresa, was Vabysmo. And again, you highlighted in your comments the strong market share and market leadership. One thing we've all been looking for from Roche is contribution to the patient Charitable Access Foundation program so that you can cement that, help patients get access and bring Vabysmo to more people.

Can you give us any comments as to when we can expect Roche -- can we expect Roche to contribute to the cap? Could it happen in '25? And would you see that as a way of accelerating Vabysmo uptake?

Sure. So maybe I'll start with the cap question and then Levi, I'll let you answer the evERA question. So I'm unfortunately going to give you a somewhat unsatisfying answer to your question because as you all know, the -- as part of our charitable contributions, we do provide donations to co-pay assistance funds. However, those donations are made by a separate part of the organization and are completely firewalled from the commercial organization.

So it would be inappropriate for me to comment what their plans are and how they plan to contribute because that is -- it is and has to be separate. That having been said, I think what we have previously communicated is that we will always attempt to do what's right.

So what I can't do is comment specifically on the results of the evERA study. But what I can -- I mean, they'll be presented at an upcoming meeting, and we look forward to all that. But what I can say is, at a high level, what we -- everything is on thesis to what we have said actually multiple times, including I, myself, have said at this event in years past, which is that if a breast cancer is dependent on the estrogen receptor, giredestrant can perform.

And that continues to be the case. And that is the case regardless of whether or not there happens to be an ESR1 mutation. So that is all on thesis. Now of course, as you go through later and later stages, by definition, you can have reduced dependence on the estrogen receptor. But if there's dependence on the estrogen receptor, giredestrant can perform.

James Quigley from Goldman Sachs. I got 2 questions, please. So first of all, on the balance between internal R&D and external R&D. So of the 65 assets you have in the -- or 65 NMEs in the pipeline, how many of these were externally sourced? And how has that changed relative to Q3 '23, I think it was -- you had 81 assets in the pipeline. How will this then change or how will this develop further over time? And what does that mean for your internal research spend? That's question number one.

And question number two, picking up on the comment, Teresa, on new areas and Roche has done this before. And across MS, across Hemlibra, and obviously Vabysmo, I think it's fair to say with those 3 assets, you are head and shoulders above the competition. So moving into those areas maybe was made a little bit easier from that respect. So as you move into obesity, as you move into CVRM, are you confident that you have the same head start that you have relative to the other competition?

Yes. So I'll start. So I don't -- it may be that Bruno has the exact numbers and can speak to those. But I can speak kind of higher level, which is that, as I mentioned in my talk, the concept of the bar is that we will go after assets that clear the bar wherever they are. And it's great if they're internal. But if they're not and they're external and they clear the bar and they otherwise are fit, then we'll go after those.

So to a first approximation, we haven't been sort of keeping tabs and using that as a guidepost to whether we go internal or external. Obviously, we certainly -- by the way, giredestrant is a homegrown molecule. Inavolisib is a homegrown molecule. Trontinemab came out of [indiscernible]. So we're very pleased with our internal pipeline and its ability to generate medicines like this, but we're not constrained by that to a first approximation in terms of where the exciting medicines might come from.

But the ratio is very stable at 60-40. It has not changed despite the turnover. It's 60% external, 40% internal. What has changed, though, is that we have more late-stage assets. So therefore, it might become more visible. Normally, we have a lot of stuff which gets onboarded as preclinic or Phase I and you will have recognized.

And the question on best-in-class obesity, would one of you gentlemen like to do that?

I'm not sure I fully got the question.

So the question was in -- when we entered prior disease areas, we have entered with really best of disease...

So we are having a growing portfolio of late-stage assets that we believe have the best in disease and can be differentiated when it reaches the market and the patients. We will do that based on some of the capabilities that Teresa talked to through combination opportunities and by being determined and curious about this growing market. So our belief is that many of the late-stage assets we have will translate into differentiated assets that have a clear role to play in the evolving obesity market.

And I think if you sort of break them down more specifically, certainly, CT-388 has potential, petrelintide absolutely has potential to be best in disease. Pegozafermin -- it will roll off the tongue next time, I promise. Again, for MASH, it is the first to move into the Phase IV -- I'm sorry, the F4 population, which is the most serious form of MASH. So I think we have assembled a pipeline which both in monotherapy and combination has the potential to be truly differentiated.

Conscious I may have the fire alarm slot, but I'll start. It's Luisa Hector from Berenberg. I have a couple of questions really on timing and efficiency. But starting with your 3 fast-track assets, which you highlighted last year, could you say a little more on how they have been accelerated versus other assets? And is 3 the right number? Might we expect to see more?

And then on efficiency and timing again, you have 8 new NMEs entering Phase III this year, but half of those don't start Phase III until next year. Is there a reason for this? Is this perhaps not the best way for us to gauge timing, speed to market, et cetera, as you push through with the efficiencies?

So to the first question, there we chose 3 programs that we felt like had particular exceptional potential for impact for the -- obviously the value of the company, et cetera. But actually, it's not as though those are the only 3 where we put all of our resources into. Actually, the learnings, as I mentioned, really -- I mean, we have many programs that are moving robustly. But I will say that it was highly instructive to pick 3 and to ask our teams to really think outside the box of what would it take to accelerate.

And there are a range of -- so there's not one single thing, but there are a range of approaches that we're taking. So they range from, for example, very tight partnership between our medical affairs colleagues and our clinical colleagues in parts of the -- in geographies where we really took that to another level.

In other cases, there were close collaborations on the manufacturing side to make sure that things could be gated appropriately. In other cases, it was about can we take interim looks and have discussions with regulatory authorities and move things. So there are a whole series of creative approaches that led to the acceleration. Now the second question that you asked, can you just remind me the...

Yes. So we're one of the fast track assets and then 4 of the 8 NMEs you highlight for Phase III don't start until next year. So why?

So actually, historically, internally, we used to get very excited about when is the first patient in to a study. And you would then move heaven and earth to get the first patient in as quickly as possible. But actually, it turns out that's not the relevant metric because very often, you could say, okay, great, we've got the first patient, and then you get through 2 or 3 months before you've enrolled enough sites.

So really, the relevant metric is when do you get a critical mass of sites stood up. Once you have that, the hockey stick -- you get to the long end of the hockey stick. So that's really what we're focusing on. So the time that the Phase III trial starts is less important than the time that enrollment finishes.

And I just want to double down on something Levi said because I think it's really important. Every single one of the fast-track assets, those teams are actively sharing their learnings back with all of the other teams. And so just the synergistic effect on the portfolio is actually pretty cool to see.

And maybe one to add on just to avoid confusion. When we show the appendix slides, for example, with the trials listed, then that -- we put it in there when it's first patient in. So what we have communicated today is when we have taken the internal decision that we will move an asset on. So there's just a couple of months, 1 or 2 quarters in between normally.

It's Michael Leuchten from Jefferies. Just if I could go back to the obesity question in terms of channels. Given the interdependence of pricing in the U.S. between the different channels, government, commercial and cash, how much flexibility really is there in the DTC channel? Like is there a scenario where you could price in the DTC side really low, say, $100 and then protect pricing in the other channels? Or is that not possible because of the set interdependence?

And then a question for Teresa on NXT007, [ 5 billion peaks ] on the slides. Can you just talk about how that fits with Hemlibra? Is that total cannibalization? Is that additional revenue potential in pockets? Like how do we frame that 5 billion versus what Hemlibra does deliver?

I mean -- maybe I'll just take that one to get it going. So in terms of 007, that we would anticipate cannibalizing all of Hemlibra. I mean if we can beat Hemlibra, which is the current standard of care, we would clearly want to shift patients on to the most appropriate therapy. I think we should also remember there remains a significant number of particularly more mild and moderate patients that are not yet on prophylactic therapy. So I think when you have a drug that potentially has the profile of NXT007, not only do you sort of take the Hemlibra share, but you can then also very credibly think about expanding your market.

I think it's probably too early to comment on the exact price point. So I'll refrain from that. What I do say and I would like to comment on is the ability to also differentiate in a cash paying segment. We clearly see that their willingness to pay also for premium solutions in that channel. So we will maintain, as I started out saying, the optionality to make sure that we are ready to offer where patients expect us to be.

Sarita from Morgan Stanley. Just a question on prasinezumab, and you mentioned the Bar across 2 trials, both of which didn't meet the primary endpoint, we saw a variety of findings, better efficacy in L-DOPA patients, better efficacy in MAO-B-treated patients, diffuse disease and then running the trials for longer over 104 weeks potentially.

So what drives the confidence that these findings across 2 trials, various findings are not due to chance? And then a second question on fenebrutinib in PPMS. Could you confirm if you've changed the endpoint to noninferiority to Ocrevus in PPMS? And are you able to do so, so close to trial ending?

So for prasinezumab, the -- so as I mentioned, after the first study, we were not convinced that the findings that we're seeing were due to an efficacy versus chance. But after the second study, which was much larger -- so it was much larger, number one. Number two, it was corroborated with additional biological correlates of activity of prasinezumab.

So then it looked like -- this is probably a real finding, and it's just a matter of can you power the study appropriately in Phase III. So hopefully, that addresses that question, but we'll talk a lot more this afternoon about the data. The second -- to the second question, we cannot comment right now on discussions around what we might or might not do in PPMS within.

Emmanuel Papadakis, Deutsche Bank. Maybe just a couple of questions on your obesity strategy. Just interested to understand how you think the future market will segment in terms of oral injectable? And then also in terms of incretin amylin monotherapy versus combinations, I understand there's many things still to be determined, but how are you expecting that will play out and how are you going to reflect that in your clinical development program, particularly in terms of time lines?

Yes. Thanks for that question. I think Teresa with her distribution of BMI clearly shows that there are two ways to truly expand the market, is treating more of the patients that are currently receiving GLP-1s or GLP-1 incretin market. And then the big opportunity to address an unmet need is to also engage patients much earlier in the disease. Teresa called out that the average BMI for patients eligible and using AOMs today is 39.

So it means that we have almost 10 BMI increments where it's not sufficiently being used. Why is that? That was a heat map in terms of what is important to the different patient segments. And there, clearly, we see that those with lower BMI in the what we named the label prevention or early treatment, they value the tolerability and the convenience and are ready to trade in some of the last few percentage points on efficacy.

We see both amylin with petrelintide and a more tolerable profile to have an opportunity for monotherapy there as well as in combination with incretins and then play in the high efficacy. But we definitely also see the opportunity for a significant share of the market, you can say, having a preference for daily orals over weekly injections. The exact share, I think, is difficult to comment on. But I think what we need to remember is that this market will be significant. So even a small share represents a large number.

So there's plenty of opportunity for oral GLP-1s to play in the early space and into the -- into maintenance potentially for the amylin and petrelintide to establish itself do the combinations with incretins, et cetera, the market will be very heterogeneous and there's not one solution that will fit all patients they are very different.

Maybe just to add one factoid on top of that. I think we generally -- sometimes we assume that patients just prefer oral treatment. But many patients would actually prefer a once-weekly injection to an everyday oral because it's just more convenient for their lifestyle. There's actually only a relatively small subset of patients that are truly needle phobic. They just won't touch a needle.

The vast majority, particularly if it's a simple subcut injection, are more than willing to do that, and it just is easier than taking a bunch of pills. So we shouldn't always assume that oral is necessarily going to be the winner. I think you're right. It's just going to be a mix depending on preference.

And I think mono is also coming back with a little more details on how we see that market pan out.

There's one over in the corner. I don't know, but he's behind you...

Naresh Chouhan from Intron Health. Just one going back to the R&D spend, please. You've got one of the highest spends in the industry. You've just shown that you've cut some of the assets and want to reduce cost per NME and drive efficiencies from time lines and AI. Help me to understand why we shouldn't expect R&D to fall in absolute terms, as a victim, please? And then secondly, on your obesity programs, how is the evolving landscape -- or has the evolving landscape -- commercial landscape reduce the size of the clinical program you anticipated maybe a year or 2 ago?

Maybe I'll start on the R&D side. So one major consideration over the next couple of years is the fact that we have indeed moved already 8 NMEs into Phase III. Some of those Phase III programs could be robust, including CT-388 and we haven't moved petrelintide, these are significant programs. And so we moved 8. We expect maybe a few more in the next several months. So that we -- in order to make those successful, we have to invest behind them. So that will be an important consideration that will require some additional fiscal discipline and trade-offs over at least the next 2 or 3 years.

And maybe on the perspective of how does the perspective of the commercial ecosystem evolve is obviously, we are observing what happens. We know that patient segments start to fragment and has very different patient preference and our CDP. So clinical development plans actually reflect this. And we'll use the portfolio we have to do the right combinations based on the patients safety observing emerging already there.

Luisa Hector again from Berenberg. Maybe to follow up on obesity. Have you debated simply trying to get to market on weight loss data as quickly as possible? And how critical are the CVOT, the comorbidities? And will they just be assumed as a given, given other drugs have that data?

Yes. So clearly, one of the lead indications will normally be chronic weight management. That's where the guidance to industry is relatively clear in terms of what is expected. And you will probably also see in many of our assets that that's where we will have the first indication. Expanding into adjacent TAs and indications really also a function of where do we see that product play a role in that evolving landscape.

If we are in the high-risk segment, you are expected to also demonstrate the benefits on comorbidities, whether it's CVOT or into MASH or chronic kidney, et cetera. So there, you will probably see a growing, you can say, pool of evidence to support commercialization of those assets. Whereas in the other end of the spectrum, it might be that it is not the, you can say, cardiovascular metabolism that will drive uptake and preference for a certain product, but actually more being other endpoints.

And we are exploring what could those other endpoints. And that's where we believe that many of the capabilities from other parts of our end-to-end DAs and TAs will help us make sure that we can make some strategic choices that will set us apart.

May we capture some questions over there before we do a second round?

Justin Smith from Bernstein. Just quickly on the BTK, Teresa, you used the word disruptive. Was that with reference to the entire market or just to tablets? If it's both, could you just help us understand where you would expect to take share from biologics and ABC?

Yes. I mean I think that fully depends on the data, right? I mean if we see -- if what we see in the Phase II plays through to the Phase III, certainly disruptive in the oral space, but potentially disruptive in the IV and subcut space as well. It looks like at least in Phase II, the efficacy is looking pretty good.

And today, just remember today, more than 1/3 don't have access to disease-modifying therapies in that space. So if an oral that would allow way more access for those patients.

Yes, absolutely. Good add.

[indiscernible] from ZKB. So also on obesity, you talked about that patients are willing to pay a premium also in the cash pay market. Do you have already some understanding of what is really needed that the patient is willing to pay a substantial premium to like cheaper solutions? And then a second question more on strategy. Would you assess that most of the internal transformation work is over now and that you are in full execution mode, but also could mean that we could see an acceleration of external deals being announced in the future?

Can I go first?

Yes.

Yes. Okay.

So I think we should try to look 10 years back and see what was the cost for medicines and relative to effectiveness of those treatments and where we are today. And that means that the price point is coming down per kilogram weight loss, et cetera. That gives room for optimizing the treatment journey through a seamless engagement and patient support solutions that could be provided by some of our colleagues in diagnostics and other places that would give a holistic approach to care and thereby also warrant a premium to a generic version of an alternative product.

I think there will be significant opportunities to explore that in the years to come. And I think we have only seen the beginning of how patients will engage with telehealth, online pharmacies and direct-to-consumer channels.

And maybe on the internal transformation question that you asked. So certainly, there are several solutions that I described where I would say -- I mean, done is always a tough term, but we are very much into the embedding into -- this is just how we do business. So those things, the cycle kind of -- the circle comes to closure. But there are one thing that if Thomas were here, we always -- he would remind us that productivity is never done. It always continues. And actually, there are still real opportunities that we see to continue to achieve productivity gains, and we're going to go after that. So that's the first point.

The second point on the internal versus external balance certainly, we are -- again, we're happy with the progress that we've made, including how harnessing these kinds of capabilities can boost our internal productivity. But it's also the case that the amount of external -- the amount of innovation in the great white world is going to be far greater than what we can do in-house. That is one thing that's not changing. So I would expect to continue to see a balance between external and internal.

It may be that we don't have to do as many late-stage deals as we have done, although if we see something that clears the Bar, certainly, we'll evaluate it. But certainly, across the continuum of our pipeline, we would continue to have external augmentation.

If I may just add one comment to your question. Obviously, we're talking about getting ready for obesity portfolio or the future portfolio. That work is at full swing. We are laser-focused to make sure we understand how this market evolves and we get the capabilities in place, as you've seen in the slide. So you could argue there's still clear execution there, but there's still also capabilities to build.

Any questions from that end? Yes, please add on.

Yihan Li from Barclays. So the first question will be on your obesity franchise. So a message from EASD conference last week is that the potentially involving trial design standard with future obesity trials might likely require active comparator arms, for example, like semaglutide or tirzepatide versus the placebo. So I know it is still in early days. I'm just wondering like what you think about it? And any implications of your trial design as you are aiming to achieve the top 3 players in the obesity market?

And a second very quick question on your oral SERD. So congrats on the top line this morning. But we note the persevERA, like first-line data is now pushed into the first quarter next year. So just wondering any commentary on this delay, please?

Yes. So I'll just briefly take the second one. All of these studies, given that they're event-driven, we have to wait for events to play out. So there's nothing particular to read into that, except that it's event-driven. On the first point, I would say we'll have time for a robust discussion on this in the afternoon.

But suffice it to say, on the one hand, there's not a categorical requirement that we do studies against active comparators. But certainly, if it becomes appropriate to do that, we are prepared to do that in certain settings. But we'll talk more about that this afternoon.

Then we do a second round here in the front row?

It's Matthew Weston at UBS again. Teresa, MFN.

I thought I was going to -- you were so close.

You're not going to get away with it. I'm sorry. President Trump's letters required or requested a response from the pharmaceutical industry by the 29th of September. I'm not expecting you to tell us what Roche will say. But I would be very interested in your view as to what investors should expect over the coming months.

Do we think that this is a period where the industry has a clear view as to what it can offer or companies have a clear view as to what they can offer, I should say? Or is this a period where investors should probably expect kind of noise, volatility and some ongoing uncertainty before some very difficult or different views come together in a compromise in due course?

So I think we are all aware of the significant conversations that are going on in the wider world today about MFN. And I'd just like to say a couple of things. One, which is just to reiterate some comments that I made earlier in the morning, which is that we, as Roche and Genentech, are committed to the U.S. We have a substantial presence there. We have over 25,000 employees. We have 13 manufacturing facilities, 15 R&D centers. We've made a $50 billion committed. We are committed to the U.S. market.

We have also said over time that we share the goal of making medicines more affordable for U.S. patients. And yet, I think we all know that the complexities of the U.S. system mean that about 50% of the cost of innovative medicine actually goes to a middleman. And so I think the industry is in active conversations as are we with the administration to try and define what passes forward actually allow us to solve the ultimate goal, which is to help make sure that more Americans have access to the medicines that they need.

And is it going to be noisy or it's all going to become clear?

I think it will all become clear.

Sachin Jain from Bank of America.. I'll take 2 on topics we've discussed. So one, a follow-on from Matt's question around evERA. I guess the simple question is, has it changed your view probability success of persevERA, I guess, was the background to the question. The reason I'm asking it is you say whilst the data is on thesis, I think there's been mixed messaging on probability of success of persevERA across the C-suite.

So just interested in your thoughts there. And then on FGF21 and 89bio, I wonder if you could talk about that relative to Bar, again, following to James' question, a lot of commercial assets, competitive landscape at the time of your launch. How do you think your asset is differentiated?

So for evERA, I would say that the results that we're seeing reinforce the thesis, the biological rationale that we have believed in and talked about on many occasions for giredestrant. We can't comment on a particular probability of success. But certainly, we're very pleased with the data, and we look forward to presenting it. So that's what we can say now. On FGF21, certainly -- I mean, I can speak to the scientific criteria. Maybe I'll let my colleagues speak to the business criteria. But certainly, from a scientific standpoint, we have a very large unmet need in MASH. And actually, that will be true regardless of how -- of the penetrants of GLP-1. We expect that to be -- remain a very large market. And actually, now that there are therapeutic solutions coming out there, the diagnosis of MASH could actually grow. So the addressable market could actually grow effectively. So clear unmet need, a large opportunity.

The FGF21 data itself, that was, of course, a linchpin for us in terms of being enthusiastic about this opportunity. The Phase II data that was seen, the effect on fibrosis, the ability to achieve resolution. Certainly, that data was for F3, but then also we know that the class can have activity kind of across the board, including F4.

So that's a therapeutic differentiation that we don't really have, particularly F4 is that the GLP-1s are not likely to play a major role in F4, and we can discuss more on that with Manu this afternoon. So from a scientific criteria, it actually checked all the boxes from a Bar standpoint. And maybe I'll let my colleagues speak to the business side.

Which makes it relatively easy because the fact that this is a potentially transformative medicine, which has a huge potential for patients, clears the Bar. And if you see just the numbers, the market right now is like CHF 2.5 billion approximately. We expect it to go easily above [ CHF 10 billion to CHF 20 billion ] in the 2030. So there's opportunity for patients there.

And maybe just to reiterate -- I mean, I think the incretin solve or attempt to address this sort of the -- through the -- addressing the underlying metabolic condition, the FGF21 analog actually goes straight towards the inflammation of fibrosis. So it's a very different way of actually attacking the disease.

Maybe also to let me add here again, we have in the appendix the updated epidemiology slides and -- based per molecule. So there you will see, for example, for giredestrant, all the different lines of treatment with driving mutations on top. You will find the estimate for what zilebesiran could do as the Phase III trials are designed, what patient population they would go after. So it's maybe worth to have a look there as well. And for MASH, I think we'll find something as well. And assumptions in there, of course, that this will change over time, the standard of care will change.

Richard Vosser from JPMorgan. Just a question on the devices. You mentioned the pilot -- the pilot facility coming on -- coming or being built now. So what comfort can we have that there will be devices for the GLP-1 and amylin franchise, given we've seen, obviously, that's been the biggest bottleneck for current players. I can imagine [ NX07 ] , you can have a device for that for the future, but just thoughts there. And maybe aligned to that, obviously, [ $750 ] million, I think, spent on a peptide facility in the U.S. You will have CDMOs, but thoughts on that capacity as well.

Yes, absolutely. So I think we've always said that with CT-388, our clinical trial and initial launch will be a balance of internal and external supply. That remains the case. We will ultimately bridge to primarily internal supply chain eventually. But that really was the purpose of building Holly Springs was to make sure that we had that large volume throughput device facility. And so I think you can be very confident that we will be entering with devices in the places where devices are necessary.

Maybe we take a final question before heading to lunch.

Kirsty Ross-Stewart from BNP Paribas. Just on the FGF21 deal that you've done, does that change the nature of your collaboration with Zealand given their interest in MASH and [indiscernible]? Just interested on the logistics there. And then just maybe a clarification on R&D spend. You've highlighted multiple new Phase III programs requiring investments.

So can I just clarify what you mean and what you can commit to in terms of R&D spend on an absolute basis? So can the current rate be maintained? Or is it fair to expect this to grow in absolute terms over the midterm?

So on the -- on the FGF21 -- sorry, can you just repeat the first part of the question?

The nature of the partnership with Zealand?

Yes. So high level, we are really excited about our partnership with Zealand, and we're excited about the FGF21. We actually think that there's certainly no fundamental disruption or alteration. If anything, it just opens additional opportunities, including additional combinatorial opportunities. And actually, that gives me a chance to emphasize something that I didn't in the last question, which is that the characteristics of pegozafermin, very safe. So a little bit of nausea diarrhea, but not a lot of dose-limiting toxicity.

So that implies -- we don't know for sure, but that implies the possibility of combinability with other assets. So we think, overall, this is a net add for all parties. Now on the R&D side, I think once again, we -- this is probably not the place where we can say specific numbers. What we can say is, number one, we are -- we've gone out and done these deals. We've made these investments. We've applied the Bar.

So we're certainly committed to driving these programs through Phase III. And we're also committed to fiscal discipline, and we have evolved practices to do that. We see other opportunities to do that, but we'll work together to make that happen. I don't know, Teresa, if you want to say anything from a P&L standpoint on that?

Yes. No, I mean, I think...

And we have been saying before, we want to defend the margins. And I think there's more room for reallocation of resources. So you should not expect because we have now a wave of Phase III studies that you will see the R&D expenses go down.

I mean, as Levi pointed out, we've been planning for this. So this isn't...

Okay. I think with that, we end our first session, and then we're heading into the 50 minutes lunch and then see you later.

Okay. Thank you.

Thank you.

[Break]

To lead the hematology/oncology product development. And I'm here to kick off the afternoon session where my colleagues and I will be sharing with you the highlights of the portfolio across therapeutic areas beginning in my case for oncology and hematology. Roche has a long history of transforming the therapeutic landscape and improving the lives of patients with cancer and blood disorders. Today, our strategic pillars include advancing best-in-class precision medicines such as our PI3K inhibitor, Itovebi and our KRAS G12C inhibitor, divarasib. We're advancing high-impact combinatorial therapies. And as you'll see shortly, we have a growing number of combinations to address non-Hodgkin's lymphoma as well as our growing portfolio of combinations in hormone receptor breast cancer.

We're investing in new technologies and novel modalities such as next-generation off-the-shelf allogeneic CAR-T therapy as well as molecular glue degraders. And we're focused on seamless end-to-end investment from discovery science through commercialization in key disease areas, namely lung, breast, hematologic malignancies and hemophilia to deliver differentiated, high-impact, high-value therapies to patients across the globe. Our organization also has a strong legacy of pioneering new drug targets and technologies. Today, we have a highly diversified portfolio in hematology/oncology, ranging from small molecules, antibodies, antibody-drug conjugate, fusion proteins, peptides, gene therapies and next-generation CAR-T.

We also recognize that innovation in cancer therapeutics requires a constant hunt for novel modalities being developed externally. Highlighted here are some of the key deals we've made over the last couple of years mentioned earlier, the CDK portfolio acquired from Regor, the HER2 TKI partnered with Zion Therapeutics, the CAR-T program through our acquisition of Poseida and antibody-drug conjugates from MediLink and Innovent. This broad array of technologies is reflected within our portfolio of agents in development for solid tumor malignancies from Phase I through registration with a particular focus on the end-to-end development in lung and breast cancers.

Turning specifically to breast cancer. Our organization really is proud to have launched the field of targeted therapy for breast cancer in transforming the lives of patients with HER2-positive disease. We continue to address the unmet needs of patients with metastatic HER2-positive breast cancer. Specifically, 50% of patients with HER2-positive disease will ultimately develop brain metastasis. We're therefore advancing ZN-1041, a highly potent, well-tolerated brain-penetrant HER2 TKI that has shown high response rates, both systemically and within brain metastases in Phase I investigation.

Today, though, a major focus for our organization, as you heard earlier, is hormone receptor positive breast cancer, which represents 70% of patients. There remains an important unmet need to reduce the risk of disease progression or recurrence and foster greater tolerability of these therapies. We're tackling HER-positive breast cancer by targeting well-validated signaling nodes that are critical for tumor development and resistance with our best-in-class SERD, giredestrant, our recently approved PI3K inhibitor, Itovebi, and our acquisition of the next-generation CDK inhibitors from Regor Therapeutics. These novel CDKs have the potential to become best-in-class molecules, overcoming resistance mediated by CDK2 adaptation and reducing toxicity often mediated by CDK6 inhibition. Our lead molecule, GDC-4198, is a potent inhibitor of both CDK4 and CDK2.

Among patients who have progressed on a CDK4/6 inhibitor, 4198 monotherapy demonstrated durable objective responses with acceptable tolerability, favorable PK profile and sustained target coverage of both CDK2 and CDK4. And at upcoming scientific meetings, we look forward to sharing our ongoing work and plans for this program. As I alluded, our strategy in HR-positive breast cancer focuses on 3 fundamental pillars: next-generation endocrine therapies and CDKs and novel targeted therapies. Within these 3 pillars, endocrine therapy, as you know, remains the backbone of treatment for HR-positive disease across both early and metastatic disease until, of course, resistance develops. Despite progress in endocrine therapy, there remain limitations. In the adjuvant setting, 50% of patients discontinue due to tolerability issues and 30% will go on to develop progressive metastatic disease.

Our ambition is to improve the standard of care across all 3 pillars through novel agents that both meaningfully improve efficacy and reduce toxicity. Let me turn now to our next-generation SERD giredestrant as well as our PI3K inhibitor, Itovebi. As compared to other SERDs currently in development, giredestrant demonstrates the highest potency, is combinable at full dose with CDK4/6 inhibitors and has not been associated with ocular toxicity, an important adverse event for patients now seen in up to 20% of patients treated with an alternative SERD in development. Our development plan for giredestrant aims to replace current standard of care endocrine therapies across early and metastatic disease as the next-generation best-in-class endocrine backbone. We expect first results, as you know, from the Phase III persevERA trial in first-line endocrine-sensitive disease in early 2026.

We anticipate an additional readout in the first-line endocrine-resistant setting in 2027, which, by the way, allows physician choice of CDK4/6 inhibitors. Additionally, we are the leading SERD trial in early disease with a head-to-head comparison in the lidERA trial comparing giredestrant to an aromatase inhibitor in the adjuvant setting. We're also evaluating combinations of giredestrant with CDK4/6 inhibitors as a sub-study in lidERA. Several other studies are ongoing in other lines of treatment, including a Phase I/II study in combination with our CDK4/2 inhibitor and a trial in ER-positive, HER2-positive metastatic disease in combination with Phesgo. This morning, as you heard, we announced the positive results from the Phase III evERA trial in second-line HR-positive metastatic breast cancer, comparing physicians' choice of an AI or fulvestrant plus everolimus to the combination of giredestrant plus everolimus.

Of note, a 100% of patients had received a prior endocrine therapy and all had received a prior CDK4/6 inhibitor. The study met its co-primary endpoints with giredestrant and everolimus demonstrating a statistically significant and clinically meaningful improvement in progression-free survival in both the intent-to-treat population and the ESR1 mutated population. Overall survival remains immature at this time, but there is a positive trend in both the ITT and ESR1 mutated cohorts. Consistent with prior studies, treatment was well tolerated, and we did not observe any new safety signals. The evERA results reinforce our confidence in the giredestrant clinical development program and show the potential for this giredestrant combination to improve patient outcomes in the second-line setting.

Full data will be presented at an upcoming medical meeting, and we are actively sharing these results now with health authorities. In the wake of these positive results, we're looking forward to the persevERA readout in Q1 of next year, followed by our trials in adjuvant therapy and endocrine-resistant disease. We believe that this compendium of studies will make giredestrant the standard of care across all estrogen-dependent breast cancer indications. Complementing giredestrant is our best-in-class PI3K inhibitor, Itovebi. For decades, we've realized the potential to target PI3K mutated in 40% of HR-positive breast cancer and 30% of HER2-positive breast cancer. Nonetheless, the available molecules previously in development and currently approved demonstrated limited efficacy and poor tolerability with a higher rate of treatment discontinuations due to adverse events.

In contrast, Itovebi is differentiated as a highly potent and selective inhibitor of PI3K alpha. A unique feature of the molecule is its selective degradation of the mutant PI3K alpha subunit, conferring greater efficacy and far greater tolerability when compared to the alternative PI3K inhibitors. Moreover, Itovebi can be combined at full doses with endocrine therapies and CDK4/6 inhibitors. In the INAVO120 study, the combination of Itovebi, palbociclib and fulvestrant demonstrated a near 60% improvement in progression-free survival in the first-line endocrine-resistant setting and served as the basis for our approval in the U.S. and EU with additional approvals on the way. At the recent ASCO meeting in June, we shared an update of INAVO120, demonstrating a statistically significant 33% improvement in overall survival for the combination, making Itovebi the first PI3K inhibitor to improve overall survival.

At that update, the time to subsequent chemotherapy was delayed by 2 years among patients receiving Itovebi. Importantly, Itovebi continued to be well tolerated with discontinuations in the single-digit range. Based on the strength of these data, we're conducting additional studies in PI3K-mutated breast cancer, including INAVO120 a head-to-head comparison of Itovebi versus palbociclib in the post-CDK setting, INAVO120 in the first-line endocrine-sensitive setting, INAVO122 in the HER2-positive disease. We recently initiated a Phase II trial of Itovebi in combination with ribociclib and letrozole as neoadjuvant therapy for patients with Stage II and III disease, and we have other trials in the adjuvant setting currently in design. As well, we're conducting studies in multiple other tumors that possess PIK3CA-mutations.

Turning now to lung cancer and our KRAS G12C inhibitor, divarasib. In preclinical studies, divarasib is 25x more potent and 50x more selective than the current G12C inhibitors on the market. These features translate into robust durable clinical activity. In our Phase Ib studies in second-line metastatic non-small cell lung cancer, divarasib monotherapy achieved an objective response of 56% as compared to 30% to 40% for sotorasib or adagrasib. At the same time, divarasib documented a median progression-free survival of 15.3 months as compared to roughly 6 months for the other molecules currently on the market. Based on these highly encouraging results, we have an ongoing comprehensive Phase III program. First, we have the Phase III KRASCENDO-I trial, which is a head-to-head study comparison of divarasib versus the choice of sotorasib or adagrasib in the second-line non-small cell lung cancer setting. We have FDA breakthrough designation for this indication, and the trial will read out next year.

Next, we have the Phase III KRASCENDO-II trial in first-line non-small cell lung cancer, which we're initiating, comparing a chemotherapy-free regimen of divarasib and pembrolizumab as compared to chemotherapy and pembrolizumab. Finally, we're expanding our Phase III program into the early curative setting of non-small cell lung cancer, initiating the KRASCENDO-III trial, which will test divarasib in the adjuvant setting.

Moving on, Tecentriq continues to deliver positive results across tumors indications. At ASCO, we shared the results of the IMforte trial, demonstrating a statistically significant and clinically meaningful overall survival benefit for the combination of Tecentriq and lurbinectedin as a first-line maintenance therapy for extensive stage small cell lung cancer, establishing a new standard of care for this difficult-to-treat form of lung cancer. At the ASCO Plenary session this year, we shared the results of the ATOMIC trial in the adjuvant setting of DNA mismatch repair deficient colon cancer, demonstrating that the addition of Tecentriq to postoperative chemotherapy conferred a 50% improvement in disease-free survival. This regimen has now been incorporated into NCCN guidelines in the U.S.

Additionally, at next month's ESMO conference, we'll be presenting results from the IMvigor011 trial. In this trial, patients underwent serial ctDNA testing following cystectomy for muscle invasive bladder cancer. Those testing positive and therefore, at a high risk of recurrence were randomized to treatment with either Tecentriq or placebo, whereas those testing negative remained on observation. Among patients with positive ctDNA testing, Tecentriq demonstrated a statistically significant and clinically meaningful improvement in both disease-free survival and overall survival. Additionally, ctDNA-negative patients experienced excellent outcomes without adjuvant therapy. This is the first prospective Phase III trial using a personalized ctDNA-guided approach to identify patients who will benefit from adjuvant therapy with Tecentriq and has the potential to change the postoperative management of bladder cancer. We believe these positive trials of Tecentriq across the three domains of small cell lung cancer, colorectal and bladder cancers will translate into continued stable sales for Tecentriq through the decade.

Now turning to hematology. We're advancing novel molecules targeting both malignant and nonmalignant conditions across the spectrum of development, all designed to meaningfully improve the lives of patients with hematologic disorders. Over the past 25 years, Roche has transformed the landscape for patients with non-Hodgkin's lymphoma, which accounts for almost half of all blood cancers. We have a comprehensive program of clinical trials in the range -- across the range of NHL subtypes. The list on the right is not an exhaustive one, but it is a strong representation of the breadth of our clinical program in NHL. Our CD79B antibody drug conjugate, Polivy has had strong uptake in the first-line DLBCL setting with over 50,000 patients treated around the globe. At this year's heme summer congresses, we presented the positive results for the POLARGO study of Polivy and R-GemOx in second-line DLBCL. And prior to that, we shared the 5-year update for the POLARIX trial in the first-line setting of DLBCL.

We launched Lunsumio monotherapy in over 60 countries. And last week, the CHMP for the European Medicines Agency recommended approval for subcutaneous Lunsumio for the treatment of follicular lymphoma. We're moving Lunsumio both into the second- and first-line treatments of follicular lymphoma, and we look forward to the readout of the second-line CELESTIMO trial of Lunsumio and lenalidomide in the first quarter of next year. Additionally, the SUNMO trial of Lunsumio and Polivy in second-line DLBCL was reported out this summer, and I'll review those results with you. Similarly, for Columvi, the third-line launch of this highly efficacious CD20/CD3 bispecific is going well with over 6,000 patients treated in 60 countries commercially.

STARGLO in the second-line setting of DLBCL has been approved in over 45 countries, including the EU, and is a category 1 recommendation by the U.S. NCCN. And our first-line SKYGLO trial of Columvi is almost fully recruited. In July, we shared the positive results of the SUNMO trial of Lunsumio and Polivy in second-line plus DLBCL. Patients randomized to Lunsumio and Polivy experienced a 59% improvement in progression-free survival with a median PFS of 11.5 months for Lunsumio, Polivy and 3.8 months for R-GemOx. The combination of Lunsumio and Polivy also conferred a robust and statistically significant improvement in objective response with a 30% absolute improvement in ORR, 70% versus 40% and a doubling of complete response, 51% versus 24%. Of note, this chemotherapy-free regimen of Lunsumio, Polivy was delivered entirely as an outpatient and was associated with a manageable safety profile with a low incidence and low severity of cytokine release syndrome.

SUNMO is the first positive Phase III study of a chemotherapy-free regimen in DLBCL, providing an important fixed duration, well-tolerated, highly efficacious outpatient regimen suitable for the community care setting, where more than half of patients in the U.S. now currently receive their care. Also this summer, we shared the Phase Ib data for the combination of Columvi with Polivy-R-CHP in first-line DLBCL, demonstrating an unprecedented response, a 100% response rate and a 96% complete response rate. These responses were highly durable and treatment was well tolerated with relatively low rates of cytokine release syndrome. The results further enhance our confidence in the ongoing Phase III SKYGLO trial. SKYGLO is the only Phase III trial in which a bispecific is being added to the more effective Polivy-R-CHP regimen, offering patients the best chance for cure for diffuse large B-cell lymphoma. Of note, recruitment for SKYGLO is now nearly complete, and we're expecting the results from this trial in 2027.

Turning to multiple myeloma. We're advancing a unique CD3 engaging bispecific cevostamab into pivotal Phase III investigation. Cevostamab is the first antibody in the clinic that targets FcRh5, which is expressed on essentially 100% of myeloma cells. Cevostamab offers the potential to address an important unmet need in multiple myeloma. Emerging data in myeloma is now moving anti-CD38 antibodies, immunomodulatory agents and proteasomes with dexamethasone, all collectively now into standard first-line therapy, often now referred to as first-line quad therapy as the standard of care. Nonetheless, myeloma remains incurable and virtually all patients will eventually develop refractory disease. This highlights the need for additional safer and efficacious options in the second and later lines of treatment. At the International Myeloma Society meeting last week, we reported the efficacy data from our Phase I trial of cevostamab in combination with pomalidomide and dexamethasone in relapsed/refractory multiple myeloma.

The combination with cevostamab demonstrated deep and durable responses with an objective response of 86% to 88% and a VGPR rate of 74% across the two doses tested. Equally notable, the safety profile was quite manageable. On the basis of these exciting results, we're initiating the Phase III CEVOLUTION trial for the combination of cevostamab, pom, dex in second-line patients. Beyond CEVOLUTION, we believe the unique target of cevostamab, FcRH5 and the documented efficacy and safety for the molecule enables this agent to be a partner of choice with a range of multiple myeloma therapies across the industry, and we're now actively exploring those combinations with various potential companies. Further building on our portfolio in heme malignancies is our entry into cell therapy with the acquisition of Poseida Therapeutics.

CAR-Ts have transformed the landscape of hematologic malignancies, but current autologous technologies are highly complex and only about 25% of patients are technically eligible -- who are technically eligible for CAR-T actually receive the therapy. Our approach developed with Poseida leverages healthy donor-derived stem cell memory T cells, enabling durable antitumor protection. As well, Poseida's nonviral gene editing technology allows stable integration of multiple CAR-T's as well as high fidelity knock-ins and knockouts to both enhance efficacy and reduce adverse events.

We believe this off-the-shelf immediately available CAR-T approach will allow us to truly democratize cellular therapy. Our lead program is a BCMA-ALLO CAR-T currently in Phase I in relapsed/refractory multi myeloma, demonstrating high response rates in heavily pretreated patients for which the FDA has granted RMAT designation. Beyond BCMA, we have a CD19, CD20 dual CAR program currently in Phase I in B-cell malignancies. We're also leveraging this technology in non-oncology indications, and my colleagues will share some of that work with you this afternoon.

Finally, switching gears to nonmalignant hematology, we're initiating Phase III trials of NXT007, our next-generations bispecific that promises to enable patients with hemophilia A for a life-free of bleeding complications. Hemlibra has set a high bar in terms of efficacy, safety and convenience. NXT007 builds on Hemlibra to enhance potency, half-life and allow for low-volume, infrequent subcu injections. NXT007 is 30-fold more potent than Hemlibra. And as you can see on the left side of the slide, an in vitro assay indicates that NXT007 promotes thrombin generation within the range of people without hemophilia. In July, we presented our dose escalation trial of NXT in cohorts B-3 and B-4, patients experienced 0 treated bleeds during the observation period, and there were no safety concerns.

These data support the growing body of evidence that NXT007 has the potential to become a new standard of care. We're now initiating three Phase III trials, including a head-to-head comparison of NXT007 to Hemlibra, reflective of our confidence in this program. This ends my presentation. I thank you for the opportunity to share our portfolio in hem/onc and I look forward to your questions later this afternoon. But now let me turn to my colleague, Hideki Garren, to review the work in neuroscience. Hideki?

Thank you, Charlie. Well, I'm super excited to talk to you about our neurology portfolio. It's really -- we're one of the leaders in neurology, both on our on-market medicines as well as in the deep and rich portfolio. In fact, I was a little bit of a misguided soul. I left Roche for 4 years. And one of the main reasons I came back is because of our portfolio. It really is the envy of the industry, as you'll see in a couple of slides here. And how do we get to that portfolio? It's through critical capabilities. One is we are looking at multiple therapeutic modalities. The one you already heard about is Brainshuttle, [ anti- amyloid-beta ] for Alzheimer's disease, which we called trontinemab. We also have a Brainshuttle for multiple sclerosis. It's one of our many modalities we're looking at.

As you heard multiple times already, we have this unique collaboration with Pharma and Dia. In fact, neurology is one of the areas where we have really used that to our advantage. One example is this traveler prescreening program for Alzheimer's disease, where we're using a blood-based biomarker to identify patients as early as possible and to roll them into the study. And I'll talk about that in a little bit in more detail. Further, we're taking that one step further, using the same blood-based biomarker, we can identify patients for prevention of Alzheimer's disease. So this is in preclinical Alzheimer's disease. Preclinical meaning there are no signs or symptoms of Alzheimer's. These are patients with biomarker-based amyloid pathology who are walking around, perfectly normal maybe one of us as well. And therefore, we may be able to identify patients and treat them, prevent Alzheimer's disease with a molecule such as trontinemab. Finally, we have -- as you heard earlier this morning, we have made end-to-end investments in both MS and Alzheimer's disease. And so we're doubling down on our R&D capabilities there.

So let me show you the pipeline I've been talking about. So here's a pipeline. As you see, there are multiple different modalities. The index is on the right lower corner for nearly every modality you can think of is within our pipeline. And the other thing you should notice is the color coding. The across the board through development, we have neuroimmunological molecules across the board in development in purple. Neurodegenerative disorders, we also have molecules in development across the board in green. And the neuromuscular disorders in orange, we also have molecules throughout development. So a broad and deep pipeline in neurology, industry leaders.

One other thing you should note is that this pipeline looks a little bit different from what you saw early in the year because two molecules have moved into Phase III. That's at the bottom on the Phase III column, bottom trontinemab, as you heard, has moved into Phase III as well as prasinezumab for Parkinson's moved to Phase III. We're also looking forward to some exciting Phase III readouts with fenebrutinib in PPMS later this year and relapsing MS next year. The other Phase III molecule Enspryng in MOG antibody disease and autoimmune encephalitis, we expect readouts next year.

So let me give you one example of our new immunology pipeline molecule. Fenebrutinib, which you've heard about many times already, but let me dig down into some details about why we're excited about this. It is the only BTK inhibitor that achieves a dramatic CSF concentration greater than IC90 for B-cells and microglia. BTK inhibitors have a dual mechanism of action, as shown on the far left panel in terms of targeting B-cells as well as macrophage and microglia. And therefore, they can reduce relapses as well as disability progression. In the middle panel, our particular BTK inhibitor fenebrutinib has an optimized pharmacokinetic profile. What that means is if you look at that red line in that graph, there is a high bioavailability because of the free plasma AUC. And the reason that's important is that it will lead to increased concentration of fenebrutinib in the periphery as well as within the CSF.

And speaking of CNS penetration as in the far right panel, you see that we have a dramatically increased CSF penetration compared to the other BTK inhibitors. So therefore, because of this profile as well as CNS penetration, we have a high confidence in the outcome of fenebrutinib in relapsing and primary progressive MS. The other aspect of fenebrutinib that's unique is that it's very highly selective. And that's shown in the left panel where the selectivity is compared to the other BTK inhibitors in development. And you can see that based on this graph, there's very high selectivity to the appropriate BTKs that we're targeting.

And it's a non-covalent molecule, meaning that potentially we could have a beneficial benefit risk profile. It's the only non-covalent BTK inhibitor in development. We've completed the Phase II study called noncovalent in relapsing MS versus placebo, and I'll show you that in the next slide. We will have 2 studies reading out early next year with fenebrutinib versus teriflunomide. And then on the bottom, we have a study called FENtrepid, which is in PPMS versus Ocrevus, which readout toward the end of this year.

So let me show you the Phase II study that's been completed. Here, we see a really dramatic decrease in relapses and Gad lesions. So the annualized relapse rate is shown on the left, where during the double-blind period, which is that blue greyed part of the graph, you see really 0 relapses within that double-blind period. Patients roll over into the open-label extension. And you can see that the relapse rate is dramatically low at 0.06 up to week 96. Further, 97% of patients remained in the open-label extension through week 48, speaking to the tolerability of the molecule.

On the right panel is the data on Gad lesions. T1Gad lesions are measured by MRI, and it's indicative of active acute relapses and acute disease in the brain. And the double-blind period, again, is that greyed out period in the first part of the graph. You see a rapid and dramatic decrease in Gad lesions with over 90% reduction by week 8. If you follow that out into the open-label extension, we have 0 new lesions at week 96. So no new Gad lesions. So really a dramatic decrease in both relapses and Gad lesions, giving us a lot of confidence in the relapsing MS study.

Let me switch gears to our neuromuscular pipeline. The first example is Elevidys, which is a micro-dystrophin gene therapy for Duchenne's muscular dystrophy. And what's the most important message to take away here is that we have a positive benefit risk profile in ambulatory DMD patients. The reason that's important is diagram on the left, these patients over time have a loss of the ambulatory skills and motor skills such that roughly around age 13, most patients are wheelchair bound. So there's a huge unmet medical need for these patients.

On the right are data that you've already seen from our EMBARK Phase III study. What you see here is a 2-year extension of that study, where we're comparing to external control because the placebo period is over. But in every instance in every measurement of motor function, the Elevidys is favored and so on the right-hand side of the graph with significance. So we demonstrated a stabilization of disease progression with durable effects in DMD patients. It's important to note that over 850 patients have been treated with Elevidys -- ambulatory DMD patients have been treated with Elevidys across clinical and real-world settings.

Another example from our neuromuscular pipeline is Emugrobart, previously known as GYM 329, and we expect Phase II results in both SMA and FSHD later this year. Emugrobart is an anti-latent myostatin inhibitor and has a unique sweeping and recycling technology. And this allows for [Audio Gap] every 4 weeks of [ SC ] dosing. Preclinically, Emugrobart has been shown to have an enhanced benefit in mouse studies compared to other anti-myostatin. As you can see, these are the 2 studies that we're waiting for readouts, Phase II studies we're waiting for readouts, the MANATEE study, which is Emugrobart on top of Evrysdi for SMA and then the MANOEUVRE study, which is Emugrobart and FSHD. Important to note that in FSHD, there are no disease-modifying treatments available.

Check this [Technical Difficulty] pipeline. First of all, you've heard a lot about prasinezumab, but let me give you one slide on the Parkinson's disease. In Parkinson's disease, there is a steady progression of neurodegeneration and loss of motor and non-motor signs and symptoms, such that by the time of diagnosis of clinical Parkinson's disease, it's already about 40% to 60% of the dopamine producing neurons within the brain have been lost. This diagram the burden on the health care system with about 1% of the population over 60 affected by Parkinson's disease, resulting in a [Technical Difficulty] we do have symptomatic treatments, primarily L-DOPA, MAO-B inhibitors, but it's important to note that they do not change the course of disease and that there's a wearing off phenomenon as patients stay on these symptomatic treatments over time.

So prasinezumab, it is our anti-alpha-synuclein antibody. And as you heard, we have achieved a Phase III Go decision with this monoclonal antibody. The pathophysiology of Parkinson's disease is shown on the left panel where alpha-synuclein, which are the red items diagram in the left panel, causes neuron degeneration by accumulating neurons as well as interferes with synaptic transmission. Prasinezumab, which is shown in green in that cartoon is designed to bind to aggregated alpha-synuclein in order to reduce that cell-to-cell transmission of alpha-synuclein. We've conducted 2 Phase II studies, PASADENA and PADOVA. And these are unique in the sense that they are large Phase II studies and that we have a primary endpoint of a clinical endpoint as opposed to a biomarker endpoint. And the wealth of that data, the clinical data from those 2 studies has led to the Phase III Go decision.

PASADENA study -- PASADENA first Phase II study started many years ago. And the advantage of that is that now we have long-term data on patients on prasinezumab. And that's shown in the left panel. And what you see is that whether patients started with prasi from the very beginning, that's the early start in dark blue or they switched from placebo to prasinezumab and that's the delayed start in light blue, you see almost flattening of the measurement, which is called MDSUPDRS Part III. One word about the endpoint it is a clinical and Part III is a motor examination by the clinician. In other words, an objective measure of disease progression.

And you see near flattening of that curve over the 4-year course of the open-label extension. So the question is, well, what happens if you were not on prasinezumab? And what you see there is in an external control, the grey line going upwards. And that external control is a robust clinical external control called PPMI, where data were collected very carefully and the same endpoint was collected. And so therefore, is a robust comparator to prasinezumab in the open-label extension. And you see this wide separation of about 65% reduction in progression.

More proximal study is PADOVA on the right, the Phase IIb study, where again, its that same endpoint, the same measurement, MDSUPDRS Part III, but we have evolved that into a time to event. The reason we evolved into time to event is because these patients were treated with prasinezumab on top of their symptomatic treatments. What we found there is in the L-DOPA subgroup, we see a separation of the curves in blue and grey in terms of time of event, which achieved a hazard ratio of 0.79 and a p-value of 0.04. Now this is a subgroup that's quite large. About 75% of the patients were on L-DOPA. So it's a large subgroup of a large Phase II study. So that result, along with multiple secondary endpoints, which I'm not showing here, which all favored prasinezumab as well as a biomarker endpoint, which is the MRI endpoint looking at neuromelanin and iron, show that prasinezumab has biological activity and is very robust. So that led to the Phase III Go decision in the context of no disease-modifying treatments available.

Another molecule early in the pipeline that we're excited about for Parkinson's disease is NLRP3. The NLRP3 inflammasome has been shown to be associated with inflammation across various therapeutic indications and in particular, Parkinson's disease. So the idea is with an oral molecule called Selnoflast to reduce the activity of NLRP3 and therefore, reduce the brain inflammation and microglia activation and progression of disease. Phase I is completed, and we're waiting and analyzing the data from the Phase I. There's also been studies in coronary artery disease and asthma as this inflammasome has been indicated in those -- implicated in those diseases as well.

Moving on to our other neurodegenerative indications, Alzheimer's disease. As you know, there have been 2 molecules approved for Alzheimer's disease, lecanemab and donanemab, but there are limitations with what's available to Alzheimer's patients, those limitations are shown in the left-hand side of this slide. Number one is that there's limited blood-brain barrier penetration of these large monoclonal antibodies. Number two is the rapid and deep reduction in amyloid load has been shown to correlate with clinical response, and that's shown in the right-hand side of the cartoon that we want to reduce that amyloid load much more rapidly in a much deeper manner. The third item down on the left is the overall safety profile. There is a side effect with anti-amyloid therapy called ARIA, amyloid-related imaging abnormalities, and that causes brain edema as well as potentially brain bleeding and in rare cases, death. We want to try to avoid that as much as possible.

Biomarkers indicative of change downstream of beta amyloid is another area that we want to look at. So how are we addressing that? We're addressing that with trontinemab. Trontinemab is our Brainshuttle coupled anti-beta monoclonal antibody. And we believe this antibody addresses all the limitations that were outlined in the prior slide. From trontinemab crosses the blood-brain barrier through a TfR receptor mediated transport mechanism and that has been shown in humans to have an eightfold increase in the CSF, plasma ratio versus standard antibodies.

On the right-hand side is data that was shown earlier this year at AAIC, where we see a dramatic decrease in amyloid. What you see on the Y-axis is the centiloid scale. Centiloid is a 0 to 100 scale in with [Technical Difficulty] in milligram dose which is the purple line, you see a dramatic reduction such by day 196, roughly 6 months, these [Technical Difficulty] level amyloid, patients below the amyloid positivity threshold at 28 weeks, 6 months. All patients had a reduction in amyloid. In other words, there were no nonresponders. We also had a pronounced effect of fluid biomarkers downstream of amyloid on CSF pTau, which decreased by 27% at 25 weeks. And safety that's shown on the right-hand table, there was less than 5% instance of ARIA-E. It's much lower than what's been shown by the competitor molecules.

We're excited to announce that we have started the Phase III study called TRONTIER I and II, and we have achieved first patient enrolled in that study. The TRONTIER 1 and II is diagram on the right-hand side. What I'll point out is what's in the green box, we have a prescreening program called TRAVELLER, where we're using a blood-based biomarker, pTau 217 to identify patients for this study. And in TRAVELLER, we already have in the first 60 days, 3,000 patients enrolled in this prescreening study. So there's clearly a lot of excitement about this study.

The study design of TRONTIER I and II is diagram here where there after the prescreening, patients enter a screening period, followed by induction with trontinemab every 4 weeks or 6 months, followed by maintenance merit every 12 weeks. The maintenance period is important because it reduces any pathologic burden that the patients might have within the brain. Amyloid is one thing we can measure amyloid plaques or one thing we measure by PET, but we can't measure the other amyloid species that might be toxic like protofibrils. And so that's why it's important to keep that every 12-week dosing schedule.

So TRONTIER I and II have started with first patient in. We also announced that we have a preclinical study, which is in the lower left part of this diagram, and that we plan to start shortly. Again, this is in preclinical Alzheimer's patients with no signs of symptoms of Alzheimer's, but with biological evidence of amyloid pathology. And those patients are also going to come from TRAVELLER. So really robust prescreening program to identify patients. A brief word about our Dia collaboration. Diagnosis of early Alzheimer's is, of course, important and is a major unmet need. 55 million people -- over 55 million people have dementia, majority of those have Alzheimer's disease, but majority of them remain undiagnosed. And PET is not available broadly and measuring CSF A beta is invasive because you have to do a lumbar puncture or spinal tap. And therefore, we're really excited about the developments at Dia with the Elecsys pTau217 assay, which should be launched next year and which is the same assay we're using in the TRAVELLER prescreening study. They result directly of our collaboration with Dia.

And my last slide because the timer has now been at 0 for a while, is early-stage molecule called nivegacetor, which is a potential first-in-class gamma-secretase modulator orally administered for Alzheimer's disease. Higher iterations for gamma-secretase inhibitors. And what's unique here is that it's a modulator. And what we've already seen in the Phase I study, which is in the middle panel, a decrease in the pathologic A beta 42 and 40 species and an increase in the non-pathogenic A beta 37 and 38 species, which may also be protective as well. So the Phase I study is complete, and we're looking forward to the Phase IIa GABriella study, which have interim data hopefully by next year. And with that, let me turn it over to my colleague, Larry Tsai, to talk to you about our exciting immunology portfolio.

It would seem I'm being encouraged to use the podium. Is that -- okay, fantastic. Thank you so much, Hideki. Hello, everyone. Good afternoon. I guess this afternoon, my jet lag bodies isn't entirely sure what time of day it is. But I am Larry Tsai. I do know that, and I am the Global Head for Product Development Immunology. Happy to be here this afternoon to give you an update on the immunology strategy and pipeline. Hopefully, you are familiar with the One Roche immunology strategy, which we rolled out last year. And -- which is really an extension of the -- one Pharma strategy, which Teresa rolled out last year, which she spoke about a little bit earlier this morning. Roche has a long legacy of innovation in immunology, dating back more than 2 decades at this point. And the strategy really aims to build on that legacy using 4 strategic levers.

First of all, optimizing pathways, both novel and validated pathways that are known to play foundational roles in immune-mediated diseases. Second, looking for rational combinations of novel and validated pathways to raise the efficacy ceiling for immune-mediated diseases. Third, using biomarkers to break down patient heterogeneity and identify endotypes, which may predict a better response to specific therapies. And finally, where the science is mature, seeking functional cures using deeper B-cell depletion and potentially immune reset.

The strategy also centers around 2 end-to-end disease areas, inflammatory bowel disease and chronic obstructive pulmonary disease, while recognizing there may be opportunities in other disease areas where there is a potential for breakthrough innovation. This is particularly important in immunology because the immune system tends not to abate anatomic boundaries. And often, you do have to follow the science wherever it leads you to other areas of unmet need. So when we look at our clinical stage pipeline, we can see that there is a good balance between programs that are relevant to our end-to-end disease areas and to areas of potential breakthrough innovation.

And I'll spend the rest of my time today taking a little bit of a deeper dive into some of these programs. Okay, microphone is working, but clicker is not. There we go. Okay. So one area of breakthrough innovation is that of immune-mediated kidney diseases. Immune-mediated kidney diseases are a common cause of chronic kidney disease and are, in fact, the third leading cause of end-stage kidney disease. Chronic kidney disease is obviously very common globally, affecting about 1 in 10 people worldwide and is associated with high health care costs, including about 1/4 of U.S. annual -- U.S. Medicare annual budget and about $140 billion annually in Europe in health care-related costs. The Roche development pipeline includes several programs which are relevant. Two choices here for clickers and I'll try this one that are relevant to immune-mediated kidney diseases, including lupus nephritis, membranous nephropathy, idiopathic nephrotic syndrome and IgA nephropathy.

Gazyva, of course, is our next-generation anti-CD20 monoclonal antibody, which has been glycoengineered to be less dependent upon complement for cytotoxicity and therefore, making it particularly well suited to diseases where there may be depletion of complement levels such as advanced lupus. And in the Phase III REGENCY study, Gazyva was shown to be associated with a 46.4% complete response rate compared to 33.1% for placebo. Those results have now been published in the New England Journal of Medicine. The results have also been submitted to health authorities earlier this year with a U.S. PDUFA target date set for next month. In addition, international and national lupus nephritis guidelines have already been updated to include Gazyva as part of first-line combination treatment for patients with lupus nephritis. Beyond lupus nephritis, Phase III data in extrarenal lupus in the ALLEGORY study and in idiopathic nephrotic syndrome in the INShore study are expected before the end of the year and Phase III MAJESTY in membranous nephropathy data are expected in 2026.

With success in these disease areas, we expect that Gazyva can become the treatment of choice for immune-mediated kidney disease. In addition, Roche is helping to write the next chapter in B-cell depletion with its T cell engaging B-cell-directed bispecifics. Lunsumio is a CD3, CD20 bispecific antibody that's already been approved for relapsed and refractory follicular lymphoma. Lunsumio has been tested in the Phase I study in lupus and lupus nephritis and showed deep B-cell depletion in the top 3 dose cohorts with an acceptable safety profile. And accordingly, Lunsumio has now been entered into a Phase II study in lupus and lupus nephritis as of earlier this year. In addition, our CD19, CD3 bispecific program is in an ongoing Phase I study in lupus and lupus nephritis. And beyond our T cell engaging bispecifics, our CD19CD20-ALLO1 CAR-T cell program filed for an IND in lupus and lupus nephritis and should be entering a Phase I study this year.

And so with these treatments and with other B-cell-directed therapies in the pipeline, we believe that Roche can continue to build a leadership position in the emerging field of immune reset for autoimmune diseases. Moving beyond B-cell depletion, sefaxersen is our first-in-class antisense oligonucleotide for selective complement suppression in IgA nephropathy. Globally, IgA nephropathy is the most common primary glomerular nephropathy and can progress to chronic kidney disease and renal failure. High levels of complement factor B have been associated with IgA nephropathy. And so sefaxersen aims to treat IgA nephropathy by down-regulating complement factor B production by inhibiting mRNA translation. And sefaxersen was shown in the Phase II study to have met its primary endpoint of a change in 24-hour urinary protein with a 43% reduction in mean proteinuria at week 29 as well as meeting secondary endpoints and showing stabilization of kidney function. The Phase III imagination study of sefaxersen in IgA nephropathy is ongoing with an interim analysis readout expected in 2026 that could lead to an accelerated approval.

Shifting gears to talk about GI and in particular, our end-to-end disease inflammatory bowel disease. IBD is a very serious problem that affects about 8 million patients worldwide. And despite significant advances in treatment over the past decade, still only about 40% of patients respond initially to standard of care therapies and only about 20% achieve a lasting remission. And so afimkibart is our anti-TL1A monoclonal antibody that's now being developed for inflammatory bowel disease, atopic dermatitis, MASH, and as announced this morning, rheumatoid arthritis now as well. TL1A sits upstream of and is a key amplifier of several different inflammatory pathways and tissue remodeling pathways and cells that express TL1A and its receptor DR3 are known to be drivers of different immune-mediated and fibrotic diseases. And so we believe that afimkibart has the potential to become a pan-immunology pipeline in the molecule.

We have -- as I mentioned, the studies are underway in inflammatory bowel disease, atopic dermatitis and MASH. We have also initiated as of this year, registrational studies in ulcerative colitis and Crohn's disease. And we are continuing to look at additional indications that could help to realize the full potential of afimkibart as a pan-immunology molecule. The Phase II study that I mentioned in rheumatoid arthritis is expected to begin later this year as well. But of course, afimkibart's lead indication is inflammatory bowel disease.

And in the Phase II TUSCANY study, afimkibart was -- demonstrated strong proof of concept in ulcerative colitis and also demonstrated its potential to be a best-in-disease molecule. The Phase III AMETRINE I and II studies in ulcerative colitis are well underway, and we are very pleased to report today that enrollment is about 3 to 6 months ahead of target. And that's due in large part to the fast track designation that Levi spoke about earlier this morning, which has enabled expedited enrollment and execution. In addition, the Phase III SIBERITE studies in Crohn's disease are underway as are the studies in atopic dermatitis in Phase II and MASH in Phase I, as I mentioned earlier.

And we continue to collect biomarker data, which will help us to explore the endotypes that I mentioned earlier that may predict better response to treatment. But since IBD is an end-to-end disease area for us, we're already thinking about the next generation of treatments, including bispecifics built on a backbone of TL1A combined with other validated mechanisms of action such as IL-23. And towards that end, we have initiated a Phase II study in ulcerative colitis of our p40 TL1A bispecific molecule, and we have additional bispecifics in preclinical development at this time. But let's not forget, although not technically a part of our immunology strategy, Roche's commitment, Roche's contribution to the fight against global antimicrobial resistance, which is part of our commitment to global health security.

Zosurabalpin is a novel macrocyclic peptide molecule that targets the LPS transport system in gram-negative bacteria. Zosurabalpin has been shown to have potent activity against Acinetobacter baumannii, including carbapenem-resistant strains, which are the most -- or the highest threat pathogens according to WHO and CDC from a bacterial pathogen standpoint. And Acinetobacter baumannii, in particular, carbapenem-resistant strains are becoming a more common cause of nosocomial infections, including hospital-acquired bacterial pneumonia and ventilator-associated pneumonia, which are associated with high morbidity and up to 40% mortality. A Phase III study of zosurabalpin in [indiscernible] is getting underway and is expecting to enroll its first patients in 2026. And with that, I will end, and I will pass the baton on to Chris Brittain, who will give us an update on what's going on in ophthalmology.

Okay. So good afternoon. Thanks for the introduction, Larry. So I'm aware that I'm the last person standing between you all and an exciting update on the cardiovascular, renal and metabolism franchise. However, I really want to make sure that you learn a lot about the ophthalmology business because there is so much exciting stuff going on. So give me my 15 minutes, please. Just as a reminder, ophthalmology, our strategy remains threefold. Number one, we continue to treat patients who have suffered vision loss, and we have the great example of Vabysmo, which you heard from Teresa and Levi this morning, which is doing really great work for patients globally. We also want to treat earlier in disease still, and that in order to prevent that vision loss happening in the first place because even though Vabysmo does great, the majority of patients once they have vision loss due to the structural damage cannot achieve the vision which they were used to before that disease impacted them.

And thirdly, we aim to have truly restorative therapies for those patients who have more end-stage vision loss. And this is the graph which you see at the bottom right. And examples of these will be our OpRegen cell therapy program. So our critical capabilities remain key to our success. So we have a number of novel mechanisms of action, which you'll hear about. And these include bispecific antibodies, of which we now have 2 in clinic and 2 declared novel MOAs, which you'll hear about. We have a concrete and strategic approach to extended durability and novel technologies. Examples here are the port delivery system implant and also, again, our OpRegen cell therapy, very innovative, very novel technologies. In terms of our digital capabilities, we have, over the years, brought in-house tens of thousands of retinal images on top of thousands of biosamples from patients and hundreds of intraocular fluid samples.

We've been using all that data to develop algorithms. We have a remote monitoring system. So we really have this robust digital capabilities system setup. And finally, end-to-end investment. We're really using our end-to-end disease areas, which are geographic atrophy and intermediate AMD and then the retinal vascular disorders, which are group together diabetic macular edema, neovascular AMD, diabetic retinopathy. And we're really investing in those heavily. And what that is enabling us to do is have a seamless transition through the stages. And then at the end, we have the opportunity to reverse and to back translate. So we provide those data from our aqueous humor samples, our ocular samples, which taken the Phase III and to identify new markers of disease and new potential targets for new therapies back to our research groups.

So in terms of our pipeline, it's exciting to share. First of all, I'll start on the right-hand side. So satralizumab in thyroid eye disease, you've heard of this program for the last couple of years. We'll be able to share some data at the American Society of Ophthalmic Plastic and Reconstructive surgeons, that's in the middle of October as ESOPRS. I'm not going to be able to share any specific data today. That data just came in-house in the last few days. The vamikibart Phase III program in uveitic macular edema, that data will be presented. Again, the 2 Phase III studies there. I'll talk a little bit more in a moment. That data will be presented at the Scientific Congress, the American Academy of Ophthalmologists in the next few weeks. We'll be sharing that data with the regulators. And those 2 programs together, we'll be discussing, as Bruno mentioned this morning, the Investor Relations call specifically for ophthalmology after the AAO meeting.

And thirdly, today, we'll be talking a little bit about showing some first-time data from our vamikibart DME program. And we are proud to say that we now have a new asset in Phase I, which is a VEGF/IL-6 bispecific molecule. So starting with susvimo in neovascular AMD. Just as a reminder, this is an intraocular implant, which is designed to continuously release a customized formulation of the anti-VEGF ranibizumab. And the actual implant enables us to have as infrequently as every 6 monthly refills of that device. Now the exciting data which we've been sharing recently is this on the left-hand side, which shows that over 7 years since the implant, about half of patients maintain driving vision. Now I'll compare that to a previous study, which was 7 years of ranibizumab. And over that 7-year time period, about 1/4 of patients were able to maintain driving vision.

So the advantage of having this continuous release of ranibizumab through the implant has really doubled that number. So quite remarkable results. Another -- and sometimes I'd like to share kind of patient, patient anecdotes, which are quite interesting. So we here in the clinic, one of our physicians are talking about a conversation he overheard. And one patient was saying, how often do you come into clinic and one patient saying every 4 weeks. And then the next patient says, oh, I come in every 10 weeks. And then the susvimo patient says, I just come in twice a year. So that is a real strong patient benefit for Susvimo.

Next piece is that the port delivery platform itself, the implant, we actually have now 2 bispecific molecules, which are ready to be used in that platform and additional -- a large number of additional preclinical assets.

Moving on to vamikibart. Now just as a reminder, IL-6 is a cytokine, which is very pro-inflammatory. You will know from many other systemic diseases, but it's also involved in the pathogenesis of uveitis and a number of retinal diseases. Vamikibart is designed for intraocular use, and it was first used and it's an intravitreally delivered product.

I've mentioned that -- I've not mentioned, but we're kind of excited to say that we've now got the uveitic macular edema Phase III data in-house. And they've shown that we have a favorable benefit risk profile in the patients that have undergone treatment, and the data will be shared with the regulators. We'll also, as I said, we'll be sharing the clinical data at the AAO, and we'll be able to discuss more clinical data on the 21st of October in our Investor Relations call.

Secondly, I'll talk a little bit about the DME study, but I'm delighted to say that we are moving forward with the VEGF IL-6 bispecific in DME. And dwelling on that for a moment, this is the diabetic macular edema study, Phase II study of vamikibart plus ranibizumab versus ranibizumab alone. So this is a superiority study. And what we see here with this data is that the overall outcomes were a 3.4 letter improvement in visual acuity when vamikibart was given on top of ranibizumab versus ranibizumab monotherapy.

And more importantly, which is the FDA approvable endpoint, which is a 15-letter gainers, we saw a 44.7 versus a 28.6 percentage of 15-letter gainers if you take vamikibart on top of anti-VEGF monotherapy. So that's exciting data from the [indiscernible] superiority study. As you note in the second bullet, we did note that there were 2 cases of occlusive retinal vasculitis. But when you take all this data together, we believe that there's a strong and robust proof of concept for the role and the benefit of adding an anti-IL-6 mechanism of action on top of VEGF. And therefore, we will be proceeding and we're in the middle of an expanded Phase I program. We'll be accelerating this VEGF-IL6 bispecific in DME. Additional benefit of this bispecific, obviously, is it's a fab sized molecule based on the DutaFab platform, and therefore, it's compatible with the port delivery implant.

And then going back to vamikibart in UME. I've already made the comments that we're excited that we've seen robust improvements in vision and anatomy in the 2 identical Phase III studies, which we've just read out, MEERKAT and SANDCAT. As a reminder, uveitic macular edema is the leading cause of vision loss in patients with uveitis. And despite a plethora of systemic immunosuppressive therapies available for this condition, there are significant unmet need for a safer option for patients. We are looking forward to sharing this data more in the American Academy of Ophthalmology, as I say, about 3 weeks' time.

Moving on to zifibancimig in neovascular AMD. This is the second molecule, which went into the port delivery implant. This is currently in the Phase I/II BURGUNDY study. As a reminder, zifibancimig is our bispecific, which targets both anti-VEGF and Ang2 and is much more potent than the -- in terms of both VEGF and Ang2 than the intravitreally delivered faricimab molecule. Therefore, we believe that zifibancimig will be able to provide us with an option in the port delivery implant, which will give us infrequently as once every year dosing. So this is proceeding very rapidly in our Phase I/II study.

Then moving on to satralizumab. As a reminder, satralizumab is a subcutaneously delivered IL-6 inhibitor. IL-6, as we know, is a key mediator of disease in thyroid eye disease. There is a significant level of unmet need that remains despite a single available treatment in thyroid eye disease. Thyroid eye disease is kind of a painful, disfiguring and potentially blinding condition, which people often forget about. And although there is an available treatment, there is a need for a treatment which is more safe. The available treatment has issues such as its teratogenic. It causes significant disruptions in the menstrual cycle. It causes deafness. It causes muscle spasms. So there's a significant opportunity there for a safer treatment. We have the data just brought in-house. I'm not going to be able to share any today, but what I can say is that we're looking forward to, again, to sharing the clinical data in the ASOPRS conference in just 2 to 3 weeks' time.

Next on is OpRegen, one of my personal favorite programs. This is the allogeneic RPE cell therapy injected under the retina for patients with geographic atrophy. This is delivered, as I say, subretinally. And what we're starting to see is over the years, we've got longer-term data from our Phase I study. And on the left-hand graph, you see the visual acuity remains stable over that 3 years compared to the fellow eye of these patients.

Now fellow eye controls aren't always optimal. But what I can say is that in the large trials, what we see consistently is that patients with GA generally lose about 5 letters of vision every year. So what we're seeing is stability of vision. So this is really exciting. And this is replicated in the anatomy on the right-hand side, whereby the size of the geographic atrophy lesion is remaining stable. And we see that is not the case in the fellow eye here, and we know from natural history of geographic atrophy that the lesions do get bigger.

So we're excited about this program. It continues in Phase II. And with extended follow-up, we're continuing to see a really robust and acceptable safety profile. Now obviously, this is delivered through a transvitreal approach currently. And therefore, there's an opportunity to improve the delivery. And Teresa talked this morning about a number of devices. And so we're excited that we've acquired 2 additional devices, which have the potential to even further improve how we deliver not just this therapy, not just the cell therapy, but other future cell therapies or other future gene therapies or other technologies that we need to deliver to the eye.

The left one is a transvitreal subretinal injector with a dual lumen. This means that instead of injecting, you go through the vitreous through the jelly in the eye, make a hole in the retina and inject your therapy. And instead of pulling out and making a second hole because it's got 2 lumens in that injector, you can just simply inject the drug without taking it out and putting it back in again. So this has a significant advantage.

The device on the right-hand side is the Orbit Subretinal Delivery System. It goes through what we call a transchoroidal approach, which is between 2 layers between what we call the retina, which is the seeing part of the eye and the choroid, which is the vascular support of the retina, goes between those 2 layers and then a needle comes and the cells are injected between the retina and the choroid. So you don't need to make a hole in the retina. So that's really important.

Again, these have the potential to make the delivery of these technologies easier and safer for patients, and we're excited to report these technologies in-house. And so this program carries on in its Phase II, and we're looking forward to sharing more data in future meetings.

And with that, I'm delighted now to hand over to our last act for the day, Manu Chakravarthy, who will talk through our cardiovascular, renal and metabolism pipeline. Thanks, Manu.

Great. Good afternoon. Can I stand here or do I have to -- it's good. Okay. Perfect. It's an honor to go after all of the -- all of my esteemed colleagues. So hopefully, we'll have another 15, 20 minutes of engaging discussion and then open up to questions. So you heard a lot from Teresa this morning about the obesity strategy. In case anybody is still wondering why is Roche in CVRM. Hopefully, this is a slide that can remind us as to why this is so important in a way that there's an obligation in some ways to really address this incredible burden to society. I heard earlier this morning that 51% of the globe is going to be either overweight -- with people with overweight or obesity.

And what we know from the epidemiology is that obesity transects and touches the lives of almost -- I wouldn't say everybody, but a great majority of the people in the world. And you can see in this graphic, the incredible overlap with many of the diseases and that's why when we really think about CVRM at Roche, we're thinking about these intersecting highly interdependent diseases. And so we see this as one continuum, whether it's CV, whether it's renal, whether it's MASH, whether it's diabetes, all of that underpinned through a common pathophysiological thread.

And so in order to address that level of heterogeneity, we really need to be thinking very broadly and very holistically about how to actually position our portfolio to really try to tackle this challenge. And so it's really around the 4 pillars that we had introduced last year. And just to reiterate some of the forward momentum that we've had since the last year, you can see with the first box, it's not just incretins, it's incretins and beyond.

We believe that our portfolio will obviously have a foundational backbone with incretins, but there are other foundational targets, too. And that was one of the reasons why we did the deal with Zealand with amylin. And then most recently, a couple of days ago with 89bio with pegozafermin, which is FGF21. So both amylin, FGF21 are foundational targets just like incretins are. In order to really address that level of heterogeneity you saw on the previous slide, we need combinations.

So without combinations, it's nearly impossible to really tackle this multi-pathophysiological spectrum of interdependent diseases. Then you heard a little bit about the comorbidities. We are singularly focused on improving the health of the society. And so in order to bend that trajectory, it's not possible to do that without addressing comorbidities. So the key comorbidities, of course, a lot of people think about are CV and renal, but you can also easily imagine other things like MASH and many other areas where there's a true unique synergy with other therapeutic areas. And you heard from my esteemed colleagues, they presented the work from neuroscience, immunology, oncology, these are potentially transsecting areas where incretins and our portfolio can be synergizing.

And the final piece is the end-to-end solutions, which -- without which we cannot really bring all these things together. So holistic solutions, which includes a very close partnership with diagnostics with our devices as well as with our digital apps and then ultimately to monitor, diagnose. And the final piece is very, very important as a physician, I can tell you that patient empowerment, especially in a chronic disease setting is critical.

People with chronic disease don't like to be reminded that they have chronic disease. And so how do you really address all of these things? And this is kind of what we have really spent a lot of time and effort and thought thinking through. You heard a little bit from Teresa this morning about how we've taken a more patient-centric approach. We also have taken an approach that we also listen to our KOLs, our key opinion leaders and other thought leaders in the field. And here is a smattering of things. And you can see that many of these things intersect with what Teresa shared with us this morning.

Patients really are looking for obviously superior weight loss, but also improved tolerability. They're also looking to improve not only 10%, 15% of their weight loss, but improving their whole -- the comorbidities, better maintenance of their weight loss, better quality of their weight loss. And so these are all things that we actually take very seriously. And this is the reason why when we get asked the question, why do you need so many different things? This is the answer to that because there is an intrinsic heterogeneity and the incredible fragmentation of this marketplace. And so we need to be in a position to really meet patients where they are.

And chronic weight loss management is a chronic weight loss journey. So we have to meet patients wherever they are, whether it's in a preventive mode, it's an early treatment mode or in really the high-risk comorbidity mode. And so this is why you need the armamentarium that we have, which is orals, incretins, non-incretins, combinations. So all these things are actually required to really compete and to really be successful in addressing the needs of patients that are living with obesity and its comorbidities.

So this is the backdrop in which we have taken a lot of pain and thought and care to construct the portfolio that you'll see today. So I was here a year ago and shared this slide. This is -- actually, 2023, there was only one, right? It was fulvestrant. In 2024, we added 4 more to our pipeline. And so this is what we had showed just approximately a year from now. And I'm really proud and gratified to be able to share with you today the portfolio as it stands in 2025. So this is the forward momentum and the progress that we have tried to articulate by learning about the market, knowing what it takes to compete and to pull together a portfolio that will address these kinds of heterogeneous needs.

So what I'll do in the next few minutes is to give you snippets and highlights of the 3 Phase III programs. Very happy to announce that we're moving CT-388 into Phase III. And we'll cover those top 3 Phase III programs. And then I'll give you a quick snippets of the remaining 4 Phase II programs with petrelintide, CT-996, emugrobart and then we'll end with CT-868. So let's start with fulvestrant. So we believe that this is a paradigm-shifting first-in-class opportunity that is truly unique. Why do we believe that? The first is the way that addresses the target. So it's way upstream of the -- what is called the RAAS pathway, that's Renin-angiotensin-aldosterone system.

So it really addresses angiotensinogen, which is made in the liver. And with an siRNA approach, you can have near a complete knockdown of this gene. So when I say near knockdown, it's like over 95%. So it's completely flattened out. And we know that the suppression of angiotensinogen directly correlates with blood pressure control. So we know that we definitely hit this target. Then the siRNA approach allows us to silence this for a long duration of time, and that's really critical. So that's one of the core reasons why you see here that we are able to reduce blood pressure, not just for a few hours in the day, but 24 hours in a day, 7 days a week for all the way up to 6 months.

So one shot every 2 years will give you 24-hour control and gives you nocturnal control. And we know from epidemiological studies that when you have nocturnal control, you will improve outcomes. So that's been a direct link. And that's one of the core reasons why we've been so focused on getting sustained blood pressure control over a long time.

So here are some data to actually prove that, right? So one, it's easy to say, well, we have nocturnal control. So we have a very comprehensive Phase III program. Levi shared with you how robust this program is. It was actually heartening to be quite honest, to see 3 Phase II studies before we actually decided to go to Phase III. And that shows the rigor of the program. So KARDIA-1 and 2 were in mild to moderate hypertension. So they're relatively healthy people with 1 or 2 antihypertensive meds.

One could have taken the view that, well, we could have actually gone to Phase III at that point in time. But we actually took a little pause to make sure that we truly understood what is the dose, what is the population and how do we actually want to study this. And that's why KARDIA-3 was actually designed. So KARDIA-3 is a population that is much sicker. So they are all people with established CVD or those with high risk of CVD, the exact population that we wanted to study with at least 2, if not maybe 3 or 4 background med. So these people are actually very poorly controlled, and that's why we call it uncontrolled hypertension. And the majority of them were in a backdrop of diuretics.

And so what you see in the middle panel is this very nice reduction about 7 to 9 millimeters of mercury in the population [Technical Difficulty] and diuretics, those that have a CVD risk. Those are the people we want to take forward. And then on the right, you'll see what I've been talking about is the importance of the nighttime control. And there, you can see clearly in the nocturnal period, you are able to really see profound suppression, both doses, 300 and 600 relative to placebo. So this gave us great confidence that we are able to achieve what we set out to do.

And so I think we already had made this announcement prior today, but just to reiterate, very happy to announce the initiation of ZENITH, which will be our Phase III 11,000-person cardiovascular outcome study. So this is not a small study. It's as robust as it gets. So a large population with a primary outcome of the composite of the essentially the 4-point -- what we call the 4-point MACE. Right? So CV death, nonfatal MI, nonfatal stroke and then heart failure event.

And then many secondary events that we will look at in all the different forms and components, including nighttime control, et cetera. So all in all, very happy that this is advancing forward. We feel very confident that this will then -- given the results that we have, it should translate to outcomes because the one thing that we know about blood pressure control is that, that's a significant predictor of cardiovascular morbidity and mortality.

Now let's transition from hypertension CV to more obesity. So really, on the left side of this thing, you had seen some of the work that we had presented last year. So to build on the work from the incretins, which by design, all of our incretin molecules have been what we call signal biased because we believe based on our existing data, both preclinical data as well as early clinical data suggests that having a biased signaling pathway enhances pharmacology. So more efficacy, potentially longer duration of action, and we will see all this borne out, hopefully, in the Phase II studies, but that's our approach to incretins.

And on the right side, you'll see now that we've added to this armamentarium, as I said, with petrelintide, which is a long-acting amylin analog. It is suitable for once-weekly dosing. We believe based on the preclinical characteristics and the preclinical data set, this is a molecule that will be differentiated because it has no fibrillation, it's stable at neutral pH, has better potency and stability and has the pharmaceutic properties for potential combinability. So all are important attributes and differentiators.

So I'm going to start with 388 because that's in our Phase III now. So again, as I said, very happy to advance this forward. We had shown you a lot of the Phase I data from the weight loss. So I'm not going to repeat that, but just remind us that we had seen up to 19% weight loss in 6 months in a cohort of people that were -- had obesity and nondiabetes. And here, what you're seeing here is the cohort with obesity and type 2 diabetes. And in the middle panel there is the curves of the A1c. So this is just over 12 weeks. Glycemic control on people that are on monotherapy of metformin. So if you treat people on a placebo with metformin, you get that gray line, but you treat people with metformin plus CT-388, you get the blue line. So that's a delta of around 2.8% relative improvement in A1c in as little as 12 weeks.

On the right, it basically underscores how potent the metabolic benefits are with CT-388. And this is the same cohort from our -- what we call the Phase Ib study over 24 weeks, obesity, again, nondiabetes, we looked at MRI-PDFF. And what was striking to at least to me, having looked at the MASH sort of field for a while is that almost 85% of people achieved a 30% relative risk reduction in PDFF. And for those that follow the MASH field, you know that 30% reduction in PDFF means that you actually have 1-stage improvement in fibrosis or at least 1- to 2-point improvement in MAFLD activity score. So very potent effect even as little as 6 months. So we are very excited about the metabolic benefits here.

So for all those reasons, along with all of the data from -- the totality of the data from Phase I because as I said, we've done it in people with and without diabetes, overweight and obesity and looked at it in all different ways. And consistently, we're seeing a very comparative profile of the weight loss. So from all of that, along with the ongoing safety monitoring with 2 existing ongoing Phase IIs, which are fairly large, between 350 to 450 people across the population, we feel very confident to be able to move this now to Phase III, which is planned to be initiated in the first half of '26.

So quickly about pegozafermin because that's our -- the third Phase III asset very quickly. For those that have followed the field, you'll appreciate that not all FGF21s are created equal, okay? Because everybody says, this is another FGF21. But again, having worked in this area for a little while, I can tell you that all FGF21s are definitely not the same. It depends on how and what the receptor potencies are, what receptors they bind to, et cetera. And so we believe that pegozafermin certainly has the characteristics to be best in disease for MASH.

And what was really exciting to us is that actually not just in between 2 panels, but the panel on the right. So in the Phase II study, it so happened that there were about 11 people that actually had a diagnosis of cirrhosis. So when we actually look at those 11 people, I know it's 11, but again, the point here is that it's a proof of principle. In nearly half of them in as little as 6 months, we were able to see a fibrosis stage improvement. So that is very hard to see, by the way, in as little as 6 months. And so to see that in an F4 population was, again, further reaffirmation of the properties of the FGF21 molecule.

And then in between there, you can see the more formalized analysis from the 24-week study ENLIVEN study, where there's about a 27% response rate for fibrosis improvement and roughly the same between 27% and 37% response rate for MASH resolution. So all in all, very excited about this, and we're looking forward to integrating with the company and then continuing the great momentum with 2 additional ongoing Phase IIIs that's ongoing with data that is to be expected in the first half of 2027.

Okay. So then moving on to our Phase II assets. So I'm going to talk about 4 very quickly. So let's start with petrelintide. So many of you have already seen this multiple ascending dose data. So currently, where we are with the program is that we have 2 ongoing Phase II studies, what we call supreme 1, that is in people with obesity nondiabetes and then supreme 2 is people with obesity plus diabetes. So both are going really well, very happy with the progress.

And then just a reminder here that in as little as 16 weeks, we see a clinically meaningful weight loss of about 8.5% with petrelintide. And what was, again, very reassuring to see this even though we had hypothesized that the tolerability would be better, always nice to see data that supports it. So you can see no vomiting, in petrelintide, excepting for the one case at 9 milligrams, which is very mild and then very little to no diarrhea, okay? These are the 2 things that really annoy people when they take incretins, for example.

The nausea is manageable, but vomiting and diarrhea are really life disturbing kind of events for most people. So we're very excited about this profile. We continue to push forward to get this into completed Phase II so that we can get to the Phase III, again, sometime in 2026. Now very quickly on the 996, just as a quick update. Again, many of this we have already presented last year, especially the exciting data from the multiple ascending dose cohort. So this was just a 4-week study in people with obesity nondiabetes. And you can see up to a 7.3% weight loss in about 4 weeks.

And that set of data was the reason why we actually initiated 2 large Phase II studies, which you see on the right. So again, roughly between 250 to 350 participants in both with obesity, with and without diabetes. So these 2 studies are now underway. It's being conducted. And again, the data to be expected in 2026. So very excited about this oral program because this is -- we believe, has the potential, again, to be quite differentiated at least from the oral space. Then you heard a little bit from my colleague, Hideki, about emugrobart.

So I'm not going to talk about the biology. He's already covered that really nicely. Just as a reminder that this is a sweeping technology and it removes the, what we call the, anti-latent myostatin. So it's really all of the ligands that bind to the activin A receptor. So just not myostatin per se, but a little bit more of a broader swath of ligands that impact the receptor. And so the overarching hypothesis here is that we would get more significant or more deeper weight loss with a clear preservation of lean mass.

So we're excited about what we call the GYMINDA study, which is the study that we're doing in combination with tirzepatide. So these are in people, again, overweight or obesity, a 48-week study with tirzepatide by itself and tirzepatide with the various doses of emugrobart. And then we will have this data again in 2026.

So last but not the least, we are also very excited about the ongoing developments and work that's going on with CT-868. So CT-868 has been designed, again, as a once-daily dually biased GLP-1/GIP and specifically for type 1 diabetes, again, a very underserved, high unmet need population. So we had already demonstrated the proof of concept of glycemic control reduction in people with type 2. And you might recall, but again, I'll just quickly remind us that we had seen about a 2.3% reduction in A1c in people with type 2 diabetes. And so we feel really good about the glycemic control potential.

And we have done 2 studies with type 1 diabetes, both ongoing, near completion. One was a mechanism of action study and the other is a formal Phase II study, dose range finding study, both of which are ongoing, and we'll have data from the Phase II study actually in the fourth quarter of 2025. But just as a reminder, why are we excited about this is that people with type 1 diabetes need to actually still lose weight, are still insulin resistant, still have a very high insulin burden. So the hyperinsulinemia actually predisposes to a very high cardiovascular risk. So this set of data in type 2, which we fully believe will translate to type 1, indicates that you're able to reduce glucose while also reducing the insulin burden, which is really, in our minds, the proof of principle that improving insulin sensitivity in people with type 1 diabetes will improve their overall outcomes as well. So very excited to see what those results bring us. But again, quite happy with the progress there.

So let me conclude with this slide, which is the slide that Teresa showed, which to me is an encapsulation of why it is? What we are doing? So let me end before I hand it off to Teresa to bring us home with 3 things. The first is we've taken a lot of thought and effort to understand what is our market needs, what are our patients' needs and what is the holistic way to approach this extremely complex and very heterogeneous disease.

So I hope you at least can leave with the impression that we have that understanding that it is a segmented and fragmented market. The second is we have made significant progress with forward momentum in our pipeline, right? So the way the pipeline has been constructed and designed is really to try to address this intrinsic heterogeneity and the interdependencies of CVRM.

And the third, which is what I'm most excited about is the fact that we are confident that we can actually execute this because of our capabilities and what unique opportunities are available at Roche. And that confidence then translates to commitment because we have to do this all in. This is not something where we can just sort of dip our toe in the water and see how it feels. We are all in into this. So we feel confident, we're committed, and I am really excited to see how this will evolve in the coming months and coming years.

So with that, let me hand it off to Teresa to bring us home, and I take questions after that.

Wow, what a day. We've covered a lot of information. But if there's one thing I hope you take away from everything we've discussed today, it's actually the word that Manu used a couple of times, momentum. We have spent a considerable amount of time over the last number of years, focusing on what it is going to take in order to change the trajectory of our future, and there's how we go. We have thoroughly embedded the Pharma strategy in our everyday operations. And we are investing and strengthening the capabilities along our enterprise from end-to-end to ensure that we can not only deliver to patients today, but we are setting ourselves up to deliver for patients around the world in the future.

We are optimizing our investments in our commercialization capabilities, not only to ensure that we can reach maximum potential with our on-market products, but they're ready to launch into these new disease areas when we get there. And we are maintaining and will maintain a stringent cost discipline to ensure that every dollar, whether it's an R&D dollar or an M&D dollar or an SG&A dollar is going to the thing that will deliver the highest return for the patients, for our people and for business.

If there is one thing I'm proud of over the last couple of years, it is the change in our portfolio. We are in a 180-degree different position than we were in a number of years ago. We have a significantly evolved Phase III portfolio and a robust Phase I/II portfolio that really well position us to deliver on our ambition of 20 transformative medicines by the end of the year and beyond.

R&D excellence is helping us figure out how to do this more efficiently with more rigor and with higher success rates and partnering supplements our impressive internal activities.

I'll leave you today with the same commitment I made to you a year ago. We will continue to operate with rigor in the science and discipline in the business, and we will continue the momentum that we've generated.

With that, thank you very much for your time and attention. I'm going to call all my colleagues back up on stage, and we can start grilling Manu with questions.

It's Michael Leuchten from Jefferies. Maybe 2 questions. One on cevostamab, just taking that asset straight in Phase III, given what we know about multiple myeloma is quite a aggressive decision. Just wondering how you feel comfortable about those range and [indiscernible] make a decision? And then second question on vamikibart, why would a bispecific targeting the same targets get around the adverse event...

Your microphone seems to be fading in and out. Can you just repeat the latter part of your question on cevostamab? I'm sorry.

Yes. So you're taking cevostamab straight into Phase III, as I understand it. And we've seen dose range work become more important with recent history on the regulatory side. Just wondering how you feel comfortable with the dose ranges out of the Phase I straight into Phase III.

Yes. Well, I think that as you saw, the 2 doses we focused on both are showing excellent tolerability and deep and durable responses. We are finalizing the choice of dose for the Phase III with health authorities, but we're honestly comfortable with both. With regard to the decision to go from Phase Ib to III, the strength of the data for cevostamab, pom, dex. And just to be aware, we've now treated 700 patients with cevostamab in aggregate. The strength of that triplet data and realizing the opportunity in second line because of the evolution of the field where all the drugs are now moved into first line and patients really need new and different, more effective and safe therapy second line, we see the -- and realize what's now available, we see the -- we believe that our probably of technical success based on these data in second line is worth moving forward on.

Okay. So on the vamikibart question, so we're moving -- we're prioritizing the VEGF-IL6 bispecific over vamikibart for a number of reasons. Safety is not the primary reason. The primary reason is actually because having a single molecule, which is a bispecific is much easier to deliver in clinic. Instead of giving 2 separate injections, you can give one separate injection -- you can just get away with just a single injection. So that's the primary reason why we're moving forward with the bispecific. In terms of is the VEGF-IL-6 bispecific safer than vamikibart, what we've seen in the early stages is that it really improves our confidence that, that specific VEGF-IL-6 molecule is looking really good. So we're very confident with that, yes.

Sarita from Morgan Stanley. Just a quick follow-up on Columvi. Have you had confirmation from the FDA post the negative ODAC that the Phase III SKYGLO trial serves as confirmation for converting the accelerated approval? And then I noticed in the data in the STARGLO trial, the OS benefit in when it's actually detrimental in the U.S. patients. So are you confident that we won't see the same trend in the frontline trial? And has the frontline trial recruited enough patients from the U.S.

And then if I could just squeeze in another question, sorry, on the FGF21. Do you expect this to move to a combination market with GLP-1? And if so, are you initiating those trials? And perhaps you could speak to the ease of co-formulation and administration.

Yes, I can start. So the ongoing discussions with the FDA, we have not yet finalized the confirmation. Obviously, the accelerated approval for Columvi remains and we are finishing our discussions with the FDA. With regard to your questions about SKYGLO, that we're completing enrollment. I can tell you without getting into the details, there is robust U.S. enrollment in that trial. So I don't think we should be running into problems in terms of U.S. enrollment within that particular first-line study.

FGF21combinations?

You mean -- so just to clarify, so GLP-1 with just generally or specifically to FGF21, sorry.

So MASH team GLP-1 with FGF21 will you start combination trials.

When -- so we haven't basically provided that level of information yet. But we -- as shown on the slide, we fully anticipate that there will be some kind of combination that we would need to do. But the exact timing, we'll come back to you on.

In terms of coformulation?

So in terms of coformulation, so I can start -- that's been a general topic of importance for us because as you saw, combinations is a critical part of our pillar. So all of our assessments have always included early CMC assessments to see is many of the things that we have in our portfolio combinable. So happy to say that while we're doing our Phase II study in parallel, we are also looking at coformulation, feasibility for petre and CT-388 for example, Similarly, we're looked at early for FGF21 and GLP-1 feasibility studies, for example. So that's part of our thinking in terms of feasibility of coformulations.

It's Luisa Hector from Berenberg. I had a question on the TL1A. Just interested to hear where you see that fitting into the treatment paradigm. And then perhaps for Manu, on obesity, could you talk a little bit around how to mitigate some of the risks of running trials today given potential dropouts on placebo arms and perhaps the expectation around the GI tolerability profile and how that could impact? And then maybe just a comment on the obesity market and how you think pricing might work in terms of maybe an induction price and then a maintenance price, whether that's something you're thinking about?

Yes. Maybe I can start in terms of the TL1A, I'm assuming the question has to do with where it fits into the treatment paradigm for MASH. Is that the question?

No, I'm sorry, in IBD.

In IBD. Okay. Yes. No, we believe that, I think, has the potential to be best in disease for ulcerative colitis and Crohn's disease. And therefore, we fully expect that it will be early in the line of therapy as a best-in-disease therapy.

Yes. So in regards to the trials. Thank you for asking that. We are acutely aware that the world that we live in now is very different than when those trials were designed 5, 6 years ago. So one of the things that we're -- we've been paying a lot of attention to is patient retention because that's one of the core challenges, as you can imagine, right? So you're running against placebo, yes, it's important from a regulatory perspective, but it's not the most patient-friendly perspective.

And so for that reason, in our trials, what we have decided is to actually allow for what we call long-term extension, meaning after a certain defined period, you collect your primary endpoint, then you convert all of those that were on placebo to active. So that allows for motivation, for retention, for, again, doing what is right ultimately. So we believe that can be a substantive benefit.

The second is partly related to your question, if I understand it is, in order to keep patients in the trial and motivated, ultimately, we fully recognize that we would need to go head-to-head. And differentiation is going to be important, and that is very much something that Morten and I think about every single day.

And so I can't get into the details of exactly what a competitor in what trial and what population today. But suffice to say, that is at the forefront of our thinking in the design of our trials is that we would need to differentiate and we would need to go head-to-head against the right and the appropriate competitors. Your third question was to Morten, I think.

Yes. I think this thing about a weight loss and a maintenance phase, I think, is starting to evolve. And of course, the question is, will you use the same agent and the same dose and the same everything to reduce the weight loss and maintain it over time. And if that's the case, then you could also argue the price will be the same for the 2. If you can use 2 different, you can say, agents or doses, then you would have the opportunity to charge a price for the weight loss phase and a different price for a different agent on the maintenance. The jury is still out if that will become the way obesity is treated over time. So I think it's too early to comment on a specific price point there. But I think there are some underlying dynamics that needs to play out to really consider such an approach.

Richard Vosser from JPMorgan. Maybe a question on fenebrutinib to start with, please. There's been discussion on lots of the BTKs about changing the endpoints towards a disability endpoint. So I just wonder whether you've had that discussion in terms of your Phase III. And also wondering on the baseline characteristics because as earlier patients are recruited, it's much more difficult to show a difference to Aubagio. So thoughts there, please. And then second question on 996. Lots of discussion on CT-388, just wondering what's going on with 996. It seemed to have got lost in the presentation.

I can address the feneubutinib question on relapsing MS studies, which is what I assume you're asking about. Yes, we're -- at this point, we're not going to change the -- we don't have plans to change the endpoint from annualized relapse rate. We will, of course, look at disability progression as the secondary endpoint, but we have high confidence that based on the Phase II study that we should be able to show a difference in relapse rates.

Do you have a follow-up?

Sorry, baseline characteristics and how similar are they to Phase II? Are they early patients? Just wondering how that might impact the results?

Yes. It's -- the baseline characteristics are relatively similar to other Phase IIs. But what's happened is over evolution of time, the relapse rates just generally have decreased. Nevertheless, because of what we see in the Phase II study, we still have high confidence that we'll see a difference.

So on 996, I mean, it's true, and there's only so much you can pack into a 15 minutes. So it's -- while it may have been buried, it's certainly not in our mind space. It occupies a significant portion of our energy and time. We believe that 996, as I said in my presentation, has the potential to be differentiated and has the characteristics based on the data that we currently have to support that. So we're currently in 2 Phase II studies. Just to reiterate, again, it's both in people with obesity, with and without diabetes. So all in all, it's about between the 2 studies, roughly 600 to 650 participants. So it's a very robust Phase IIb program. We believe that data from that program will inform on the dose, the regimen and also, of course, trying to really creatively think through what are the other areas we can start to really differentiate from endpoint perspective. And all of that will then inform our Phase III decision, which we, as I said before, anticipate in 2026.

Sachin Jain, Bank of America. Two on CNS and one obesity. On fene, whilst we're focused on efficacy, could you just update on the liver safety in your total package? I think there's been one confirmed case of iso, does that impact your regulatory sort of discussions if you get in efficacy? Secondly, on tronti, I think the primary completion of tronti studies is mid-'28. Have you had any discussions of potentially earlier look in the study on amyloid for a potential earlier filing?

And then just one on obesity on 388. Any high-level commentary on the Phase II interim that's driven the Phase III start? Did you sort of get your dose maximization on safety? And have you told us what dose and titration regimen you're going into Phase III with?

So let me start with fene. And yes, we've seen slight elevation in liver function. And it's only in a couple of patients, they're mild to mild plus in nature. They were reversible and did not lead to any consequences. We did have one highest loss case, as you pointed out, but it is not a great concern. And then on the tronti. Tronti, yes, as far as earlier filing, we, of course, open-minded to really get this drug, promising drug to patients as soon as possible. But currently, we don't have plans to accelerate filing of that on that study. Having said that, the FDA has stated that amyloid PET is a validated surrogate biomarker.

So on the 388, firstly, very important for ourself to be reminded that the studies are still ongoing. So the study remains blinded. So what we have made the decision on is the extensive totality of the data. So that's what I was trying to emphasize during the presentation is we've done a fairly long and extensive Phase I/Ib program, that has gone up to 6 months essentially. And many of that data, you've seen it either in previous meetings or at the ADA this year. So that totality of the data gives us a lot of confidence in terms of what do we really think is the trajectory? Where do we really believe are the safety tolerability issues. But we also believe that -- which is a known fact at this point, which is going lower and slower in the titration will actually help us.

There is ongoing safety monitoring in any study that you would do, and that further reassures us of the profile that we anticipate. So when you take it all together, that's the totality of the information on which we've made the decision to go to Phase III. It's not because we've looked at the data or anything like that. So again, I want to emphasize the data remains blinded, and we need to keep that so for data integrity.

So is dosing being decided? Are you waiting for the Phase II to decide the dosing in Phase III?

So again, based on the PK data, the PK/PD modeling, again, the information that we already have gives us confidence of what the dose is -- at least the dose limits ought to be. So it's not that we have to specifically wait, but we do want to wait to see what is the final doses, if you will.

Can we start in the midsection? Is there in the back anyone? No, then we go to front.

Colin White, UBS. My first question was on FGF21. In 89bio studies, it seemed the placebo rate was quite low compared to some of the other studies that we've seen. I just wondered if you could comment on your thinking about the reasoning for that and then what gives you confidence when you look at all the data that's been presented that your FGF21 is differentiated from the others?

And then the second question I had was on fenebrutinib and the FENtrepid study. So it's designed originally to show superiority. I was just wondering when you look at the data from FENopta on disability and obviously, you know the effect of OCREVUS from your studies, if there's anything you've seen in the clinical data to give you reason to believe that on 12-week disability progression, it might be faster acting.

So I got the first part of your question about the placebo response, so I'm happy to answer that. What was your second part?

On FGF21, I mean, what makes you think your FGF21 is differentiated from the others in development?

Okay. So on the placebo response, so one thing to keep in mind again is that's why it's -- kudos to the 89bio team in terms of how they conducted the study. It was very rigorously done. So it was a biopsy-based study, the biopsy endpoint and the way that the biopsy endpoints were read were very rigorous. And so that rigor allows to basically get to what would be a true treatment effect size. So again, really kudos to them for designing a trial that allowed us to get to that very clear result.

So in terms of the differentiation with others, I mean, I cannot get into the nitty-gritty, obviously, for obvious reasons. But on a broadest level, it does matter whether you go take the approach of targeting based on a ligand-based approach or a receptor-based approach. So these are all ligand-based approaches.

And so we believe, again, that based on the data that we have seen and the data that's available for -- that's in the public domain, we do think that the data speaks for itself. It does look quite robust. The nausea and the vomiting rates are there, but it's very low, in fact, lower rates of diarrhea than others may have seen.

So when you take the totality of the efficacy, the tolerability, the safety, the convenience of dosing, when you put it all together and the targeting modality, it gives us a lot of confidence that this will be differentiated and the potential for best in disease.

And on the question of fenebrutinib and impact on disability progression, because of the mechanism of action where it's a dual mechanism action targeting B cells as well as microglia macrophages and the high CSF concentration we achieved, potentially, this could have a very strong impact on disability, even potentially higher than OCREVUS. Nevertheless, if it is as equal to OCREVUS in terms of impact on disability progression, it's an oral molecule. So we think this will be a benefit to patients.

James Quigley from Goldman Sachs. Two questions. So one on a much later point on the Phase III design or potential Phase III design for CT-388, but one of your competitors has gone down the traditional route in terms of escalating up to the highest dose for maximum weight loss. The other competitor has taken approach in Phase III of a more flexible dosing design. So where are you landing in terms of how you're thinking about that for the Phase III design for 388, and also, are you going to start a cardiovascular outcomes trial straight away as well?

And the second question, again, on fenebrutinib. The PK/PD data that you've shown today, again, highlights a stronger profile. But when you look at the Phase II data for evobrutinib, there you also had an annualized relapse rate of 0.08 in -- at the highest dose. So have you interrogated gut analysis to work out, why they can still get such a low relapse rate?

And then similarly, given the baseline characteristics are similar or similar as you mentioned, have you run any analysis of where you think the fenebrutinib profile will come versus the placebo or the [ bio ] arms in the failed trials for evobrutinib and tolebrutinib?

I can take the 388 question first. So we haven't, as I said before, disclosed the exact details but we will in the coming months when we actually start the Phase III program. So what we have today disclosed is the fact that we are going to Phase III. So we're finalizing the design. So obviously, I cannot speak to the specificity of what you're asking. But in terms of principle, I can tell you that those are all very much front and center of our minds. Just as I mentioned to Luisa, which is that the long-term extension part of the design, the head-to-head component, the regimen is equally important for us.

So providing the appropriate level of flexibility, appropriate level of maximization or optimization, if you will, of the tolerability profile because that's another way we believe we can be differentiated. These are all going to be key considerations in the design. And at this point in time, not able to give you more details than that, but that's the general principle in which we would want to do these designs.

And in terms of fenebrutinib and impact on relapses, as you saw, a really low annualized relapse rate, 0.06 at week 96, translates to one relapse every 17 years or so. And so really dramatic reduction in relapses. The important thing to remember about fenebrutinib is the benefit risk ratio is a highly selective BTK inhibitor, which is the advantage it has over the other BTK inhibitors and that is non-covalent, so potentially a better benefit risk profile.

Can we take question from the right side.

It's Justin with Bernstein. Two, please. Firstly, on zilebesiran, given the outcomes data, you said at 2030, could you just remind us on the LOE? And then the second one on [indiscernible]. I know it's not an exact comp true cap, but can you just help us understand if we should be expecting the sales to inflect? And if so, when?

When we'll see the inflection?

So I mean, as you remember, we have an entire program in the whole PIK3 mutations, and we start with a very small patient population. So that's probably -- for that population is a smaller CHF 100 million. But once the other trials read out, which you're going to have in the course of this decade, we'll see those inflection points and move further into frontline. So we'll see the potential of that hopefully being in that CHF 1 billion to CHF 2 billion range as we have anticipated coming then through once those trials come through.

Would you mind repeating the LOE part. We didn't quite get it.

Yes, sorry. I guess I kind of struggle with why Astra aren't doing an outcomes trial, you're doing an outcomes trial. You said today your outcomes data was 2030. So could you just remind us when the patent exclusivity expires on the drug? Or if I got running to the stick and why you are -- you obviously believe in outcomes and Astra don't, I'd love to understand why...

So I mean, I don't want to speak on what the exact data is, but we can get that information for you. But I think in terms of where we think the data will read out, just given the number of people that we just sort of alluded to, 11,000 people, 4-point MACE, it would have to be -- they would have to be followed for at least 2.5 years, minimum. So that's the earliest, but it's event-driven, right? So we would have to wait to see. So that's kind of the end of the bookend, if you will, that we have provided just to make sure that we have a sufficient number of events. We can get back to you on the exact LOE date.

Yihan Li from Barclays. So first question actually is a general question on Alzheimer's disease. So we saw like the trontinemab showed very promising and differentiated profile in Alzheimer's. But we also know like your competitor is running a trial for GLP-1s in Alzheimer's. So we're just wondering what is your view on GLP-1's potential in Alzheimer's in prospective trials? And do you see a rationale for GLP-1 plus anti-amyloid combo for Alzheimer's disease. And our second question actually is a very quick clarification question in NXT007. So you noted in your slides that you will run the Phase III head-to-head trial with Factor VIII. So we are just wondering like can we know which specific Factor VIII will be used in the trial, like potentially out to? Yes, that's all the question.

Why don't I -- I'll take your last question first while people work through the early ones. So we haven't provided yet the details. As you know, we have 3 Phase III trials for NXT, one in the pediatric population, one comparing to Factor VIII and one comparing to Hemlibra. And obviously, all 3 are critical to establish the basis of evidence for this to become a standard of care. We do look forward to providing those additional details, but that's probably as much as I can say right now.

And maybe while Manu and Hideki are starting through at a high level. We have to see the data on the GLP-1s and Alzheimer's. But if it happens to look good, this is -- we, of course, also believe that we have potentially a competitive set of assets in that space. And certainly, we think trontinemab could be very much competitive in that space. So certainly, it will be interesting, but everything depends on the data and whether this, in fact, plays out.

Yes, I couldn't have said it better. I mean I would only add to that is that it's always good to remind ourselves that sometimes Alzheimer's, what it's called Alzheimer's is also called type 3 diabetes because the pathophysiology is very, very similar. It's insulin resistance, neuroprotection, neuroinflammation. So one of the other things that we had highlighted worth highlighting again is the synergy that we have within our own TAs. It's a very unique opportunity for Roche to be able to mix and match the portfolio to go after the most unmet needs. And I think we'll all agree that Alzheimer's is an unmet need. So having that opportunity in a data-driven way is kind of how we will approach it.

Next question goes to Emmanuel.

Emmaneul Papadakis. Maybe a follow-up on NXT on the safety side, 2 out of 30 patients in Phase II seeing clinically significant antidrug antibodies is a fairly meaningful rate. So could you just talk a little bit about potential impact that could have in the real-world setting? How do you think that will be monitored -- because obviously, the risk of patient losing response could be pretty catastrophic.

Second question on SUNMO in light of what's happened with STARGLO. Have you had any agency feedback yet on the lack of OS benefit at the interim analysis and the potential importance of showing that at the final? And indeed, could you comment on the potential for an early readout on SKYGLO at interim?

Sure. So I think your first question was on the antibody drug -- antibodies for NXT. So as you rightly point out, we've seen 2 patients with ADAs that influence PK. One patient was on the lowest dose in the dose escalation study and did go off study. The second actually remained on treatment has had 0 bleeding events. And in fact, we're going to have an additional data cut of more patients to be shared at a future medical meeting. But what I can say is the totality of evidence gives us continued confidence to move into Phase III with that robust portfolio that we think will establish this as a standard of care and best-in-class treatment. With regard to the discussions, you asked about the feedback about STARGLO and the OS. I couldn't quite make...

SUNMO, which I think you planned for...

SUNMO, yes. So you're asking specifically about -- well, the study was designed to look at PFS. That's -- that was its PFS. PFS was the primary endpoint as discussed with health authorities. And obviously, it did achieve that robustly with a 59% improvement and a marked improvement in terms of months. The OS is a secondary endpoint. It's certainly going in the right direction. That's historically been what the health authorities have looked for. I think you asked about U.S. data. One thing that was presented in the presentation in July was the PFS within North America, which was a hazard ratio for PFS of 0.15. So certainly result looks like it's going in the right direction. So we are moving forward with our plans and discussions with health authorities. Your third question was on.

In terms of SKYGLO?

First SKYGLO...

Yes.

So the first one.

I think you said the readout would be '27 anytime sooner earlier interim.

Yes. I don't think -- I'm not really able to share more details other than the fact that the study is finishing enrollment. It's robustly. The feedback we get from TAEs was that they really -- they're very impressed with the Phase Ib data with a 96% CR rate, which gives us greater confidence. And right now, what we can say is we're expecting a readout in '27.

Okay. I think this is already final questions. We have to stop here, but I think all our key management will be available for the upper row, so you have another 30 minutes to drill down on whatever is on your mind.

Let me quickly use the occasion also to thank a couple of people because this then would not have been possible with the contribution of many, and just to call out here for the people from the IR team who have been working here since 4, 5 weeks, I think day and night, some of them to make it work. It's [ Jan Philip Schwartz ] and [indiscernible]who were responsible for both morning sessions and also manage the overall slide deck and who were working really, I can say, tirelessly day and night and throughout the last few weekends.

It's [ Rafael Pablolowsky ] who was responsible for the oncology deck. It's Loren Kalm, who was responsible for the neurology slide deck and also preparing macro topics and keeping Q&A up to date for speaker preparation. It's [ Anita Tang ] who was responsible for the immunology and ophthalmology slide decks and speaker preparation and also manage the epidemiology slides, which you can find in the appendix.

[ Julia Broiler ] and Sabine Borngr ber, who are responsible for the CVRM decks and also all the alignment needed around it with all the external partners as well. This was quite a task. And last but not least, and I want to thank the IR back office team that's Melanie Wolf, Beatrice Hau and Eva Losert for all the excellent organization.

And of course, there are many other colleagues from -- within the organization and different departments and who have contributed, which I simply could not all call out by name. This would be a long, long list. But we really want to say a big thank you from our end here as well.

And with that, I think we are closing the session. As said, you're welcome to stay for an upper row and get additional answers, whatever is on your mind. Thanks.