Roche Holding AG (ROG) Earnings Call Transcript & Summary

Ladies and gentlemen, welcome to Roche Angiogenesis 2022 Meeting. My name is Marco, and I'm the technical operator for today's call. Kindly note that the webinar is being recorded. [Operator Instructions] At this time, it is my pleasure to introduce you to Bruno Eschli, Head of Investor Relations. Bruno, the stage is yours.

Thanks, Marco. Can I have the first slide, please? Next one, please. So welcome to our event, and let me quickly guide you through today's agenda. We have 3 speakers with us. So first of all, Nilesh Mehta, our global franchise head in ophthalmology. He will present an update on our overall franchise strategy. And then we have Christopher Brittain with us, our Vice President and Global Head of Ophthalmology Product Development, who will provide you an update on our growing clinical development pipeline. And finally, last but not least, we have Nancy Holekamp with us, a well-known retina specialist and also a Vabysmo clinical investigator, who will lead us through the 2-year data of Vabysmo and Susvimo, which we have just presented over the weekend at the angiogenesis meeting. So overall, we have, I think, planned for 1 hour and 15 minutes, and we will have a Q&A session with all the speakers present following the presentations. I would also want to highlight here that we have a little online poll 5 questions only at the beginning of the Q&A session, which I would ask you to maybe fill in. Next slide, please. I also wanted to quickly use this occasion to give an update. We have a busy year ahead of us when it comes to pivotal readouts in 2022. And there is also a first green tick here on the regulatory side, which we have said, which is for Vabysmo in nAMD and DME, the U.S. approval already achieved. And I also wanted to point out that on the ophthalmology side, there is 2 additional readouts -- pivotal readouts to come later in the second half this year, which will do Susvimo in DME and diabetic retinopathy. And on the bottom in the box, I also wanted to use this occasion here to highlight some events where we have been able to fix dates. The next event, virtual IR event, will be following the MDA conference, is scheduled for the 16th of March, where we will present our latest data for Evrysdi and providing an update on our neuroscience pipeline. And then in the -- at the 16th of May, we will have our first little Roche ESG Day, where we will focus on the access to health care and equally represent both our divisions, the dia side and the pharma side. Next slide, please. Before handing over to our first speaker, I wanted to quickly share some recent external surveys from -- amongst ophthalmologists, which I found quite interesting, which focus on the remaining needs in AMD and DME. And if you look to the left side, this early result was collected amongst 990 U.S. and international ophthalmologists at the ASRS conference last year. And the question was, what do you think is the biggest unmet need in wet AMD therapeutics. And what you can see here is that the need for longer-lasting treatments was named first with around 40% and the need for new drug pathways was named second with around 30%. And then if you go to the right side, you see here, this was a survey conducted by Deutsche Bank last October. This was a survey amongst 41 U.S. ophthalmologists. And they were asked again about unmet medical needs in AMD and DME. And again, dosing frequency was mentioned here on the #1 position, followed by efficacy and then also by patient access and to a lesser extent by safety. So overall, I would say, a pretty consistent picture here on the unmet need side. Next slide, please. And this is my final slide. It just shows another poll, which is just 2 weeks old and was published by FirstWorld Pharma. This is following our Vabysmo approval in the U.S. And when asked about which proportion of patients who would get -- which proportion of patients would get a meaningful benefit from fewer eye injections. The answer was, again, 72%. And what is also interesting here on the right side is that there would be a benefit for new patients as well as for patients who could be switched from other treatments. And with that intro, I would like to hand over to Nilesh, our Global Head of Ophthalmology, for his presentation. Nilesh, please?

Thank you, Bruno, and good morning and good afternoon to everybody. Allow me to add my warm welcome to you all, and thank you for joining the call on Valentine's Day. My name is Nilesh Mehta. And over the next few slides, I'd like to cover the unmet need in retina, provide an overview of the ophthalmology market with a focus on retina, discuss the recent approvals of both Susvimo and Vabysmo, which we think take us a step closer to addressing some of these unmet needs and then close with an overview of our ongoing approach to ophthalmology innovation. Next slide, please. So as you can see, despite the tremendous advances in retinal care since the advent of anti-VEGF therapies some 15 years ago, we firmly believe there remains a significant unmet need and an opportunity to improve upon patient outcomes. Let me explain a little bit deeper. So on the left-hand side here, what we see is that even in clinical trials where people with retinal disease are treated rigorously, many don't achieve 20/40 vision, which is a threshold that's usually required for driving. In the data recently presented over the weekend, only about half of patients achieve this level of vision. Then further, when we move into the real world, the application of IVT treatments often cannot keep pace with disease requirements and many more people lose vision over time. We all know that these populations will become ever more prominent in society as the world ages and the diabetes epidemic continues. On the next slide, we show that the advances in retinal treatment and the epi factors mentioned really have made retina a key segment in the ophthalmology market. We estimate the full market to be worth about 22 billion and the retina forming about 2/3 of that at about 13. The outlook is for growth to about 15 billion by 2025. Tailwinds being innovation and epi and the headwinds being biosimilar impact. We move to the next slide. When we focus for a moment on the real-world outcomes we see, so we see from data on the left of the page here that health care systems really vary in their ability to deliver sufficient IVT treatments, and these are really correlated with the outcomes people achieve. Today, only about 50% of patients can be extended to Q3 month dosing with current anti-VEGF IVT therapies. We believe, really, this is where both Susvimo and Vabysmo have a role to play in allowing people with retinal disease to maintain their vision gains with fewer treatments and in a manner that health care systems and providers can deliver. Next slide, please. So now I'm really very pleased to update you on the recent events for the Vabysmo, faricimab, and you would have noted the publication of the 1-year primary data published in the Lancet. This is the first time that 2 major indications, neovascular AMD and DME, have been studied together in Phase III, presented jointly and published simultaneously. The publications highlight the data across the 4 Phase III studies with greater than 3,200 patients. They show that faricimab when dosed up to 16 weeks was consistently across the study shown to deliver vision outcomes that were non-inferior to aflibercept dosed every 8 weeks. And this is the first time that such a durability, supported by consistent anatomic findings, has been demonstrated in a large randomized controlled setting. And on the next slide, we see that just a week after the Lancet publication, the FDA granted approval for Vabysmo for the treatment of both neovascular AMD and diabetic macular edema and with dosing flexibility from Q4 to 16. Faricimab is under review with health authorities, including the EMA and others around the world as we speak. You're all familiar that this is the first bispecific antibody for intraocular use that simultaneously neutralizes both angiopoetin2 and VEGF. It's engineered using what is proprietary CrossMAb technology, which is developed in-house within pRED and represents more than 10 years of focused efforts. We've just presented the 2-year data from DME at angiogenesis over the weekend and look forward to sharing the 2-year neovascular AMD data later in the year. There are ongoing extension studies with RHONE-X and AVONELLE-X for both indications, and these will provide 4 years of data. Further, not to be forgotten, we have ongoing studies in retinal vein occlusion, which is the third largest indication in retina. On the next slide, I'd like to move to Susvimo. So also another focused development over more than 10 years, and it's the first and only eye implanted device producing -- providing continuous delivery of anti-VEGF and requires twice-yearly dosing, which is equivalent to 12 monthly doses of ranibizumab. Susvimo was approved by the FDA just at the back end of last year. So as you can tell, it's been a fairly busy time for us in ophthalmology. And we again have ongoing discussions with the EMA as well as other health authorities around the world. There's an extensive ongoing Phase III trial program covering diabetic macular edema, diabetic retinopathy and also testing dosing up to 9 months in neovascular AMD. So moving to the next slide. What does this really mean? So as outlined by addressing key unmet needs, we have an opportunity to improve sustained vision outcomes for people with retinal pathologies with treatments that health care systems can provide and which patients can manage in a day or 2. With Susvimo, patients can attain outcome certainty with just 2 refills a year following a surgical implant procedure and with Vabysmo about 3/4 of patients can achieve Q12 4x a year dosing in an IVT format. And this comes without compromising visual outcomes associated with more frequent A-VEGF dosing, which is either every month in the Susvimo studies or every 2 months studied in the Vabysmo studies. Finally, if I go to my final slide, so we're really very pleased with the introduction of both Susvimo and Vabysmo in the U.S. market, but our ongoing efforts won't stop there. And I wanted to conclude by sharing some of the 4 pillars that guide our efforts and continue to address unmet need. So first, there are the ongoing efforts towards novel modes of action and pathways, including DutaFab, which you'll hear more about from Chris as well as the geographic atrophy. These aim to get more people to better outcomes by addressing pathways beyond anti-VEGF. Then a number of long-acting approaches with port delivery system as a platform at first, but also expanding into long-acting IVTs as well as into gene therapy, which holds so much more promise to get real-world outcomes closer to those seen in clinical trials. In addition to our significant internal R&D efforts, we're active to bring in and partner with external innovations and the recent foray into cell-based therapies are just 2 examples of that. And finally, with the improvements to imaging and clinical biomarkers, we believe that the field is really poised for more specific therapeutics and management approaches in the future. And all of this makes this such an exciting time to be in ophthalmology. Thank you for attention. I'll hand over to Chris to overview some of these in greater detail.

Great. Thank you, Nilesh. So as you've heard from Nilesh, one of the opportunities that we've been giving here at Roche is to recognize that ophthalmology has actually become 1 of the 3 core areas for investment, the others being oncology and neurology. In the next few minutes and before Dr. Holekamp talks us through the newest data, I want to explain how we're looking at the future opportunities which we intend to utilize to improve patient outcomes and reduce the burden of retinal disease on them, the caregivers and health care systems. So looking at the next slide. After the launch of Susvimo in the U.S. last year, the launch of Vabysmo just in the last couple of weeks. And some of the exciting quotes which we heard over the weekend as we presented some of these 2-year data at the Angiogenesis Conference, such as very impressive data. I can't wait to get my hands on this drug and makes us think we have a new gold standard. I'm actually really pleased that we're going to have a continued flow of exciting data in the near and upcoming future. So starting, as you can see in 2023, we're going to potentially be filing, as Nilesh has already implied potential filing opportunities for Vabysmo in both CRV, central retinal vein, and branch retinal vein occlusion. And for Susvimo in patients with diabetic macular edema and diabetic retinopathy with treatment intervals going out as infrequently as 9 months with the diabetic retinopathy study. And then beyond that, we have readouts in the coming years, 2 of our exciting programs in geographic atrophy, just to call out 2, the [ MDH1 ] and the Ionis complement factor B inhibitor, about which we will talk a little bit more later on. And as you can see, there are a number of additional Phase I programs in both diabetic macular edema, GA and neovascular AMD. Now if we just move on to the next slide. As a reminder, which Nilesh has already spoken to, Vabysmo is a true bispecific molecule targets 2 distinct pathways, aims to reduce vascular leakage and inhibit neovascularization through the established anti-VEGF pathway, while the anti-angiopoietin 2 pathway also enables stabilization of vessels, reduced vascular leakage and reduced inflammation. As you'll see with Dr. Holekamp's presentation, this dual inhibition has led to even more durable efficacy in the long term in our Phase III DME program. So moving on to the next slide. The reason I wanted to really highlight the importance of bispecifics is that we believe these have got the potential to have a strong impact across our portfolio with 2 broad-brush approaches that we are currently undertaking. Firstly, the expansion of our portfolio with what we describe as platform technologies; and secondly the expansion of our portfolio with new indications. In particular, we're starting to move from the monotherapy Fab approach, which we've had and patients have had incredible success in treating their disease with through ranibizumab, Lucentis through a bispecific approach using our DutaFab platform. We're going to talk a little bit more about the DutaFab platform in a second. Secondly, investigating the use of multiple molecules or multiple different molecules in the port delivery platform. And thirdly focusing on the significant unmet need, which remains for patients with geographic atrophy. On the next slide, I just want to spend a minute talking a little bit about the DutaFab platform. As you can see, DutaFabs enable a binding of multiple targets from 2 or more distinct pathways. And these are both highly potent and stable, and this, in turn, allows high dosages for both intravitreal injections and compatibility with the port delivery system. The most advanced example is our VEGF and DutaFab, which is in Phase I currently in patients with neovascular AMD and which we'll be moving into the port delivery system. The next slide. This moves us on to the port delivery platform, which as a reminder, is designed to continuously deliver therapeutics through first-order passive diffusion kinetics. The concentration of ranibizumab in Susvimo, as we know, is 100 milligrams per ml. So I hope you can start to see why new therapies that are potentially higher in concentration and bispecific in nature have the potential to take this platform play to the next level. And then moving on to the next slide to discuss briefly geographic atrophy. As you know, geographic atrophy is the -- is an advanced form of age-related macular degeneration and affects the older generation at exponentially increasing rates, rising from about 1% at the age of 70 to over 10% of patients in their 80s. As a multifactorial disease, it's clear from the number of complement and non-complement associated genetic markers that the impact of environmental stresses and pathological findings, including the presence of lipids and amyloid within the retina that this is a disease that will likely require combination therapies in the future. And on the next slide, you will see how we are approaching this. We're approaching it by having currently 4 programs ongoing, 3 of which are discussed here. The first one at the top utilizes the anti-HtrA1 serine protease as an intravitreal injection, and that is in Phase II. The second program is our partnership with Ionis, which utilizes a subcutaneous injection of complement factor B. By targeting the liver, which is the source of complement factor B, this approach potentially enables simultaneous bilateral treatment effects for patients with bilateral geographic atrophy. At the bottom, you would have seen in the last few months our recent deal was with Lineage Cell Therapeutics. And we're really excited about the potential of this allogeneic RPE cell therapy to enable us to take a step towards restoration retinal structure and improvement in vision. This would be a real step change because as currently the current study programs and therapeutics in studies are aimed simply reducing the rate of disease progression. And next slide. So in summary, I'm sure I hope to share with you that with our execution of the launches of the Vabysmo and Susvimo are on track, our expansive pipeline is also setting us up for a really exciting future of possible treatment options for patients with retinal disease. And now it's actually a great pleasure to ask Nancy Holekamp, retinal specialist, Professor of Clinical Ophthalmology and Visual Sciences at Washington University School of Medicine in St. Louis and the Director of Retinal Services at the Pepose Vision Institute. And most importantly, a very experienced principal investigator across many of our programs, including Susvimo and for Vabysmo to talk through some of the new 2-year data. Thank you, Dr. Holekamp.

Well, thank you, Chris. Thank you for that introduction. It is my pleasure to be here to talk about the Vabysmo DME 2-year data set and the Susvimo neovasuclar AMD 2-year data set. Next slide. We're going to begin with faricimab in diabetic macular edema. Again, the 2-year results from the Phase III YOSEMITE and RHINE trials. This was presented by Jack Wells on Saturday at the Angiogenesis Meeting on behalf of the YOSEMITE and RHINE investigators, and I was quite pleased to be a YOSEMITE investigator. Next slide. So first, we're going to look at the clinical trial design. YOSEMITE and RHINE investigated faricimab dosed either Q8 weeks or dosed as a treat and extend dosing posology called a personalized treatment interval with up to Q16 week intervals. Both of these trials were large Phase III randomized double-masked, active comparator controlled trials. And in fact, combined, they looked at over 1,800 patients with center-involved DME and best corrected visual acuity between 20/40 and 23/20. These patients were randomized equally to 1 of 3 treatment arms. At the top, we see faricimab dosed Q8 weeks after 6 loading doses. We then have faricimab dosed according to this personalized treatment interval after 4 loading doses. And the active comparator was on-label dosing of aflibercept every 8 weeks after 5 loading doses. Of importance to this clinical trial design is the faricimab 8-week arm, which offers a direct comparison to the aflibercept Q8-week dosing arm with the only difference of that additional loading dose. We also see that the primary endpoint was an average of change in best corrected visual acuity, averaged over weeks 48, 52 and 56, at the study end, which we will share today, is the average of change in best corrected visual acuity averaged over weeks 92, 96 and 100. Next slide. So just a word about this personalized treatment interval algorithm. This is a protocol-driven treat-and-extend regimen. So essentially, it's doing what we, clinicians, we retina specialists do in our clinics in the real world under the auspices of a clinical trial. And importantly, decisions are based on both changes in central subfield thickness or anatomical changes and best corrected visual acuity. So there's an initiation phase where patients receive 4 initial Q4 week dosing. And if their CST is less than 325 microns, they go to the PTI phase. So if the CST is improved, you maintain the treatment interval. If the CST is stable compared to the reference CST, you can extend the interval by 4 weeks up to Q16 weeks unless there is a significant decline in vision of 10 letters or more. But if the CST is worsening, if vision is unchanged, you can maintain the interval. But if vision is worse, you can reduce the interval by either 4 weeks or 8 weeks. And again, this was a way of approximating what, we, retina specialists, are doing in the real world. Next slide, please. So it's important to note that baseline patient characteristics were well balanced across the treatment arms and that the patients enrolled in YOSEMITE and RHINE really mirror or match what we see in clinical practice. Patients with DME tend to be in their early 60s. They experience moderate vision loss, and the median vision here was equivalent to about 20/50. And the retinas are very thick with the CST measurements being in the high 400s. Next slide. So we've seen this data before. This is the 1-year data that was just published in Lancet. And we see vision gains with faricimab dosed every Q8 weeks and PTI up to Q16 weeks, while they were non-inferior to aflibercept dosed Q8 weeks. And if we build the slide, we'll see that these vision gains are maintained through year 2. And importantly, these visual acuity lines are on top of each other, and we see that there are no clinically meaningful differences when we compare all the arms across both trials. And in fact, visual acuity improvements were robust, ranging from 9.4 letters to 11.4 letters. Next slide, please. But now we want to look at durability. And we want to look at durability in the faricimab PTI arm, and we already know this from the data released at week 52, that approximately 50% of patients in the faricimab PTI arm could be dosed at Q16 weeks. Next slide. But now we have data through week 96. And we see that the proportion of patients who achieved faricimab Q16-week dosing actually increased to greater than 60% by week 96. And if we look at the percentage of patients who could be dosed at those extended intervals of Q12 weeks and Q16 weeks, the number was 78.1% of patients in both YOSEMITE and RHINE. When we look at the median number of injections in year 2, that means that the PTI arm was receiving 3 injections on average in year 2. Now a critical question is for those patients who achieved Q12 week or Q16-week dosing at week 52, what percentage of those patients could maintain that extended dosing interval until week 96. And that answer is 79%. And if we only look at Q16-week dosing and ask ourselves what percentage of patients could stay at that interval without having to contract or shorten the dosing interval, that number was 76%. So we see excellent durability of these extended treatment dosing intervals through week 96. Next slide. When we talk about extended dosing intervals, we want to make sure that we're not placing patients at risk for not being the big winners, not getting that greater than or equal to 15 ETDRS letter gain at 2 years. And on the left-hand side of this slide, we see that these results are balanced across the treatment arms of both studies. On the right-hand side, we're answering a slightly different question. We really don't want to have fewer doses and expose our patients to severe vision loss as determined by losing greater than or equal to 15 ETDRS letters at 2 years. And again, we see that's not the case. The vision is maintained and preserved across all treatment arms, including faricimab PTI arm at 2 years. Next slide, please. Now we're going to take some time to look at the anatomical results. And this was already known at year 1 that there were greater reductions in CST with faricimab dosed Q8 weeks and PTI up to Q16 weeks versus aflibercept dosed at Q8 weeks, and we have a direct comparison between the faricimab Q8-week arm and the aflibercept Q8-week arm, and those comparisons in both studies favor the faricimab treated arm. If we go to the next slide, we'll see that these greater reductions in CST in both faricimab arms were maintained through year 2 and that the anatomical results favor both arms of faricimab in the YOSEMITE and RHINE clinical trials. Next slide. So we're going to look at other secondary anatomical endpoints and just a note that where you see an asteric, there is a nominal p-value that's less than 0.05 when compared to aflibercept dosed every 8 weeks. And here on this slide, we see that more patients achieved absence of DME with faricimab dosed Q 8 weeks or PTI up to Q16 weeks compared to aflibercept through year 2. And again, I draw your attention to the dark blue lines comparing to aflibercept because those are Q8-week dosing arms for both drugs. Next slide. More patients achieved absence of intraretinal fluid with faricimab, either dosed Q8 weeks or PTI up to 16 weeks compared to aflibercept through year 2. Next slide. But here, we look at the rates of absence of subretinal fluid through year 2, and they were high and comparable across all treatment arms. But I have to say that DME is not really a disease of subretinal fluid, it is much more typically associated with intraretinal fluid. Next slide. So here, we see the proportion of patients with a greater than or equal to 2-step improvement in the diabetic retinopathy severity score, and this is consistent across studies and treatment arms. And we tend to learn from large clinical trials. And I think the teaching point here is that DRSS improvement is likely related more to VEGF than Ang2 inhibition. Next slide. Now we'll look at safety. Faricimab was well tolerated through the end of the study. Here, we're looking at events per 100 patient years, and we see that ocular AEs and serious ocular AEs were well balanced across the treatment arms when the data was pooled between the 2 studies and that the confidence intervals are overlapping. Next slide. But inflammation and retinal occlusive events have become critically important to clinicians understanding the safety of new molecular entities. And again, here on these tables, we see events listed per 100 patient years. And I want to call out intraocular inflammation events. We see these rates are low, balanced across the treatment arms with overlapping confidence intervals. The endophthalmitis events occurred in all 3 arms and were low and infrequent. And retinal vasculitis and occlusive retinal vasculitis events were not seen in any of the treatment arms of either study. When we look at retinal occlusive events, these numbers were small in all arms of the study and were not associated with any inflammation. Next slide. So in summary, over 2 years, faricimab demonstrated durable efficacy through disease control with up to Q 16-week dosing. And I think this really speaks to faricimab targeting 2 distinct disease pathways, which really lends credence to the thought that there is more vascular stability, more durable therapy and that we're able to maintain long-term vision gains, at least through 2 years as seen in these studies because of targeting this second pathway. When we look at the durable vision gains, the 1-year BCVA gains were maintained through year 2 but with less frequent dosing in the PTI arm. 78% of patients were dosed at Q12 weeks or Q16 weeks. And when we look at the Q16-week dosing alone, it was 62% of patients. When we looked at anatomical outcomes, the initial improvements were maintained over 2 years. But when we looked at key secondary analyses involving change in CST, absence of DME or absence of intraretinal fluid, these results favored faricimab. Faricimab was well tolerated. And importantly, there were no cases of retinal vasculitis or occlusive retinal vasculitis. And there is a long-term extension study, the RHONE-X that will continue to generate both efficacy and safety data through the next 4 years. So if we now take the next slide, we will shift gears slightly. And again, it's my pleasure to share with you the 2-year outcomes from the Phase III ARCHWAY trial, management of neovascular age-related macular degeneration using the port delivery system with ranibizumab. This was presented on Saturday at the Angiogenesis Meeting by Charlie Wykoff on behalf of the ARCHWAY investigators, and I'm proud to say that I was one of the Archway investigators. Next slide. Again, it's important to begin with the ARCHWAY clinical trial design. It was designed to evaluate the efficacy and safety of continuous drug delivery with the port delivery system refilled every 24 weeks. And again, this is a first look at continuous drug delivery as opposed to the pulsatile drug delivery we have been experiencing with frequent injections. In this study, there are just 2 arms, the port delivery system with 100 milligrams per ml of a special formulation of ranibizumab compared to intravitreal ranibizumab injections given every 4 weeks, which has been the gold standard for more than 15 years. At screening, both arms of the study received intravitreal ranibizumab injections, but at day 1, the PDS arm receives the implant with the initial fill of drug. That drug is refilled at month 24, but the 2 visits prior to the refill are called supplemental visits where patients are evaluated to see if they need supplemental ranibizumab injections, and we'll be looking at these visits in the upcoming slides. Now we already know that the primary endpoint was average of the mean change in best corrected visual acuity at week 36 and 40. And we already know from the reports in the Lancet that the PDS was non-inferior and equivalent to monthly ranibizumab. But if we advance the slide, what we're going to present now is the end-of-study results going out to week 96, where patients would roll over to a portal extension study. And this extended follow-up gives us additional refill exchange procedures, and it gives us additional information on safety and additional information on the need for supplemental injections. And we're going to look at key secondary endpoints through the second year essentially of the ARCHWAY study. Next slide. So again, the baseline demographics and ocular characteristics were well balanced, and the average age of patients in ARCHWAY was around 75 or in the mid-70s. And importantly, patients had received on average 5 anti-VEGF injections prior to enrollment in the study. Next slide. So this is data that has been presented before. It shows that the PDS dosed every 24 weeks maintain vision through week 96. But here, we're just showing the primary end point where the visual acuity results were found to be non-inferior and equivalent when average between months 36 weeks -- 36 and 40. There's 1 exception, and that's an expected transient post-surgical drop in vision, and this is real. And I do counsel and advise my patients that because it's a real surgery, they will have blurred vision for 4 to 8 weeks until they recover. But then we see that the lines are on top of each other. And if we build the slide, you will see that these visual acuity results are maintained through week 96, and we have 3 additional refill exchange procedures. Next slide, please. Per protocol, only noninferiority and not clinical equivalents was assessed, and the PDS was found to be non-inferior to monthly ranibizumab when averaged over weeks 60 and 64 and averaged over weeks 88 and 92. And we see that the confidence intervals are overlapping. We also see that the difference in adjusted means at these 2 time points is less than 1 letter difference. Next slide, please. So now looks -- let's look at anatomical outcomes. And when we look at retinal thickness measured as the center point thickness, again, measured from the ILM to the RPE through week 96, we see that these lines are essentially on top of each other, and they include 4 refill exchange procedures. And you may see that the lines begin to separate a bit between week 80 and week 96. And I'll just draw your attention to the vertical access where it's measured in 50-micron increments. So this difference is likely 20 microns or less. And as a clinician, that would not be significant to me or actionable to me. And it's reassuring to recall the earlier slides that show that the visual acuity lines were essentially on top of each other with less than a letter difference at key endpoints. Next slide. When we look at retinal thickness, but now look at the central subfield thickness measured from the ILM to Bruch's membrane through week 96, again, we see through refill exchange procedures and out to week 96, these lines are essentially on top of each other. However, now we can see perhaps the separation at week 80 to 96. Again, this difference is small and is likely 20 to 25 microns or less. And as we think back to the visual acuity results we just reviewed, it makes us rethink the presence fluid vis-a-vis visual acuity, that maybe control of disease through continuous therapy is as effective as maintaining vision despite these anatomical differences. Next slide. Now we're going to talk about the supplemental treatment. So 95% of patients did not meet the supplemental treatment criteria. That criteria was a decrease of 15 letters from the best recorded BCVA in the study or an increase of greater than or equal to 150 microns in CST measured from the lowest CST measurement or a combination of both anatomy and vision, where an increase of 100 microns in CST was associated with a decrease of 10 letters or more in best corrected visual acuity. But now we have 4 PDS treatment intervals to assess, and we see that the results are consistent through week 96, that essentially 95% of patients did not meet this retreatment criteria or supplemental treatment criteria and vision and anatomy were able to be controlled. Next slide. Systemic safety of the PDS dose Q24 weeks was generally comparable with monthly ranibizumab as expected. Next slide, please. But now we're going to look at ocular adverse events of special interest. And these are identified as adverse events of special interest because they emerged as we introduce a surgical procedure into the management of wet macular degeneration. And what we'll find is that these adverse events of special interest in the PDS arm were well characterized in the clinical trial setting and were generally manageable. When we look at this table, we see that the #1 adverse event was cataract, which is not surprising in this patient population. We also see adverse events in the PDS arm that we don't see in the intravitreal injection arms such as conjunctival bleb or conjunctival filtering bleb leak. The rate of vitreous hemorrhage was small, and they generally cleared, not requiring vitrectomy surgery. But we're now going to talk about conjunctival erosions, conjunctival retractions and endophthalmitis. And importantly, endophthalmitis occurred in 4 PDS patients and 1 monthly ranibizumab patient. And the FDA has issued a boxed warning for the PDS because it has been associated with a threefold higher rate of endophthalmitis compared with monthly intravitreal injections of ranibizumab. Now 3 cases of endophthalmitis were associated with conjunctival retraction. That in review of surgical videos, we could identify the root cause for the conjunctival retraction. This indicates that early detection and appropriate management with potentially surgical repair of conjunctival retractions or erosions could reduce the risk of endophthalmitis. And because of learnings from the ARCHWAY clinical trial, we can make recommendations for follow-up of these patients, again using slit-lamp examinations, perhaps a cobalt blue light and fluorescein stain or serial longitudinal photographs of the conjunctiva overlying the implant to optimize patient outcomes. Next slide, please. So in summary, the port delivery system refilled every 24 weeks, maintained vision and anatomic outcomes comparable with the gold standard for over 15 years, which is monthly ranibizumab injections. And this has now been demonstrated through 2 years of follow-up in the Phase III ARCHWAY clinical trial. And what we saw was comparable vision and controlled retinal thickness through 96 weeks. And with the extended follow-up, we can see that the PDS was noninferior to monthly ranibizumab for BCVA change at multiple points during the study. When we look at the durability of the port delivery system, we see greater than 95% of patients who were able to go the full 24 weeks without requiring supplemental treatments with ranibizumab injection. When we look at the safety profile, it's one of the advantages of Phase III clinical trials is that you can have a well-characterized safety profile. Importantly, with extended follow-up, the safety profile was generally unchanged from the primary analysis and no new safety signals emerged. The adverse events were generally manageable, but importantly, key learnings are being continually implemented to optimize patient outcomes. But importantly, the PDS was associated with a threefold higher rate of endophthalmitis compared to monthly intravitreal injections, but meticulous implementation of the improved instructions for use with the PDS will be key for maximizing patient outcomes in the real world. And I believe that, that is my last slide, and I would be happy to take any questions regarding both Vabysmo or Susvimo in the 2-year data sets. Thank you for your attention.

Thanks, Nancy. We will now start the Q&A session. We have also currently just to let you know 185 people joining. The first question will come from Wimal Kapadia from Bernstein.

Great. So just 2, please. So can I just first -- you mentioned that subretinal fluid is not a major issue in DME. But can I just ask how we should think about intraretinal fluid with respect to long-term outcomes in patients? Now clearly, there's a benefit there for Vabysmo of Eylea. But based on the 2-year data, would you argue this benefit has been proven? Or could we actually see this benefit expand further beyond the 2-year period? And then my second question, I appreciate it's not your data, but just given the debate on higher doses of anti-VEGF not driving differences in visual outcomes. I would appreciate your thoughts on the Eylea high-dose trial data. At first glance, it seems like it's slightly better on BCVA, but it's more of a debate on drying capability. So just what are your thoughts on this data and how importantly, look into here, how you think about Vabysmo stacks up against this high-dose Eylea data.

Sure. So Nancy, maybe do you want to take the intraretinal fluid question first, and then I'll take a stab at high dose.

Sure. So I actually have a great interest in fluid compartment and intraretinal fluid is damaging to vision, period, whether it's in AMD or DME, intraretinal fluid is the real driver of visual acuity outcomes in my mind. And so I think it is critically important to understand that DME is an intraretinal fluid disease. And it is quite impactful that the anatomical results really favored faricimab when it comes to the absence of DME or the absence of intraretinal fluid. So I think that this is a key finding from the YOSEMITE and RHINE clinical trials.

Thank you, Wimal. On the other data, a key point to note is really, fundamentally, this is where we're kind of different in our approach to unmet need. We actually think that the additional pathways is where we will get more benefit, the data that we've seen thus far from a Phase II setting that didn't meet its primary end point from what I understood. So yes, I think you're going to see different agents with different impacts to CST. But I think when we look at the totality of the data in Phase III setting with the Vabysmo, 3,200 patients showing consistent durability across the studies in the arms and the supportive CST and anatomic features, we're really confident that this is a data set that stands up to scrutiny.

The next question would come from Simon Baker from Redburn.

Two, if I may, please. On YOSEMITE and RHINE, on the PTI dose frequency. I was just wondering if there were any factors that led to Q16 dosing other than baseline disease. Is there any other factors that were determining which patients were in that group? And then secondly, a slightly broad question. You mentioned as one of the of planks of Roche's ophthalmology development efforts gene therapy. I just wonder if you could bring us up to date with where you are there. As far as I know, the only project in ophthalmology that Spark [indiscernible] disease. So I just wonder if you could tell us where we are there, given that was, of course, where Spark began its gene therapy efforts.

Chris, do you want to take the gene therapy and then maybe we can discuss PTI after that?

Sure. Thanks for the question. I think broadly, as you will have seen in the announcement at the end of last year, we did have a gene therapy interest with 4D Molecular Therapeutics. We've terminated that relationship and wish the company good luck in the future. I think the gene therapy remains a key platform of investigation for us within retinal conditions. And Spark remains very interested, and we're collaborating across our early research groups with Spark there. So I can confidently say that we remain very interested in the potential for gene therapy to have an impactful contribution to therapies in the future for retinal diseases. No updates on any additional programs as of today. But we remain very interested in the gene therapy platform.

I think -- I just kind of take a stab at the 17 RHINE factors leading to influencing the Q16, and then maybe Nancy I'll ask you to comment perhaps on the impact of that in real clinical practice. When we transitioned individuals from YOSEMITE and RHINE into the Q16, we actually took a number of factors in, which included both assessments of visual acuity and retinal thickness over time. So patients can move upwards and downwards dependent on the state of their vision and the anatomical status of their retina. So that's -- so these are the factors which clinicians use in the real world. And that's why I say, as you see, we have 80% are able to achieve Q12 or greater vision. But anatomy and vision were the main factors. I mean, Nancy, in terms of real-world clinical practice, do you want to comment?

Sure, I'd be happy to. So I can say in my own clinical practice, I have patients who are in Q12-week dosing posology for their diabetic macular edema, and it's like clockwork. And I have some patients who are in a Q16-week dosing posology, and they absolutely have to have their injections at those time points. But in clinical practice, it's very hard for me to know what the underlying either patient characteristics or ocular characteristics are that lend patients to these extended dosing intervals. And this is really the advantage of having these incredibly large prospective clinical trials. I think that, that is a great question for a post hoc analysis to see what the patient profile is that does well with these extended dosing intervals with faricimab. And so I think we have a lot more to learn from these large studies, and I look forward to those analyses.

Next one would be Sachin Jain from Bank of America.

It's a really simple one for Dr. Holekamp. I'm just trying to get a bit of a perspective on the clinical relevance of these assets. So I wonder if you could just share with us your rough split of existing usage in AMD and DME and Avastin and just take a stab how you think that shift in the next 2 or 3 years with the introduction of Vabysmo and Susvimo. I just want to get a sense of how quickly you think that shift happens. Is it new patients? Or are there very frequent or uncontrolled patients that are subject to a quick switch within the next, say, 6, 12 months?

So I'd be happy to take this one. I can tell you that I think the adoption of faricimab, of Vabysmo will be very, very rapid. It's an injection. It's easy for clinicians to reach into their refrigerators and pull out an injection. So I think it will be a very quick adoption. I think patients will initially be switched because that's the model that we saw when people switch from Lucentis to Eylea in the past. But I can tell you that, again, I really like Phase III clinical trial data because YOSEMITE and RHINE show us that still like almost 20% of patients will be dosed Q4 weeks or Q8 weeks. And this is what gets me really excited about the PDS being a platform. And that Chris mentioned that hopefully soon, we'll have the PDS for treating DME. Because I think even with Vabysmo, there will be a small percentage of patients, perhaps 20%, that might want the extended dosing durability of 6 months should the data prove good in Phase III clinical trial. So I think the uptake with Vabysmo will be very rapid based on the year 2 data. But again, I'm looking to the future for that small percentage of patients that still might want a different platform. Now I'll turn to A&D. I am very excited about the Susvimo platform. In fact, I operated on 3 patients 2 weeks ago and implanted Susvimo, and they were very excited about it and looking forward to doing more. But I think this uptake will be slower because it's a new surgical procedure. But I think the more physicians do, the more physicians will do. And that the uptake will accelerate as physicians get more experience with this new surgical procedure.

Did we answer your question, Sachin?

That's perfect. I just wondered like a push for any share assumptions in your own practice now and how quick do you think that shifts.

So I can tell you that today, February 14, is the first day that I can buy Vabysmo from my provider. And I have placed an order for it, and I'm trying -- I'm starting to get authorization for use. So I am an early adopter and very much looking forward to it, but I was in the clinical trials. But again, I think other clinicians who weren't in the clinical trials, I think there will be an early adoption in DME because the data is so convincing, particularly the anatomical data. And to summarize, I think Susvimo will have a slower uptake that will accelerate as more physicians get experience with this new surgical procedure.

Next question would be from Michael Leuchten, UBS.

Just a quick question for, I think, Dr. Holekamp, maybe for Christopher. When we look at the PDS patients that are being treated today, how would you describe the characteristics? I'm assuming a surgical intervention is a bit of a tougher hit to a patient than straight up injection. So maybe the 3 patients that you mentioned, Dr. Holekamp, what's the typical patient that gets a PDS today? And how do you think that will develop over time?

Maybe a quick comment first, and then I'll pass to Nancy in terms of the clinical practice. I would say that as we've expressed over previous calls that we continue to, and as Nancy's described, we continue to ensure that surgical -- surgeons are trained very, very extensively because this is an important and new surgical technique for the port delivery system. So the patient selection is an integral part. They have to be committed and they have to be of -- the type whereby managing an implant is carefully considered. So I think we have a two-pronged approach in terms of the trading. We have extensive case reviews, discussions, virtual reality simulators. And then we go on to in kind of peer-to-peer discussions to share how that trading takes place and to share direct experiences from case to case. So I think that's really important. And we will continue to learn how best to select these patients. But Nancy, perhaps from your -- first, describe how you went through your 3 recent implants?

Well, I will say that these 3 patients, they were aged 81, 80 and 78 years old. They had all had multiple, multiple injections, and they were quite enthusiastic about getting an implant. Now when it comes to discussing the Susvimo implant. In this age group, 2 out of the 3 patients had already had cataract surgery. So they understood that they had a lens implant. This is a drug delivery implant. However, for each patient, I had special informed consent that included safety data from the clinical trial. So I used not only standard surgical consent, but also special informed consent based on the data we learned from ARCHWAY. But your question, you said, well, how would people balance a surgical procedure versus an intravitreal injection. And I have to say that we have routinely underestimated the anxiety and the burden that goes along with these injections. And patients are really quite enthusiastic about not going through that anymore, about having an indwelling long-term drug delivery system where it just gets refilled every 6 months. And they are being told of the safety profile, and they're really still quite enthusiastic about it. So again, when we say balancing injections to an implant from our standpoint as a physician, I can tell you it seems like there's a big difference there, but we're underestimating the anxiety and discomfort that goes with those injections.

Did we answer all your questions, Michael?

Yes, that's great.

Next in the row would be Peter Welford from Jefferies.

So just sticking with Susvimo for a minute. I wonder if you can just talk a little bit more about the monitoring you were talking about for the safety events. In particular, I guess what I'm trying to understand is the sort of burden, I guess, on the patient for that in terms of does this require repeat visits. How many days, I guess, do you have to go back around -- how regularly does this require following up? Just trying to get an assessment, I guess, of how much we're actually reducing the visits or what the amount of monitoring that was required on a regular basis to reduce these side effects, particularly, I guess, at the moment, as experience grows. And then secondly, and forgive me, this is an ignorant question, but just to understand, with Susvimo, would you immediately go ahead and do bilateral implants? Or would you do 1 eye and then the other? How do you think about adoption of that drug -- that med device in the case of patients in that case? And then secondly, just one for Roche. Just understanding the port delivery system longer term with, I think, the DutaFabs, et cetera. Can you just perhaps outline for us how predictable, I guess, once you have the Phase I data -- now you've got the PDS approved with Susvimo. How predictable can you to be able to guess the pharmacokinetics, I guess, and understand the future drugs, how they could potentially fit in the PDS and the duration of dosing you've got? Or is it very much a case that for every drug, the PDS platform and how it applies and how the side effects, I guess, and frequency of dosing could very much differ from drug to drug, if that makes any sense?

Nancy, do you want to go first? And then Chris, you want to follow up on the platform maybe?

Sure. So I believe your first question was how am I going to follow these patients. And there are 2 key points. One is to follow the disease activity and the other is to follow any safety signals. And the data was so convincing that 95% of patients don't require supplemental injections that I really don't think that I need to follow patients for disease activity. I'm going to trust the data from the ARCHWAY clinical trial. Now it's true, and you framed this question very, very well that, is this really going to decrease the treatment burden. And the ARCHWAY clinical trial actually had to see patients monthly. And so clinical trials aren't the practice of medicine. And so it will be up to physicians to decide how to follow patients from a safety perspective to look at the implant. And this, of course, will be done on an individualized basis. So if the surgery was technically perfect and I don't see conjunctival retractions or erosions and I'm very happy based on my past experience with the PDS that my Susvimo patient looks good, I will not really see them that frequently, but educate them to call me for redness, irritation, blurred vision, floaters, anything at all. But it's quite true, clinical trials are not the practice of medicine and physicians using Susvimo will likely determine the follow-up schedule based on an individualized approach. Now I think your second question is what do you think the uptake would be if people want to buy lateral Susvimo. We never do cataract surgery on the same day. So this won't be happening in quick succession. But I can tell you that a colleague of mine in Iowa, who's in the clinical trials, his first Susvimo cases were actually on the fellow eye of patients from the clinical trial because they were so pleased with the PDS in the clinical trial that they wanted Susvimo in the fellow eye. And of 2 out of the 3 patients that I did 2 weeks ago are already asking about when they can have it in their other eye. So I think if it's a successful treatment for patients, I think patients will be requesting it. The third patient I did, she had a scleral buckle and vitrectomy in the past, and so she's not a fellow eye Susvimo candidate because those patients are outside of what was studied in the ARCHWAY clinical trial. So some patients will be good candidates for bilateral Susvimo and some patients won't be good candidates.

And I think one of the elements of patient feedback which we keep getting from investigators around the world is that there are patients in the clinical trials, such as ARCHWAY, keep asking when can they have their second eye treated. Obviously, we have greater restrictions on active clinical trials, but that is an important area that the patients have a desire for. So in terms of the second question on the DutaFabs and potential new molecules within the port delivery platform. So I think what we look for in new molecules to go into the port delivery implant, those which are stable and can be concentrated at high concentrations. And I think the DutaFabs, as an example, are a great platform in themselves, which are able to do that. As I said, we are potentially able to concentrate them at much higher concentrations than ranibizumab. So the answer to your question about how we can use the ranibizumab data to address how long these refill -- ultimately, your question is about how long can we go between refills with potential new therapeutics. I think we're still exploring. So I can't share any specific data. But what I can say is that I do believe it won't be incremental change that we'd be looking at as we look for new therapeutics in the port delivery implant. And I think the second element of that is, as we saw with faricimab going not an incremental improvement in durability, but a real step change in durability, 75%, 80% in the DME studies going to Q12 or more, it's difficult to model the impact of that for a DutaFab within the port delivery system. So I do believe that that's something where we can't model but we look forward to seeing the clinical data. But takeaways, there are elements that we can model and -- but broadly, we'll be looking for a step change in durability for these patients. So exciting platform. And outside of the DutaFabs, yes, we will be looking at other molecular platforms, which could potentially go into the port delivery system as well.

Peter, did we address all your questions?

Yes, that's great.

Okay. Then we would have a final question, which comes from [ Christopher Bianco. Christopher? ]

Can you hear me?

Yes.

This is Steve Scala from Cowen. Two questions. They're both market-oriented. First, on PDS, in the U.S., in what percent of the addressable population is reimbursement set up with payers since presumably there are drug and procedure components? And are there any intractable obstacles to setting this reimbursement up? We have heard that reimbursement or lack thereof has planted the adoption of PDS so far. And then secondly, on faricimab. The incidence of serious AEs is very low, but it's still higher than Eylea. I imagine your competitor uses this as a competitive or counter detail. I'm curious what Roche's counter to that counter detail is. And if this isn't the competitor's assertion, what has been the competitor's pushback so far?

Thank you, Christopher. A couple of points. On the Susvimo, as with all new products, we're going through all the reimbursement pieces and getting to permanent J codes. Right now there are temporary J codes. But we don't see any unusual obstacles in the way towards getting to those. They obviously take a number of months to put into place. And that's the same for Vabysmo right now. So Vabysmo will be having a temporary J code. And then over the coming months, we'll be looking to transit to a permanent J-code, which will obviously help practices. Most practices in the U.S. are adept to dealing with these temporary situations. That's happened for all previous launches as well. So they will be working through that. Now to your point on Vabysmo, to the clinical safety profile, I'll get Chris to comment or even Nancy further. But we've seen very few cases. These are very low numbers, so it's very difficult to talk about differences in the future. One thing we have done is looked at this, knowing the sensitivity of this in the community. We've looked at this very carefully. And we've been really transparent. We've been sharing the data since year 1. We shared that data within a couple of months of having the data, and we've been sharing the year 2 data. So the way I would frame this is that we're very confident but not complacent, and we'll continue to look at this as we move forward. As to what competitors are asserting about us, they will make many claims. I think we come back to the data and 3,200 patients across 4 Phase III studies and 2 indications simultaneously is where I think people will look. I don't know if Nancy or Chris have additional comments to add to that.

Maybe one comment and then over to you, Nancy. So I think as Nilesh said, we're extremely sensitive to this issue, which I think you referred to by SAEs around inflammation. And all of the -- what we actually did is we looked at with our reading center at all the reported cases of information. We went back and looked at the images. And there were, as Nancy highlighted, no cases of retinal vascular or occlusive retinitis across all of our programs as reported by physicians. So I think we're not anticipating any surprises as we saw with other recently launched anti-VEGFs. But Nancy, perhaps over to you.

Well, thank you, Chris. I was just going to reiterate what you said. What we really care about are those cases of occlusive vasculitis because those are really untreatable and can result in severe vision loss. And they're really poorly understood. And I don't see any of that in YOSEMITE or RHINE. And I appreciate that Roche has been very transparent because if you're not, it's only going to show up later. And so I really -- we all appreciate how transparent they have been and the extra effort to look in those cases. But I am reassured that there are no cases of inflammation associated with the retinal vasculitis or occlusive retinal vasculitis.

Okay. Then I would go on. We have 2 additional hands raised. The next 1 would go to Matthew Weston from Credit Suisse. Matthew?

It's a simple question for Dr. Holekamp, please. It follows on from Sachin's question about commercial uptake. I'd be very interested if you think that you and your colleagues have a warehouse of patients who may be poor responders to Eylea or who would, for various clinical reasons, you would be looking for new therapies as a source of patients to start faricimab on rapidly and whether or not we should expect essentially that warehouse to see very rapid uptake as soon as the drug is available. And then as clinicians evaluate how it's performed in that group, we can then see a more normal launch characteristic after that. I'd be very interested in that dynamic, please.

Sure. I'm happy to answer that. In fact, when this was discussed on the panel at Angiogenesis, I had my retina colleagues, including Jeff Heier saying, yes, I'm going to use this, I have patients, I have a backlog of patients that I've been waiting to switch to a different drug with a different mechanism of action. And so I think we will see a rapid uptake because we do have these patients. It is an injection, and we start with switching. And I think that we will see this rapid uptake. It will likely occur in patients for whom we know will get reimbursement. And so the permanent J code will certainly help with the uptake as soon as that is achieved. But absolutely 100%, and I think my colleagues feel similarly that there will be a rapid uptake, and we will begin with switch patients, and we do have those in our clinic.

And then Dr. Holekamp, if I could push you. Can you give us any guesstimate of what proportion of your patients that is? Is that a mid-single digit number? Is it bigger than that? Is it a handful? I'd be very interested.

Oh, definitely more than a handful. And when we talk about high VEGF need patients who need dosing every 4 weeks or every 6 weeks, that is, I would say, somewhere between 10% and 20% of the patients that I see. And those patients are absolutely looking for a treatment with a different mechanism of action that may help them decrease the treatment burden. And I think we saw that in the [ path ] survey, too, that we're looking for a different mechanism of action for a lot of our patients who still get frequent injections.

And so -- and the final question for today will go to Marietta Miemietz.

And apologies, I actually typed my questions into the Q&A. So I suppose you don't see them. I have 1 question for Nancy, just specifically on Slide 57. The rate of cataract looks increased. So I was just wondering, I mean, is that just an artifact? Or is there actually a mechanism by which PDS can damage the lens and how would that potentially evolve over time? And then just a couple sort of like market segmentation questions where I'll obviously be interested in everybody's perspective. One is you said about 50% of patients in the real world don't make the 20/40 visual acuity. And I was just wondering, if you look at the segmentation of those who don't make it, what proportion would you say are reasonably close to that and really incentivized to get there, so they can drive at what portion don't really care if they're a little bit above or below because they don't drive for other reasons? And what proportion maybe is sort of like so low that they also have an incentive to just really drive up their visual acuity just in order to be able to live a normal life? And the final question, if I can just sneak that on Slide 7. I was just wondering why the survey participants did not seem to think that better drying agents are needed. Is that because they actually think the existing agents are already so good at removing the retinal fluid? Or is drying basically not considered as critical to the disease progress as we always thought?

That's quite a lot there. So maybe, Nancy, do you want to take all the cataract question. Chris and I can follow up on some of the others.

Sure. So at the primary analysis, there was a numerical increase in cataract in the PDS arm compared to the monthly ranibizumab arm. And I don't know if statistics were run to see if the numbers that we have at end of study are significantly different. But you may not know this, but I've done cataract research in depth for 20 years. And the progression of cataract, it's a nuclear sclerotic cataract. It's very common in people in their 70s. And so we expect to see both groups getting cataract. The difference between the PDS arm and the monthly ranibizumab arm. Again, I don't know if it's statistically significant, but here's the key point. None of the cataracts were due to trauma from the drug implant. So none of them were traumatic cataracts. So I think that maybe some statistics can be done, sometimes turning on the infusion line during the surgical procedure itself may cause some additional vitreous liquefaction. Vitreous liquefaction is associated with nuclear sclerotic cataract. But I think the important point here is that none of them were due to trauma from the drug delivery implant.

Thanks, Nancy. Marietta, to respond to some of your other questions. So I'll start with the ASRS data where the need for fluid was a little bit lower. I think it's in the way the question is posed is one of your greatest needs. And I think here, you can see that durability of agents and new mechanism of action comes high. Obviously, those agents need to be able to show clinical benefit, and often that is supported by anatomic benefits consistently. And that is something we've seen in the studies outlined. To your question about 20/40. I mean interesting, I don't have all the segments that you mentioned. We look at this as really a place where as a threshold for vision that we'd love to achieve for many patients. They often come in with much lower vision into the clinical trials. And if we get them to 20/40 or so, that's seen as quite successful. It's about 70, 73 letters. And most countries around the world use that as a threshold for driving. We see that as a proxy for getting to success in a meaningful way. And I think a lot of countries define that as also one of the features for independence of living, where people get to the shops and do activities of daily living using the car. I hope that addresses your questions.

I was just wondering, somebody really -- sort of if somebody is just below -- I mean if they can basically get reasonably close to 20/40 with their existing treatment, do they still have an incentive to actually switch to sort of like make that numerical threshold or they say, well, it doesn't really matter if I'm a little bit above or below? I mean I'm just trying to understand how much inertia there might be in the market, how many patients might be like, yes, whatever, I'll just stick with the treatment I've always had, I have longer dosing intervals than I really should, just accept that I'm not going to get the best possible outcome. But it doesn't really matter because in any case, I don't drive any more and I can't really read the paper. I just -- I can still -- I don't know, cook a cup of tea, and that's good enough for me.

Nancy, do you want to take...

Can I comment on that? It's really interesting. I find that my patients across the spectrum of vision loss are highly motivated to protect and preserve as much vision as they can. And so I really don't divide it into patients getting to be 20/40. Everybody, if they understand what's at stake, which is really quality of life, every patient along that spectrum is motivated. So I would just say that I wouldn't compartmentalize it like that. Patients are highly, highly motivated to see better and see better with fewer injections.

Thank you. I think with that, we are at the end of our call. I would like to thank all the speakers again for their hard work and dedication to this field. And I would also say thank you to the audience for your interest in Roche's emerging ophthalmology franchise, which we are building, and looking forward to provide you future updates and wishing you a good day. Bye-bye.