Roche Holding AG (ROG) Earnings Call Transcript

[Audio Gap] Angiogenesis 2021 Webinar. My name is Henrik, and I'm the technical operator for today's call. Kindly note that the webinar is being recorded. [Operator Instructions] At this time, it's my pleasure to introduce you to Karl Mahler, Head of Investor Relations and Group Planning. Karl, the stage is yours.

Yes. Thanks a lot, Henrik. I can see that you have a strong interest in the call. I think we have -- you told me, Henrik, that it will be about 250 people minimum preregistered, so high interest. Thanks for your interest in our ophthalmology franchise. I hope that you're all safe and well. We have basically 1 hour to discuss our data, 30 minutes from our side on presentation, 30 minutes Q&A. This is a Zoom call. You can raise your hand, as Henrik explained. We will have the telephone line. I can see that we already have 5 participants raise the hand. [Operator Instructions] And of course, you also can send me an email if you want, karl.mahler@roche.com. Then I will read your questions to the audience. If you could kindly go to things about -- yes. So ophthalmology and strategy update will be given by Atul, and Atul is our Global Franchise Head for the Ophthalmology, so basically the commercial arm of our ophthalmology franchise. Then he will be followed by Chris Brittain. He's basically the [ pawn don ] to Atul, what the product development is concerned. He'll give us an update on the pipeline. Then we have -- very pleased that we have Nancy Holekamp with us. She is a Retina Specialist and a Faricimab Clinical Investigator, so she was working on the DME and nAMD data Phase III results of these. And thanks that you found time to join us here today. Next slide. And before I start with the event -- and of course I want to do some advertisement. It's a good opportunity here. First is that we will have 2 more events, one will be on Evrysdi. There is the SUNFISH Part 2 in Type 2, Type 3 SMA patients. This will be done because of the so-called MDA meeting. This is the 19th of March. And again, starting Q1 -- it's a diagnostic event, the 23rd of March. That is all key people in Diagnostics presenting their business case, Thomas Schinecker -- and this is really a cool event because we have a complete new software and you can enter the room and you can go up the stairs, and you can take a seat virtual and look at the screen. You can go to any kind of rooms. You can take a megaphone and say, "Everybody, listen please to me." It is really a very cool setup. So I'm very much looking forward to this new setup, which we are very excited to share with you. The pipeline is building up nicely over the past year. So we have about -- hopefully by the end of 2021, 22 NMEs. And the focus of today, you can see with the green ticks, is on all of the deliverables in ophthalmology, so namely faricimab and the Port Delivery System. If you could kindly please have the next chart. So faricimab and PDS are addressing key medical needs. There was a very positive feedback over the meeting over the weekend, the Angiogenesis Meeting. And these data refers to pivotal trials only, so-called landmark trials with a reputable trial setup. [ They use ] criteria and so on. We have only listed those which have met also the primary endpoints. So what you can see on this chart is basically 3 important messages from my point of view, 3 basically key takeaways. The first one is the benefit for patients who was kept with short infusion intervals. You can see that at the left upper part here. The next key takeaway is with faricimab. We have, for the first time, either 50% of patients maintained for 4 months or 80% at 3 months. So in summary, it's a significant -- more patient benefit longer. And with the PDS, the upper right part corner here, close to 100% for 6 months. And by exploring more opportunities, and Atul will talk about this data on opportunities for longer than 6 months, because from the pivotal trial, we actually know that they can go substantially longer. So last but not least, I wanted to thank all the speakers. I also wanted to thank Lisa and Claudia for taking the lead on this event. And with this one, Atul, over to you, please.

Thanks, Karl. Good morning, and good afternoon, everyone. Many thanks for joining us today. It's great to see the growing interest in our exciting ophthalmology portfolio. So today, I'm going to share with you 3 key factors shaping our ophthalmology strategy. Number one, why we focus on reducing treatment burden in the real world; two, how faricimab and PDS complement each other as a portfolio; and finally, how Roche is building an industry-leading retina pipeline. And we are lucky to have Dr. Holekamp with us today, since she's an investigator on both faricimab and PDS programs. And it will be great to hear Dr. Holekamp's perspective, both as an investigator and as a clinician managing a very busy practice. Now as you're well aware, the retina market was about $13 billion in 2020. We expect the market to continue to grow, driven by aging population, the unfortunate but rapid increase in diabetic eye disease and exciting product innovation. Interestingly -- it's very unusual for biopharma -- the entire category of $13 billion is made up of a single class of agents, anti-VEGF monotherapy, which makes it very ripe for disruption through innovation. Next slide, please. So the ASRS runs a very well-regarded Annual Preferences and Trends Survey with its members from around the world. Every year, the #1 unmet medical need identified by the retina specialists is the need to reduce treatment burden for patients. Now despite the best efforts of our physicians and patients, unfortunately we are not able to replicate the clinical trial vision outcomes in the real world, due to treatment burden. Now there are many reasons for this, but the key issue is that most patients, given their advanced stage in neovascular AMD and comorbidities in DME, find it really, really hard to make it to the retina clinics on a monthly by monthly basis. In the past year, this has become painfully obvious due to the COVID pandemic. You can see on your left, U.S. data, with currently approved therapies, that patients who received less than 7 injections tend to lose vision. On your right, you see that the same is true in Europe. You see relatively better outcomes in U.K., where patients receive more injections. But already by year 2, treatment fatigue kicks in and patients start losing vision. Next slide, please. Now improved durability, we believe, will go a long way in improving real-world outcomes. We see faricimab and PDS platform as 2 critical components of a very complementary portfolio. Our strategy is to offer best-in-class durability and more choice to patients and retina specialists who haven't really had too many choices beyond anti-VEGF injections for the past 15 years. So I know a big question is, how do we expect the 2 products to be used? We would largely be guided by physician and patient choice, even as our teams have built a very comprehensive positioning strategy. Now for more than 70% of AMD and DME patients, faricimab is potentially the first-line intravitreal injection of choice that will enable maintenance of just 4 injections a year. For the 30% of patients that need to come in monthly or bimonthly for IVT injections, or for patients who do not prefer IVT injections, PDS is a great option providing continuous delivery with just 2 refills a year. The PDS platform is an exciting new paradigm, and we are planning to explore extended durability beyond the 6 months we already have seen in Archway. Next slide, please. In terms of pipeline strategy, our focus is squarely on blinding retinal disorders with high unmet need. In addition to the high prevalence of retinal disorders, we are also developing gene therapies for monogenic conditions. In the near term, we are concentrating on filing and global launches of faricimab and PDS and building a global retina presence on the back of 2 best-in-class products. In Phase II, we have an oral drug that may potentially allow early treatment of diabetic retinopathy. We also have 2 different approaches to GA. The first one is an anti-Ang2-A1 antibody targeting non-complement pathway and the other one is a systemic complement factor B inhibitor. In Phase I, we have 2 bispecific DutaFabs for AMD. One of them combines a novel MOA with anti-VEGF and the other one is our anti-VEGF Ang-2 DutaFab that is being accelerated into the PDS platform. One other thing, keep an eye on the Phase I study of the new molecular entity in DME. More to come on this later this year. And with Spark as a part of the Roche group now, we foresee further acceleration of gene therapies in our ophthalmology portfolio. Next slide, please. A key pillar of our ophthalmology strategy is our personalized health care initiative. We are focusing on 3 areas: remote vision monitoring using our patient-focused app, MyVisionTrack, to enable at-home detection of vision changes. We believe remote vision monitoring tools will be important to deliver the full benefits of PDS and faricimab. Our scientists are also using our internal clinical data, which is almost about 3 million images, and external data sets to come up with retinal imaging algorithms that will help predict disease progression. Finally, we plan to release the MyVisionTrack tool as we launch PDS and faricimab to allow patients and physicians to remotely monitor their vision at home. Next slide, please. So we last met in May. Since then, our ophthalmology strategy execution is making rapid progression with 3 positive Phase III readouts in the last 10 months. For the first time in retina, we will submit a joint filing of faricimab for the AMD and DME indications in the first half of this year, and PDS is well on track for U.S. approval this year. The PDS, DME and DR Phase III studies are recruiting very rapidly, and we are also initiating ex-U.S. PDS studies that will look to further extend durability. And finally, we'll start the faricimab RVO Phase III studies later this year. We are particularly pleased to accelerate our VEGF/Ang-2 DutaFab program that essentially combines the benefits of faricimab with PDS, and that's already in Phase I. In conclusion, the Roche ophthalmology team is very excited about bringing faricimab, PDS and our 10 new molecular entities to the retina community and patients with blinding diseases. Thank you for your attention. And now I would like to hand over to Chris to provide a more detailed look into the clinical data.

Thanks, Atul, and good afternoon and good evening to everybody on the line, and thanks again for joining. If we could have the next slide, please, Lisa. Thank you. So what I'd like to do today is just for the next kind of 5 to 8 minutes or so, just really go through a little bit of the data from the Port Delivery System, which was presented at Angiogenesis, and also set the stage for why we're entering or why we entered faricimab and a little bit of the background of faricimab and the importance of the novel and exciting mechanism of action that we've got there. So really starting with the Archway Phase III design for the Port Delivery System. As you all recall from the presentations over the last few months, the Archway was a randomized controlled trial. And as you recall, the Port Delivery System is the permanent refillable ocular implant, which is -- in this RCT against intravitreal ranibizumab given every month, was given refills every 6 months. If we go on to the next slide, please. So the last time that we presented efficacy outcomes, we had them up to week 40. So this new data is bringing patients a complete data set for everybody out to week 60, and a subset of efficacy out to week 72, and you can see the patient numbers on the bottom there. And this is really exciting data. What we see here is continued long-term, high-quality efficacy data. Just as a reminder, patients, when they entered the study, they'd already had a median of 5 intravitreal anti-VEGF injections before entering the study. And therefore, this flat line is exactly what we'd hope for and exactly what we expected. So maintenance of vision is really continuing well out to week 72. If we could go to the next slide, please. And then again, here, we see the outcomes in terms of retinal thickness. And similarly, it's great to see the flat line continuing all the way to week 72. And here, we have 2 complete refill cycles, and both very little variability in retinal thickness and very good patient retention, as you can see from the numbers along the bottom. So we've seen great BCVA data, great central retinal thickness, maintenance data. And if we can go on to the next slide, what we see here is other aspects of the treatment burden. On the left-hand side, we can see that there are up to 5x fewer treatments given in the PDS arm. So the treatments given in the PDS arm include the initial fill and subsequent refills and any supplementary treatments. And in the ranibizumab arm, obviously, patients were given a monthly intravitreal ranibizumab. On the right-hand side, we can see the consistency of treatment effect in terms of need for supplementary injections. We already knew from previously presented data that only 1.8% of patients required supplementary refills in the first 6 months. And that level maintained -- was maintained at a very low -- that rate was maintained at a very low level, with 95% of patients in the second treatment interval remaining supplemental treatment-free. Data is ongoing in terms of collection for intervals 3 and 4. If we go to the next slide, please. On the safety side, a couple of -- 3 points of note here. First of all, I think we can see that very few additional adverse events have happened since the previous datacut, which is clear on the left-hand side of each of these charts, onset after 40 weeks. We note a slight numerical difference in terms of cataracts. And I'll draw people's attention to the fact that, a, these were predominantly -- progression of preexisting cataracts for patients. And importantly, there were no cases of traumatic cataracts in this aging population. The single case of endophthalmitis after week 40 in the PDS arm, we looked at that in close detail, and that patient had a -- this was this patient's second incident of endophthalmitis in the study eye. On further examination, this patient had what is known as HLA-B27 positivity, which is a protein on white cells related to high risk of systemic inflammation, such as uveitis, such as ankylosing spondylitis. Of note, there was one -- there was the first case of endophthalmitis in the intravitreal ranibizumab arm. The last piece that I want to draw attention to, which Carl -- Dr. Regillo spoke about in great detail in Angiogenesis is this -- 3 patients in this datacut underwent implant dislocation. As Dr. Regillo discussed, we've had now a total of 6 cases of implant dislocation across over 400 patients in our study programs. Now importantly, we've done an in-depth analysis, and we have, first of all, found the root cause, we believe, which was the length of the incision. The incision historically was between 3.5 and 3.7 millimeters, and 5 out of 6 of these cases of implant dislocation were related to patients whose incision was over 3.7 millimeters. So we've identified those patients and have mitigated that risk by updating the surgical instructions for use back in June of this year. So we believe that -- and since that time, we've seen no further cases of implant dislocation. So we're comfortable, in a manner similar to other surgical procedures, that we continue to optimize the surgery around this device. And we believe in terms of the dislocation, we've got the root cause and we've taken effective measures to reduce the risk of that. So if we go to the next slide, please. So in terms of -- we've now seen the strong efficacy in terms of BCVA, anatomical outcomes, acceptable safety profile. And now what do patients think of it? Well, I think this is amongst the most exciting data that we've got, the Patient Preference Questionnaire. Patients with the Port Delivery System were asked whether they prefer the PDS or their previous intravitreal injections. And as you can see here, over 93% of patients did actually prefer the Port Delivery System, and only 1% preferred the intravitreal injections. So when it comes down to patient choice, I think this is a clear signal that the patients do really -- are very happy with the Port Delivery System as a treatment. So what's next for the Port Delivery System? Well, Atul has already discussed, we are ongoing rapid enrollment in the DME studies and the diabetic retinopathy studies. We've got ex-U.S. studies to be initiated in neovascular AMD and we are expecting the U.S. approval in 2021. So if we next go -- next slide, please. Now since the success of the Archway study, we've actually progressed the Port Delivery System as a platform technology, really focusing on its use, as you can see here, with DutaFabs. Now DutaFabs are, as Atul mentioned, a combination of the bispecific element that we see with molecules like faricimab and also they're the size of Fab fragments like ranibizumab. And we do have 2 of these DutaFabs with novel mechanisms of action in clinic today, and we're excited to be accelerating the angiopoietin-2 anti-VEGF DutaFab into the Port Delivery System for future use. So stay tuned for that. Now changing tack slightly. This -- as we've seen over the past few years, and Atul talked about earlier on in terms of the impact of treatment burden on patients' ability to have consistent and correct treatment, we're really looking at ways to address how to reduce that treatment burden and how to improve time between doses. So our first look was at the HARBOR study and the READ-3 study in neovascular AMD and DME, respectively. And what we saw here, on the left-hand side, is the HARBOR study, where we had a 4x higher dose of ranibizumab versus the classic 0.5 milligram dose, and we saw, a, no treatment benefit; and b, no upside in terms of durability. And for that reason, we continue to explore new mechanisms of action. So if we could go to the next slide. And this is where we refocused on the angiopoietin-2 pathway and the Tie2 pathway. Now just briefly here, in a healthy retina, angiopoietin-1 stabilizes the vasculature. And when there's an angiogenic switch such as occurs with -- in times of ischemia, such as patients with diabetic retinopathy, diabetic macular edema, retinal vein occlusion, neovascular AMD, we see angiopoietin-2 levels rise. And these levels of angiopoietin-2, it displaces the Ang-1 and this -- the impact of these high levels are destabilization of pericytes, leukocyte migration, which causes inflammation, and also an increased sensitivity to VEGF. So therefore, if we go to the next slide, we worked on development of faricimab, and this is the faricimab molecule. So we target the Ang-2 on one arm. We target VEGF-A on the other arm. And these combined, both independently and simultaneously, both molecules. And at the same time, we have the modified Fc fragment, which reduces the inflammatory potential of the molecule and reduces systemic exposure. So with that background, I'm really excited to introduce Dr. Nancy Holekamp, who's the Director of the Retinal Services of the Pepose Vision Institute at St. Louis, Missouri. Now Dr. Holekamp has been our -- an investigator on countless trials for many companies over the years, but, of note today, she's been a very active investigator on the Port Delivery System programs and also the faricimab programs. So without further ado, thank you very much, indeed. And Dr. Holekamp, over to you.

Well, thank you, Chris, and good morning, good afternoon, good evening to everyone on the call today. As an investigator for the YOSEMITE and TENAYA clinical trials, it is my great privilege and pleasure to discuss the study results on behalf of all of the investigators for YOSEMITE and RHINE, and TENAYA and LUCERNE. And I will be presenting the data that was presented at the Angiogenesis Meeting last weekend. Importantly, this is the first time that pivotal Phase III study data has been released simultaneously for 2 major retinal diseases. I can't recall this being done in my 25 years of retina practice before. These studies were similar in their primary endpoint; that is, best corrected visual acuity change from baseline. And in YOSEMITE and RHINE, non-inferiority was achieved for faricimab when dosed up to Q16 weeks compared to aflibercept dosed Q8 weeks. And for TENAYA and LUCERNE, non-inferiority was achieved for faricimab up to Q16 weeks versus aflibercept dosed Q8 weeks. And I think that these are really important durability results. They were consistent results between both sets of studies, and you'll see that secondary analysis results were also consistent with the main analyses. Now speaking about durability. Again, Atul mentioned the treatment burden that is just so ominous for patients with these 2 diseases. In YOSEMITE and RHINE, there was an arm called PTI, which is essentially treat and extend. And about 50% of patients could be extended to Q16 weeks and 70% of patients could be extended to 12 weeks or more. When we look at TENAYA and LUCERNE, we found that 45% of patients were maintained on Q16-week dosing and about 80% of patients were maintained on 12 weeks or longer dosing intervals. And that's really remarkable new information for my field. And then when we look at safety, of course, physicians are very keen to look at safety profiles. There were no new or unexpected safety signals identified. And specifically, in these 4 large clinical trials, there were no cases of retinal vasculitis or retinal occlusive events. Next slide. So this is the global footprint for the YOSEMITE and RHINE clinical trials. And they're the largest registration clinical trials in diabetic macular edema, including around 2,000 patients across 353 study sites. And over the weekend, there was a presentation of additional slides with the study details, which I'll just summarize here briefly. Thanks to randomization, the baseline demographics and ocular characteristics were similar, not only across treatment arms, but also across the studies. And the baseline diabetic retinopathy severity was well-balanced across the treatment arms. Now while there were some missed doses due to COVID-19, sensitivity analyses were consistent with the primary analysis, and the discontinuation rates were low and similar to what we would expect from Phase III studies in this population. Next slide, please. So here's the study design, and patients were randomized equally into 1 of 3 arms. First, there's the active comparator, which is aflibercept dosed on label. Then there's a faricimab Q8 week arm. So after 6 loading doses, patients received faricimab every 8 weeks. And then we see the faricimab 6-mg PTI arm, which we will discuss further on the next slide. The primary endpoint was change from baseline in best corrected visual acuity, averaged over 3 time points, weeks 48, 52 and 56, with the final visit being at week 100. And importantly, there is a long-term extension study, which will give us data with faricimab in DME up to 4 years. Also importantly, this study looked at both treatment-naive patients and previously treated patients, but the previously treated patients were capped at about 25% of enrollment. Next slide, please. So here's a little information about that personalized treatment interval, or PTI, arm of the study. So it's well-established that physicians like myself use a treat-and-extend protocol when using anti-VEGF agents. And the PTI arm was really based on the treat-and-extend concept to individualize the dosing interval, but do so within the confines of a randomized clinical trial. So it was automated, standardized, objective, but also preserved study masking. So after 4 loading doses, if patients met the criteria of a CST value of less than the threshold for entering the study, which was 325 microns, then they could be considered for treat-and-extend. Next slide, please. So here are the results. This is a forest plot that shows the primary endpoint was met, the BCVA gains with faricimab dosed Q8 weeks or up to Q16 weeks were non-inferior to aflibercept dosed Q8 weeks, and we see the noninferiority margin at minus 4 letters. Next slide, please. So these results are the main change in best corrected visual acuity from baseline, and we see consistent improvement in BCVA, from baseline all the way through week 56. And we see that these lines are essentially on top of each other. And when we average the BCVA change at weeks 48, 52 and 56, we see significant substantial improvement, ranging from 10.3 letters to 11.8 letters, across all arms of the study. Next slide, please. And when we take that population of treatment-naive DME patients, we see essentially the same results: consistent, continual visual acuity improvement; the lines are essentially on top of each other; and we see that averaged over weeks 48, 52 and 56, the significant visual acuity gains range from 10.5 letters to 11.7 letters on an ETDRS eye chart. That's more than 2 lines of visual acuity improvement. Next slide, please. So let's look at the anatomical results. We see the change in central subfield thickness, that's on the OCT, from baseline through week 56 consistently favored faricimab. And we see that there was a matched phase for dosing early in the study, and we see a separation of the lines through week 16 at least in YOSEMITE, and a signal at least in RHINE. And when we look at the Q8 week dosing of aflibercept and we look at the Q8 week dosing of faricimab, we see the expected sawtooth pattern. But even here, the CST results favor faricimab. Next slide, please. So importantly, when it comes to durability, I want to focus your attention on the orange highlights. It shows that greater than 70% of patients were on at least Q12-week dosing intervals by week 52, and about 50% of those patients had extended out to Q16-week dosing. Next slide. So now we're going to look at a couple of anatomical outcomes. And I think these really speak to the added advantage of targeting the Ang-2 pathway. When retina specialists treat DME, we're really -- our goal is to treat to dry. So we're looking at the absence of DME. And here, the absence of DME is defined as a CST of less than the threshold level for entering into the study, which was 325 microns. And we see it every time point through week 56: these results favor faricimab, whether it was dose Q8 weeks or, according to the PTI, extend and -- treat-and-extend treatment paradigm. Next slide, please. Another way of looking at DME is actually looking at the OCT and looking for an absence of intraretinal fluid, which retina specialists understand is very damaging to vision. And here we have time points of weeks 16, 48, 52 and 56. And again, this data favors faricimab dosed either Q8 weeks or according to the personalized treatment interval. We see higher percentages of faricimab-treated patients who are without this very damaging intraretinal fluid. Next slide, please. Now to be complete, we're showing data here on the absence of subretinal fluid. We see it's high across the board throughout all the arms of the clinical trial. But subretinal fluid really isn't a big component of diabetic macular edema, so these results are not surprising to me. On the right-hand side of the slide, we see data for that 2-step improvement in the diabetic retinopathy severity score at week 52, and the results are comparable across all arms, but I really look forward to seeing the full 2-year data set in this regard. Next slide, please. I have 3 slides on safety, and the first one deals with adverse events of intraocular inflammation. And what you'll see is that these rates are low and comparable across the arms. When we look specifically at endophthalmitis, the rates are lower than what we would expect from a large clinical trial such as YOSEMITE and RHINE. And we see that the IOI event rates were about 1.3% for faricimab and 0.6% for aflibercept, respectively. Next slide, please. Now importantly, there were no cases of retinal vasculitis in either of these very large studies. And when we look at retinal occlusive events through week 56, none of them were associated with intraocular inflammation. And I want to remind everyone that these are diabetic patients with DME, and DME is a well-known biomarker for some of our sickest patients. And these rates are low and balanced across all arms of the study. Next slide, please. And then finally, other important safety events were infrequent. And when we look at the APTC events, again we see that they're well-balanced across all arms of both studies. Next slide, please. So in summary, YOSEMITE and RHINE demonstrates the advantages of faricimab in both durability potential and anatomic endpoints. The hypothesis was that faricimab could improve vascular stability via simultaneous neutralization of both the Ang-2 and VEGF-A pathways, and I think that these clinical studies do confirm that hypothesis. Both studies met their primary endpoint, in that faricimab dosed Q8 weeks or up to Q16 weeks was noninferior to aflibercept dosed Q8 weeks. What I really found impressive was the anatomical outcomes, with 50% of patients on Q16-week dosing and 72% of patients on 12-week dosing or longer. The CST outcomes favored faricimab. The absence of DME favored faricimab. The absence of intraretinal fluid favored faricimab again. And these results were reproducible across YOSEMITE and RHINE in studies with nearly 2,000 patients, and it's really the large size of this study that gives us confidence about these results. And again, safety is always very important. Faricimab was well-tolerated, and very importantly, there were no cases of vasculitis or occlusive retinitis reported. So I look forward to the complete 2-year data set that will be coming out of YOSEMITE and RHINE. And I also look forward to the ongoing extension studies that will give us a full 4 years of data of faricimab in diabetic macular edema. Next slide, please. So now we're going to turn our attention to TENAYA and LUCERNE for neovascular AMD. Again, this is the global footprint of these studies that enrolled over 1,300 patients at over 270 study sites. And at the Angiogenesis Meeting, additional slides were used to present study details. But again, thanks to randomization, the baseline demographics and ocular characteristics were well-balanced across the treatment arms and studies. The COVID-19 discontinuation rates were low, and the robustness of the primary analysis was then analyzed through supplemental and sensitivity analyses, and all of these results were consistent with the primary analysis that I'll share with you now. Next slide, please. So here is the study design for TENAYA and LUCERNE. Here we have 2 groups. Patients were randomized equally. One is an active comparator with aflibercept being dosed by the standard of care and on label. And faricimab is in an arm that has 4 initial loading doses, then at week 16, there's a break in treatment. Then at weeks 20 and 24, disease activity assessment occurs based on best corrected visual acuity and CST criteria, but also the investigator's evaluation. And it's at week 20 and 24 that patients either fall into a Q8-week dosing arm, a Q12-week dosing arm or a Q16-week dosing arm, and they remain on that dosing interval for the remainder of the first year. And the primary endpoint is the change in best corrected visual acuity, again averaged over weeks 40, 44 and 48. It is a 2-year study, and the final visit is at year 112. Next slide, please. So here are the primary endpoint results. The forest plot shows us that the BCVA gains with faricimab dosed up to Q16 weeks was non-inferior to aflibercept dosed every 8 weeks. Again, we see the non-inferiority margin at minus 4 letters. Next slide, please. Here, we see the mean best corrected visual acuity change from baseline. And we see an early and impressive improvement in vision that is maintained over the first 48 weeks of the study. And again, we see that the lines are very similar to each other, and we -- results in anywhere from 5.1 to 6.6 letters of visual acuity improvement. And if -- for people who think that this is not a lot of visual acuity improvement, I want to remind people that patients could have good visual acuity upon entry into the study of 20/32, so we're likely seeing a bit of a ceiling effect with these visual acuity results. But again, the early visual acuity gains were sustained through the first 48 weeks of treatment. Next slide, please. Now importantly, when we are extending treatment intervals to Q16 weeks in patients with wet macular degeneration, we want to be sure that we're giving every patient the opportunity to be the big winners, to gain that greater than or equal to 15 letters of visual acuity. And we also want to make sure that we're not exposing patients to a risk of significant vision loss. We want to avoid that 15-letter BCVA letter loss. And we see that with up to Q16-week dosing, the results in both of these vision endpoints is identical to aflibercept dosed Q8 weeks. Next slide, please. This is the only slide on anatomy that I have to show you today, and we see meaningful reductions in central subfield thickness on the OCT with faricimab dosed up to Q16 weeks. We do have some matched phase dosing in the first 3 doses. We see perhaps a slight signal of better drying effect from the faricimab arm, but again, this is the only slide that I have on data right now to show you regarding anatomy. Next slide, please. But here we see the durability results. And again, I want to draw your attention to what's highlighted in orange, and we see almost 80% of patients could be dosed at Q12 weeks or Q16 weeks. And again, about 45% of patients could be dosed at about 40 -- about 45% of patients could be dosed at Q16 weeks. This is important because it reduces the median number of injections. It reduces the treatment burden. There were 8 injections for aflibercept and 6 injections, by week 48, for faricimab. Next slide, please. So here we have the safety information. And again, adverse events of intraocular inflammation were low and well-balanced across the arms of the study. And here, when we look at this table, we see that the IOI event rates were about 2% for faricimab and 1.2% for aflibercept. And in my experience in clinical trials, these are low and expected and acceptable for injection treatments. Next slide, please. Importantly, there were no cases of retinal vasculitis in any of the arms of the studies. And when we look at retinal occlusive events, we only see one such event, it's in the faricimab Q16-week arm. There was one case of a retinal artery embolism. This was a patient who had an asymptomatic Hollenhorst plaque, and it was not associated with inflammation, and it was not felt to be related to the faricimab injection. Next slide, please. When we look at other important safety events, they're infrequent. We see lots of zeros on the top table. And when we look at patients with any APTC event, we again see that these numbers are small and well-balanced across the treatment arms. Next slide, please. So in conclusion, we see that faricimab demonstrated greater than or equal to Q12-week dosing intervals in almost 80% of patients in the first year. And in doing so, it met its primary endpoint in that non-inferiority was met with faricimab dosed up to Q16 weeks compared to aflibercept dosed at Q8 weeks. But we also saw increased durability. About 45% of patients were dosed at Q16 weeks and 79% of patients had a dosing interval of Q12 weeks or Q16 weeks. We also saw meaningful reductions in CST, and I look forward to additional analyses of this anatomical data. The results were reproducible across the TENAYA and LUCERNE studies, involving over 1,300 patients. Again, the large size of these clinical trials really bring confidence when it comes to the efficacy endpoints as well as safety, and faricimab appeared to be well tolerated with no events of retinal vascular occlusive disease. I look forward to the year 2 data set for TENAYA and LUCERNE, as well as the extension studies that will give us up to 4 years of data on faricimab in AMD. And with that, I will turn things over to Karl for the Q&A.

Thank you. Thank you, Nancy. Maybe, Henrik, you could kindly remind us on how the participant [ members ] can ask questions.

[Operator Instructions]

Okay. Thanks a lot. Nancy, we get questions from investors also how meaningful these data are now for patients, because of -- there's a lot of figures and percentages and so on. But how do you believe, how important are these things for patients? I mean, so you have seen a lot in your life as an ophthalmologist. So how would you frame it?

These are significant. These are impressive. I saw the extended interval dosing to decrease the treatment burden. I mean we're pushing the envelope. We used to dose monthly. Then we went to Q8 weeks. Now we're -- have studies pushing it out to Q12 weeks. This is the first registration clinical trial data at Q16 weeks, which I just want to remind everyone is about 4 months. That's substantial, not only for physicians who have to give these injections so frequently, but for patients. Decreasing the treatment burden is really the name of the game. We have to find a better way to improve and maintain vision in patients. And if we can do this with fewer injections, everybody wins.

Thank you. So I can see that we have now a lot of participants who raise the hand. So maybe we can kindly limit the number of questions to two, if that is possible? On the one side, maybe we can try to keep our answers short from our side, so that we have -- at least a bit of a chance to get through the list of questions. With this one, Wimal, you would be the first one.

Great. Wimal Kapadia from Bernstein. So can I just first touch on the inflammation. So I appreciate the differences were relatively small. But can I just ask what do you think were the potential drivers for the higher intraocular inflammation between faricimab and Eylea? Is it simply just a greater volume of drug? Or is there a mechanistic driver for this? I'm just curious, as that could help us inform how to think about the safety for the Eylea high dose trials that are ongoing? And then my second question is, you mentioned that subretinal fluid is not a major issue in DME. But can I just ask how we should think about subretinal and intraretinal fluid with respect to long-term outcomes in patients? What matters more to visual acuity long-term, and why? And I ask because the benefit seen in faricimab over Eylea is in the intraretinal fluid. So does that drive major positive outcomes for visual acuity longer term?

Yes. I think Nancy, this is going to you.

Sure. I'll address the inflammation question first. So you're assuming that there is a significant difference. In these clinical trials, that is not a significant difference. If we had seen inflammation rates exceeding 4%, I think I'd have some reason for concern. But having done so many clinical trials, I can tell you that in other registration trials for multiple agents, the inflammation rate at 1 year has been 2% in other trials for other agents. It's been 1% for other trials and other agents. Everything you're asking about is within the normal expected range and doesn't raise any red flags for me. As for what's causing the inflammation? Well, that's the $64,000 question. Nobody really knows. We know that it occurs, and these events are so rare and so isolated, it's very difficult to study rare events. So I can't tell you if it's related to the drug. I can't tell you if it's related to the host. I can't tell you if it's related to the volume. All I can tell you is that the results of the studies I've shown today are well within the expected range and do not raise any safety concerns for me as a clinician. So that's for the first question. The second question deals with subretinal fluid and intraretinal fluid. And I've done a lot of research in this area. And in age-related macular degeneration, it does matter. Intraretinal fluid is more damaging to vision. Subretinal fluid is less problematic as long as eyes are undergoing active treatment. But what I said in my presentation is correct. Subretinal fluid is not a large component of diabetic macular edema. We just don't see much. And I'm going to encourage Roche/Genentech to include the baseline levels of subretinal fluid in that one graph that I showed, because if you could see that very few patients, even at baseline had subretinal fluid, you'll see that the results really just aren't that impressive. So the take-home message is that in AMD, there is subretinal fluid and intraretinal fluid that matters and has to be considered, and the compartment is important. But for DME, it's really an intraretinal fluid disease and the ability to eliminate intraretinal fluid is of paramount importance. And I find that the data that I showed you on anatomy today was very compelling.

If I could add one comment -- it's Chris. Just add one comment there. Let's just say that we only received the new vascular AMD data about 3 weeks ago, so we are prioritizing some of these additional analyses to be looking at the intraretinal fluid and the subretinal fluid for neovascular AMD, which we will share at upcoming congresses. Yes.

Richard, you would be next.

Two questions, please. One, just on inflammation and at the Congress, there was some discussion about what caused the inflammation for Beovu: that it might be because it was a single chain sort of antibody fragment rather than a full chain -- full antibody. So I suppose the question is, what do you think about that? And also from a Roche perspective, is this a concern for the DutaFabs? They're clearly very much smaller. Would that be a concern going forward? And then the second question, there were also presentations at the meeting on 2 gene therapy products. Those look very much better than previous attempts. So just thoughts from Dr. Holekamp about what you would need to see to -- in terms of durability, efficacy and safety for that to be a meaningful treatment option for your patients?

Yes. Thank you for your questions, Richard. I can already save some time for us because actually, I don't think it would be appropriate for us to comment on Beovu because that is really not our trial and our drug, and I don't want to bring here anybody into difficulties. But maybe we can take the other questions. I think Nancy, the question is again to you.

Sure. Let me address gene therapy. I have to say, I'm very bullish on gene therapy. I'm very excited about anything that helps me take better care of patients. But I can tell you that gene therapy has a longer runway. I think what we saw at the Angiogenesis Meeting was still very early data. And although I'm very excited about it, there are still some hurdles. Delivering gene therapy with a vitrectomy is a much larger surgery than putting in a PDS. I've done both. I can also tell you that there are rates of inflammation, that we're all highly sensitized now to intraocular inflammation and the long-term consequences. And so I think those hurdles have to be overcome. But at the end of the day, gene therapy will still only serve a subset of our patients with AMD and diabetic macular edema. And we will still need our injections and the Port Delivery System and other agents to take care of our patients. I'm all for putting lots of arrows in my quiver, but gene therapy will be just another arrow for me.

And maybe on DutaFabs?

Yes. Maybe just a brief comment on the DutaFabs. I think it's a really important question and one which we pay close attention to. We're -- as I said, we're in Phase I with 2 different DutaFabs, and we'll continue to be looking very closely at both the -- any low-grade inflammation rates and pre-existing auto antibodies. So something which we look very, very closely at, but we've got nothing further to share at this time.

Yes. Emmanuel Papadakis will be next.

Emmanuel Papadakis from Deutsche Bank. Maybe a follow-up on the gene therapy question, but I'd ask it from the Roche perspective. You've not been particularly vocal about why you would not consider developing your own gene therapy solution in this space. So perhaps you could just give us the Roche perspective on that opportunity? And then perhaps a question for Dr. Holekamp. Again, thinking about other options in development, just love to hear your view on how you see the positioning of PDS relative to long-acting polymer VEGF options in development. A few cases of dislocations surely highlight some of the risks of undergoing a surgical procedure. What's likely to be the medical preference for those 2 options in patients looking for very long dosing intervals?

So Emmanuel, maybe I'll answer the first question in terms of gene therapy. Our, actually, view matches up with what Dr. Holekamp just said. We believe for polygenic conditions like wet AMD, like DME, you need lots of arrows in your quiver. So we are looking at gene therapy options for neovascular AMD. Let's put it this way, we are looking for potentially second-generation options too there, where we can combine some modalities and new mechanisms of action there. You can see in our pipeline, we are fairly active with monogenic retinal conditions. So you already have LUXTURNA on the market through Spark. We have 2 Phase I programs with 4DMT in Phase I that are recruiting nicely. And then we have a choroideremia program with Spark, again in Phase I. So we have 3 Phase I programs for monogenic retinal conditions. So we look forward to the modality with Spark. Again, in the Roche Holding portfolio, we believe we can accelerate gene therapies even faster in ophthalmology. Dr. Holekamp, you want to talk about positioning?

Sure. I think...

Before you, Nancy, answer the question, we have just for the completeness, there is another hand raised, which had more or less the same question on the preferences, patient preferences, which is Michael Leuchten from UBS. I just wanted to also say that there is another question going the same line from somebody else. Please go ahead.

Oh, thank you, Karl. So I will tell you that I was an investigator in the PDS clinical trials as well as YOSEMITE and TENAYA. And that patient preference data that Chris presented is 100% absolutely accurate. These patients love their PDS. On Monday, I saw a patient for a non study visit. She has a PDS in 1 eye. She has to come in monthly for injections in her fellow eye. Every single visit she asks me, when can I have a PDS in my fellow eye? So if we were to take patient preference into account, I think that you would see a huge boon, a huge movement toward the PDS. Patients absolutely love it. But comparing the PDS to anti-VEGF designer drugs, that's what I'll call them, anti-VEGF 2.0, the things that are in the pipeline. I think a substantial differentiator is that these designer drugs still have a percentage of patients that might go 4 months or even 6 months. But I can tell you that with the PDS, it is almost 100% of patients will go 6 months. As a treating physician, at the beginning when I start a course of treatment for a patient, I don't know who is who. I don't know who is going to be in that percentage that might be able to be extended out. But as a patient seeing me for the first time, I know there's almost 100% chance that they will do well with the Port Delivery System and not need any rescue or supplemental injections. So I don't want you to forget about that differentiator, and I'm going to close by saying one other thing. Physicians practice medicine. Drug companies, FDA, they do clinical trials, but we practice medicine. And I think physicians will find a way to position PDS with these other agents, including faricimab, that are coming down the pipeline.

Peter Welford would be the next one from Jefferies. Peter?

Karl, you unmuted me. Jo Walton from Credit Suisse. I'm sorry. My 2 questions. Firstly, on PDS. I thought at the conference, they said that 5 of the 6 dislocations have happened before the surgery, and one further one had happened since the surgery protocol had been changed. Well, I was just wondering how many patients have been added after that protocol change? So just how confident you were that you had reduced the level of problem? And whether Dr. Holekamp could say, how urgent you would see it to remove it if you found there had been dislocation? Again, there appeared at the conference to be some doctors who thought that would be very important, and another one who said he had one and actually hadn't bothered shifting it. My second question was just what you might expect to see in the second year of data? We've seen, particularly in DME, you thought that the biology was really strong. And yet if we look at the vision, the vision was no different between the 2. Do you think that there is any chance that when we get out to the end of the second year that Ang-2 component might actually drive a difference in the vision? Because I can see you're very enthused by that biology difference, but the patient presumably only sees the vision.

Maybe I'll take the first part of the PDS question, just to clarify that. So the surgical technique was updated in our instructions for use in June of 2020. And with patients, we've not seen any dislocations in patients who've undergone implantation since the update of that surgical technique. Clearly, there are patients still out there who have had the old technique, and we're looking at those, and that's where that disparity in that number in terms of timing of the dislocation. So I think -- it's also just important to note that of all the dislocations, all those 6 in total, after the dislocation their vision has done fine. These patients do really, really pretty well. Because it's a simple closing procedure, we take the device out. And actually, anecdotally, occasionally, they actually ask for the implant to be put back in. So maybe I'll kind of pass over to Nancy in terms of the second year of the DME data and perspective on that.

This question on vision that came up also making a difference. I mean, yes. [ It's ] asked.