EyePoint, Inc. (EYPT) Earnings Call Transcript & Summary

Ladies and gentlemen, thank you for standing by. Welcome to EyePoint LUGANO Top Line Data Readout. [Operator Instructions] Please be advised that today's conference is being recorded. I would like now to turn the conference over to Dr. Jay Duker, President and Chief Executive Officer of EyePoint. Please go ahead.

Thank you very much. Good morning, everybody. Thanks for giving us your time this morning. I'm pleased to report the Top Line Data from the first of our two Wet AMD pivotal trials for DURAVYU, the vorolanib intravitreal insert. These are our legal disclaimers, and this is the agenda for today's meeting. EyePoint is a leader in sustained drug delivery for retinal disease. We are currently in two Phase III programs in the two largest retinal disease markets, Wet AMD and Diabetic Macular Edema. Beyond today's discussion, we also have multiple potential catalysts over the next 16 months, including LUCIA Phase III data, a potential Wet AMD NDA filing, DME Top Line Data expected in about a year. We have had excellent execution in our four Phase III trials with rapid enrollment and a commercial cGMP facility in place for our U.S. launch. DURAVYU has significant advantages, and we believe it represents a differentiated program. We have broad clinical programs, as I've already mentioned, Wet AMD and DME. We show a very favorable safety profile. Our trial design is clinically relevant. And one potential advantage is this multi-MOA mechanistic edge that RTKI may show in clinical trials. We not only block all the VEGF receptors in PDGF, but we also block the JAK1 receptor, which gives us the ability to block the downstream effects of IL-6. Our Durasert technology has been approved in four prior FDA products. This slide shows you at the top, our Wet AMD program, at the bottom, the DME program. And today's focus is on LUGANO, the first of two Phase III Wet AMD programs. This is the design of both LUGANO and LUCIA trials. In LUGANO, we enrolled 432 patients. They were randomized to two arms 1:1 DURAVYU 2.7 milligrams or on-label Aflibercept control. This is the first study that looked at repeat dosing of DURAVYU. It was given every 6 months. The primary endpoint is 1-year efficacy and safety. The studies are planned to go on for 2 years for safety. The primary endpoint of both studies is non-inferiority mean change in vision from day 1 to average week 52, week 56 versus the Aflibercept control and the non-inferiority margin is minus 4.5 letters. Key secondary endpoints are listed on this slide as well. This is the schematic of the study. A couple of things I'd like to highlight. First of all, all of the eyes in the trial were loaded with monthly Eylea. At week 8, the eyes that were randomized to DURAVYU got a DURAVYU injection about 30 minutes after their third Eylea. DURAVYU was then redosed every 6 months, while the Eylea control was redosed every other month. The reduction in treatment burden calculation is performed after the Eylea load. And note that the week 56 injections don't count. They're in the second year. And therefore, by design, the DURAVYU eyes each will receive two DURAVYU in the first year after the load, and the Eylea eyes should receive five Eylea after the load. And here are the results. First of all, baseline demographics, nothing really to note here. As is true in most Wet AMD trials, it's predominantly a female population. We enrolled both treatment-naive and previously treated eyes in about a 75:25 ratio. And the previously treated eyes were a heavily pretreated group with, on average, over seven injections a year leading into the trial. For those of you who recall the DAVIO 2 previously treated cohort, they came in with more injections. But remember, in DAVIO 2, injections were given right up until the time of randomization. In this trial, none of the eyes could have been treated with an anti-VEGF within 8 weeks of screening. Further baseline characteristics, best corrected visual acuity, DURAVYU eyes started about 11.5 better and the total CNV areas was a little bit larger in DURAVYU. The larger the CNV in general, the worse the eyes will do, but analysis so far do not suggest either of those factors played into the results. I have highlighted a couple of things in the medical history. First of all, about 25% of the DURAVYU eyes gave a history of having Dry AMD going into the study versus about 22% of the Aflibercept arm, but almost twice the number of patients in the DURAVYU arm gave a history of having Glaucoma. This slide shows the top line results. We believe that LUGANO demonstrated clinically meaningful results that reinforce its potential of DURAVYU to improve the treatment paradigm in Wet AMD. When DURAVYU was studied for non-inferiority, by excluding an asymmetric cohort of only nine eyes or 4%, DURAVYU was non-inferior. However, in the full data set, confounded by these nine asymmetric patients who experienced visual loss greater than 15 letters from non-Wet AMD etiologies, DURAVYU was not non-inferior. When looking at the secondary endpoints, 42% reduction in treatment burden through week 56. Remember, the ceiling was 60%, and this was highly statistically superior to standard of care. About 76% of the eyes in DURAVYU were supplement free up to week 32. That was 6 months after the first DURAVYU went in. 54% remained supplement free through the end of the first year, meaning over half of the eyes in the DURAVYU arm were controlled exclusively by DURAVYU. Anatomic control was excellent. The CST difference at week 56 was only 4 microns, and we showed continued favorable safety profile. So let's look at the visual acuities first. What you're seeing on this slide is the graph of the visual acuities, the top gray being the control arm and DURAVYU is in purple. And you'll see when we excluded this asymmetric cohort of nine eyes, DURAVYU was non-inferior to on-label Eylea. So who are these asymmetric cohorts? Remember, these are elderly patients, and they can lose vision for reasons other than Wet Age-related Macular Degeneration. That is to be expected. In the DURAVYU arm, however, nine patients lost greater than 15 letters of vision due to non Wet AMD etiologies. That included six patients who lost significant vision from Geographic Atrophy, Dry AMD, two patients lost vision from Glaucoma and one lost vision following a retinal detachment. Interestingly, all nine showed good anatomic control of their Wet AMD, and six of these nine eyes received a supplement at some point in the study, but none of the supplementation improved their vision. These nine eyes accounted for a total of 258 total losses of letters. Why are we calling this asymmetric? Because in the Aflibercept control group, surprisingly, none of these eyes lost more than 15 letters due to a non-Wet AMD diagnosis. So looking at the graph up right, the dotted line is the DURAVYU eye's visual acuity results when the nine asymmetric eyes are removed. Notice about 5-letter improvement that is maintained throughout the trial. The solid purple line is the visual acuities of the nine patients who lost vision due to non-Wet AMD diagnoses. Notice they never really gained vision, and they steadily lost vision throughout the trial. When looking at their anatomy, what's surprising at first glance is that their OCTs were actually thinner than the rest of the cohort. The cohort overall showed good control of fluid. But when you think about it, Geographic Atrophy is atrophy. And it's no surprise that those nine eyes were actually had thinner maculas than the rest of the cohort. This was not due to active Wet AMD, however. This slide shows the full data set showing that the primary endpoint was not achieved. There was approximately 3.8 letter difference between the two groups. Recall, if the nine eyes were removed, it dropped to a 2.4 letter difference. The bottom graph is the CST on OCT central subfield thickness and notice again, at the end of the trial, there was only a 4-micron difference. So why are we calling this asymmetric? This slide shows you the number of 15 letter losers that occurred in prior studies that involve 2-milligram Eylea. And you can see looking at the right side of this chart, it covers around 4% to 5% and averages about 4%. In LUGANO, at the bottom, it was only 0.5%. So this was highly unusual in an elderly population that there weren't additional patients who lost significant vision due to non-Wet AMD diagnosis. So what did our preliminary evaluation conclude about this? So the first question is, was this asymmetries in adverse events? Is that what resulted in the visual outcome result? And the answer appears to be no. Cataracts were evenly balanced between the two arms. Intraocular pressure increase evenly balanced. We had very little intraocular inflammation or IOI, only one patient in each group and the IOI was not severe. These are investigator-reported percentages of Dry Age-related Macular Degeneration. And you can see while no statistical difference, there was a trend for more Dry AMD overall in the Aflibercept arm. There was one patient in each arm that had a detached retina and that percentage is typical for this type of trial. However, the eye that had the detached retina in the Aflibercept arm gained a letter. The eye in the DURAVYU arm lost 40 letters that appears to be due to the onset of a cataract in an epiretinal membrane unrelated to wet AMD. The second question might be, were these eyes undersupplemented? Did our supplement criteria cause the result? And the answer appears to be no. Supplementation worked as expected. The top line of this graph on this slide represents the visual acuities of DURAVYU patients who received a supplement, without the nine asymmetric patients. You can see at top left, the visit prior to supplement, these patients had about a 3-letter gain in their vision. But at the supplement visit, they were minus 2 letters in their vision, which is about a 5-letter difference, which is exactly what we tried to capture in our supplement criteria. By the second visit post supplement, the eyes were back to where they were a month prior to supplementation. The bottom graph is the patients in the nine cohort who got supplemented. It was only six. And presumably, the other three did not get supplemented because the investigator deemed that the visual loss was not due to a VEGF-mediated disease, not due to Wet AMD, and therefore, supplement wouldn't help. And in fact, the visual acuities that the supplement visit in these eyes were already down 14 letters. You notice what happened post supplement, essentially no change in the visual acuity, confirming that this visual loss was not due to Wet AMD. How about Geographic Atrophy? Well, I already showed you the investigator's assessment of Geographic Atrophy. What you're seeing here is the reading center assessment in a masked fashion. At baseline, the two groups were evenly matched with geographic atrophy, 8% in DURAVYU versus 10% in control. At the end of the trial, no surprise, this happens in Wet AMD eyes. There was progression of geographic atrophy. Numerically, however, it was higher in the control arm and therefore, suggesting that DURAVYU showed no trend of worsening of Geographic Atrophy. So we conclude that neither the supplement criteria or any DURAVYU-induced AEs drove the primary result. How about the secondary endpoints? So let's talk about reduction in treatment burden first. And again, we had a 42% reduction in treatment burden. This averages to about 1.1 supplement injections per year in the DURAVYU eyes. But from a clinical perspective, if this drug is approved and available to clinicians, they might expect two fewer injections per year in their DURAVYU patients, which we believe would be an outstanding result for patients with Wet AMD. The supplement-free rates demonstrate our potency and durability. Up to week 32, that is 6 months after the first DURAVYU, 76% of the patients were supplement free and 96% had received only 1 or 0 supplements with 99% getting 2 or fewer supplements. At the end of the first year, it was 54% supplement free with almost 80% getting either 0 or 1 supplement. At week 56, almost 90% of the Wet AMD eyes that were treated with DURAVYU were stable with 2 or fewer supplements. That is four injections annually or fewer controlled almost 90% of the Wet AMD eyes. This slide shows you the subgroup analysis for the unsupplemented eyes in both DURAVYU and control. This was over half the population, 54%. And you can see in this subgroup, DURAVYU was non-inferior to on-label Aflibercept with only a minus 1.8 letter difference. And again, looking at the anatomic control, it was excellent with only a 3-micron difference at week 56. So this slide also helped to show that our supplement criteria appeared to be adequate. If you believe that some of these unsupplemented eyes were losing control on DURAVYU, then you would have expected the OCTs to get worse as the study progressed. And you can see that they didn't do that. They remain stable. So both stability in vision and stability in anatomy in over half the eyes with our drug alone. So anatomic control, you've seen these graphs already, but just to reiterate, a 4-micron difference at week 56 with the typical sawtooth pattern that we see with every other month Eylea and lack of a sawtooth pattern in DURAVYU. Safety. AE profile, we believe, confirms a favorable safety profile with repeat dosing. We had no cases of insert migration into the anterior chamber, no anterior chamber opacities, no free floating drug particles were observed, no cases of retinal vasculitis, no cases of severe IOI. And in fact, as I already mentioned, there was only one case in each arm of IOI. And in the DURAVYU arm, it was treated for 2 weeks with topical medications and got better. Less than 1% rate of retinal detachment in endophthalmitis in each arm. There was a mismatch in floaters, but all cases of floaters were mild or moderate and didn't require treatment or have any impact on vision. And as we previously reported, there was a low discontinuation rate of about 6% in each arm. This is the entire AE chart for all adverse events that occurred over 2%. As I mentioned, we had about 2.5x the rate of floaters as Aflibercept, but they really didn't cause any visual decrease. None of the patients reported wanting to have the inserts removed, really didn't seem to be an issue with repeat dosing. We will be digging into that a little bit more. But based on the temporal association, we think that some of these floaters were probably visualization of the insert. Neovascular age-related degeneration, you'd expect with more supplements, we would have a higher rate, and that accounts also for the Retinal Hemorrhages, the Retinal Edema and the Subretinal Fluid. Those are all events that one would see with active Wet AMD. Notice again, no cataract intraocular pressure, Dry Age-related Macular Degeneration, no difference. So what are our conclusions? First of all, we believe these results are clinically meaningful and potentially impactful commercially if approved. While we didn't hit the non-inferior margin in the total population, we were non-inferior when we excluded this asymmetric cohort of nine eyes who experienced severe visual loss unrelated to Wet AMD. We believe this study shows favorable safety profile with redosing, 42% statistically superior to standard of care reduction in treatment burden. 80% of the eyes could go up to week 56 with 0 or 1 supplement and the anatomic stability was excellent with only a 4-micron difference, confirming both our potency and our durability. So why do we have confidence? Well, first of all, we showed that half the patients could be supplement free at 1 year with stable vision and OCT, which potentially means 50% of the Wet AMD patients out there could receive just two injections a year of our drug showing stable vision and anatomy. This evidence gives strong potency data and durability data based on the OCTs. This is the first TKI to show a vision gain in Phase III. And in the totality of the results, we believe we have demonstrated the potential to change the treatment paradigm. We also have a lot of confidence in LUCIA. The strength of the ad hoc analysis and the secondary endpoints gives us confidence in hitting the primary endpoint and the secondary endpoints in the upcoming LUCIA trial. We expect to show that data in the fourth quarter of this year. So if positive, we expect to engage the FDA in a discussion of our Wet AMD data package, and we hope to submit that in the first half of 2027, and we look forward to advancement of the DME program with both pivotal trials reading out in Q4 of '27. Thank you very much, and happy to entertain any questions that you may have.

[Operator Instructions] Our first question will come from Tessa Romero with JPMorgan.

So I wanted to ask a regulatory question here. What are the pushes and pulls that we should be thinking about around approvability potential as we think ahead and your expectation to file in the fourth quarter. Isn't it the case that you will need two positive non-inferiority trials to file? And then I have one follow-up.

Yes. We don't believe that, that will be necessary. As I'm sure many of you are aware, there are at least two instances that we're aware of in the ophthalmic division of the FDA, where there were approvals of a drug with one Phase III that was positive and one Phase III that was negative, the Apellis drug and Outlook. And there are a myriad of other studies like this outside of ophthalmology with approval. Also, remember that just a few months ago, the agency updated their guidelines around single trial approval. And what was important with a single trial is supportive results on distinct prespecified secondary endpoints. I think we showed that. Consistency among the subsets, I think we showed that, high-quality conduct with missing minimal data and a mechanistic MOA that directly targets the major driver of the pathophysiology. So I think if LUCIA is positive, we're going to have a strong package to submit to the FDA, and we are confident that if LUCIA is positive, that the FDA will look very favorably on the package.

Okay. And just as a follow-up here, what gives you confidence that you won't have a confounding cohort in LUCIA as well?

Well, I think, again, looking at the data, having eyes in an elderly population that lose vision from Geographic Atrophy, which is unrelated to Wet AMD therapy or other diagnosis is not unusual. Having none of them in the control group is highly unusual. And so again, those type of diagnoses, additional anti-VEGFs don't stop glaucoma. They don't stop the cataract after retinal detachment repair. And so what we would expect is not that these would not occur at all in LUCIA, but they would be balanced. And had they been balanced, then we would have hit the primary. Now I didn't show a slide on this, but we've already done an analysis like that, where if we had taken all the 15 letter losers and just split them evenly between the two arms. And did that randomly, and we did that over 10x randomly and each time we were non-inferior. So we believe that it's likely that this was due to chance alone and that we won't see a similar split in LUCIA. But beyond that, if you just look at the efficacy and you look at things like the control of the anatomy, our drug did a great job controlling it. If you look at the supplement free, 54% were treated with our drug alone with non-inferiority change in visual acuity and excellent control of the anatomy. We have a highly potent drug that is durable. And that if approved, we believe doctors will welcome its use, and we believe this can really help a lot of patients if approved. So we're going into LUCIA with a lot of confidence.

And the next question will come from Faisal Khurshid with Jefferies.

Can you just confirm the outcome on the primary endpoint, including what the bottom range of the confidence interval was for BCVA in the ITT population?

I'm going to ask our Chief Medical Officer, Ramiro Ribeiro, who's here with me to comment on those numbers.

Faisal, so when we look at the full analysis set, including all patients, the difference in BCVA was about 3.8 letters and the confidence interval crossed the 4.5 letter at about 5.7. What is very interesting is that when we remove the nine patients, again, this is only 4% of the patient population, then the difference between the groups is 2.4 letters. So those nine patients represent 1.4 letters. And then we are below the 4.5 confidence interval for the non-inferiority. So again, I think this represents compelling evidence that the effect of those 4% patients driving the results. And then one very interesting analysis that we have done also instead of looking at the mean change, we look at the median change. And again, with that type of analysis, we also observed the impact of these extreme values. And then with the median analysis, we also show non-inferiority against only Aflibercept.

Got it. And then just as a follow-up. So does that mean that like for that 2.4 number that the bottom end of the confidence interval for that was like pretty close to 4.5. And if so, how should we think about that going into LUCIA considering that like past precedent and what AMD has shown variability between identical trials?

Yes. So I wouldn't call it very close. It was around -- I think our limit was around 4. And so I don't think that's very close. And so obviously, if we get that result in LUCIA, that would be favorable.

The next question will come from Sun -- I'm sorry, Yatin Suneja with Guggenheim.

Thank you, Jay, for walking us through in details. I mean, definitely, those nine patients, a little bit unlucky on those nine. I mean remarkable control on OCT. So the question I have for you is maybe just repeat, if you could. When physicians treat, how do they -- like what is more relevant to them? Is it the OCT or the BCVA? And then with regard to the FDA, are you going to have an interaction with them before the next study reads out just to get a sense of how they might be handling these nine patients? If at all, would love to sort of get some clarity on the FDA front, how they will treat this study in terms of those nine patients? And then what about the previous DAVIO studies, right? Could those be used as a supporting evidence even though the patient population was different? If you can maybe just sort of frame the full package for us.

Sure. Let me work my way backwards. Yes, the full package will get submitted and so DAVIO 2 will be confirming evidence, especially vis-a-vis the safety. So the second question you had was about regulatory interaction. It's too early to really have a clear pathway. Obviously, we've been working diligently over the last few days since the study was unmasked to understand the results. And we will, in the future, have a more clear path and discussions with the FDA, but it's too soon to know that. As for how we treat in the real world, I think Tarek Hassan, Tarek, if you're here, can you say hello? Are you on the call?

Yes, I am, absolutely.

Tarek, so you're obviously still a highly regarded, very busy practicing retina specialist. The question was, when a patient comes in for a visit with Wet AMD, what's more relevant to you, anatomic control or visual acuity?

Well, it's obviously both. When we think about the things that we can control within the bounds of a treatment that we may have in our hands, it's ultimately really the anatomy that we have the most impact upon that. So we deliver the drug to do the best we can with the disease at play, in this case, Wet AMD. So we want to dry the eye as much as possible. We want to control the growth of the wet AMD lesion. We hope it translates into a visual improvement, which it does in most cases, when the primary driver of vision loss is that. Of course, we care about vision. It's just that so many other confounding things like we saw in this trial, for example, can affect the vision, and that's out of our hands. We wish we had more drugs and more therapies for these other things. But in short, we care about both, but we have in our hands, something that we can try to fix anatomy, and we hope the vision correlates.

And the next question is going to come from Steve Seedhouse with Cantor.

I had two. The first one is just are you planning to amend or update the statistical analysis plan in LUCIA just to prespecify this type of secondary analysis, excluding these etiology patients just to build some perspective sort of credibility for the case that you'll ultimately make to the FDA?

I'm going to ask Ramiro to answer that question.

Yes. Thanks, Steve. I think it's very fair to take some learnings from the Phase III study and then apply for an upcoming Phase III study readout. So we're typically going to spend the next 2 months looking into the LUGANO data, trying to learn as much as possible so that we can apply any learnings into the LUCIA readout.

Okay. And then, Jay, I think I heard you mention in response to a prior question that the cost margin for the confidence interval was about 4 letters associated with that 2.4 letter difference in that ad hoc analysis. But when you exclude the nine patients, obviously, that are outliers, presumably that reduces the variance quite a lot in that subgroup. So how much did that affect sort of that 4-letter outcome? And when you did that simulation, I think you mentioned splitting up those nine patients across the two arms. In that -- in those 10 simulations, how close does the non-inferiority margin get to the 4.5 letter threshold?

Yes. Good question, Steven. And again, I'm going to ask Ramiro to address them.

Yes. So let me just recap the first. On the full analysis set with all patients, the difference between DURAVYU and Aflibercept was 3.8 letters and the lower bound of the confidence interval was 5.7. Once we remove those nine patients that lost vision not because of wet AMD, then the difference between DURAVYU and Aflibercept was 2.4 letters and the confidence interval, the lower bound was 4 letters. Once we did all those simulations that Jay referred to, where we split in half the patients that lost 15 letters, of course, varies a little bit, but then the confidence interval for all those 10 simulations was less than 4.5 letters as low as 3.7. But of course, on those 10 simulations, it varies a little bit.

And our next question will come from Yigal Nochomovitz with Citigroup.