So, welcome everyone to day one of Citi's Back to School Biopharma Conference here in New York City. Great to be back in New York. I am Yigal Nochomovitz, one of the biotech analysts here at Citi. Our first session, it is my great pleasure to welcome the management of EyePoint, Inc. We have Jay Duker, President and CEO, George Elston, EVP and CFO, and Ramiro Ribeiro, the CMO of the company. Welcome all of you. Thank you very much.
Thanks for having us.
Of course. Obviously, a lot has happened. You have had some phase III data. You have another very important phase III result coming up soon in the fall. Jay, maybe just to level set, if you could just introduce the company, introduce the goals of your lead program. Then I know we were going to have an important discussion about the LUGANO results and then what to expect, of course, for LUCIA in a few weeks.
Sure. Thanks for the introduction and thanks for the invitation. At EyePoint, we are a company that specializes in drug delivery to the back of the eye. Our lead product is called DURAVYU. It is the vorolanib intravitreal insert. Vorolanib is a small molecule tyrosine kinase inhibitor, which has activity against all the VEGF receptors, PDGF, and JAK1, giving it an anti-inflammatory component. It is in our proprietary delivery system that allows it to have therapeutic levels in humans for at least six months with a single injection. As you mentioned, we are currently in two phase III programs, wet AMD and diabetic macular edema, or DME. We have four phase III trials ongoing, and the first trial, the LUGANO trial on wet AMD, read out several weeks ago.
The second trial, the LUCIA trial, started about two months after LUGANO, so we expect to read out about two months after the LUGANO readout. Then we have two fully enrolled DME trials, COMO and CAPRI, which fully enrolled approximately a month ago. Those will have readout in Q4 of 2027.
Okay. Let's start with LUGANO. I think it would be a good idea to summarize the results there, what you observed and, of course, some of the asymmetry that you saw in the results that led to just missing on the primary endpoint. There was certainly some mitigating factors there in terms of the way the randomization played out.
Sure. The trial enrolled 432 patients with wet age-related macular degeneration. Approximately 75% were treatment naive, and 25% were treatment experienced. They were randomized on day one to two groups. They either received 2.7 mg of DURAVYU every six months or EYLEA on-label control. Both arms were treated with monthly EYLEA at the start of the trial. Each then, all the patients were seen monthly, and every other month, the EYLEA arm received another EYLEA injection on schedule. Every six months, the DURAVYU arm received another DURAVYU on schedule. Both arms could qualify for a supplemental injection if they met specific criteria for disease activity. The trial execution was quite good thanks to Ramiro and his team. We had overall about a 6% dropout rate in both arms. None of the dropouts in the DURAVYU arm were due to the drug itself.
As you noted, the top line, which was non-inferiority change in visual acuity between day one and weeks 52 and 56 averaged. The top-line result for all patients was that we were not non-inferior. However, the most glaring unusual aspect of the trial was the behavior of the control group at what I'm referring to as the low-end division. In all previous studies of 2 mg EYLEA on label in phase III, approximately 4%-5% of the EYLEA eyes lost 15 or more letters, either due to wet AMD activity or due to other things. Remember, these are elderly patients, and they can lose vision for things like cataract, glaucoma, retinal detachment, geographic atrophy. So the expectation is 4%-5%. It's varied between 3% to as high as 6%.
If you look at our phase II data, in phase II, 4.3% of the DURAVYU eyes lost 15 or more letters, and about 7%, actually, of the EYLEA eyes lost 15 or more letters. That gave us quite a bit of confidence going into phase III. But when we looked at the results, only 0.5% of the control eyes lost 15 or more letters. Just a very unusual result. Of course, that result really had nothing to do with our drug. Those eyes, obviously, they were control. They did not receive our drug. They received EYLEA on label. If you look at our inclusion and exclusion criteria, they really were not that much different than prior trials. Other than, I hate to say it, but bad luck.
It looks like a series of eyes that lost vision due to other things were randomized into our arm, and none of them got randomized into the EYLEA arm. Specifically, when we looked at, overall, the 15 letter losers in the LUGANO trial, it was about 4%, which is what one would expect. But it was virtually all of them except one was in our arm. Now, some of them were due to wet AMD that still had activity, but we had three retina specialists independently take a look at the eyes, and they all agreed that the primary loss of vision in nine of them was not wet AMD. In fact, those nine eyes, what we call this asymmetric cohort, they actually had good control of their wet AMD at the end of the trial. There was no wet AMD activity.
Based on the CST.
Based on OCT, CST, color photographs, fluorescein angiography, all those tests we do to determine is there active wet AMD or not. One of them lost vision due to retinal detachment. Retinal detachment was successfully repaired, but that patient had not had cataract surgery, and if you repair retinal detachment in an elderly patient using a vitrectomy technique, you get a cataract very rapidly, and that patient did get the cataract.
Easily obtained.
Yeah. It had nothing to do with our drug. And of course, when you look at the rate of retinal detachment, there was one in each arm, which is actually a little bit lower than one would expect.
Six of the patients lost vision due to geographic atrophy. And we've now done some extensive analysis of the geographic atrophy in the trial. And the conclusion is we have no evidence that our drug causes new geographic atrophy or speeds the growth of preexisting geographic atrophy. This is something we will show publicly shortly. But we looked at things like the rate of new GA, and of course, geographic atrophy was allowed in the trial to a certain degree, and at the beginning of the trial, there was 10% of the eyes in the control group had GA, according to the reading center, 8% in the DURAVYU arm, and at the end of the trial, it was 29% versus 24%. So at least statistically-
Sorry. It was just identical.
Yes. It was equivalent.
Agreed. Yeah.
Numerically higher in the control arm. The rate of growth, again, if you look at the rate of growth of GA untreated, your studies will say anywhere from 2 mm - 2.2 mm squared per year. The rate in our arm was less than that.
Okay.
You can also use the other eye as a comparison. For the eyes that started with GA, in both eyes, if you hypothesize that somehow we sped GA up, then it should show up as a difference between the two eyes. There wasn't a difference between the two eyes in progression.
That's interesting. That's data you're going to
Yes. We'll be showing that.
I don't think that specific statement was in the original.
No, we're going to be showing that publicly shortly.
Okay. That's quite compelling. Okay.
Yeah. Finally, the location of the GA overall started closer to the fovea. To summarize, again, not looking at all at our drug, just looking at the performance of the control arm, it would be hard to believe that we're going to see that lower rate of 15 letter losers in the next trial.
Sorry, Jay? So I guess, okay, that makes a lot of sense. Aside from just sort of having bad luck again, which one shouldn't expect, is there anything else that, what else would give us confidence in LUCIA delivering sort of a, quote, "normal result"? It is a little bit of a larger trial.
It is. I'd like to just mention how well we did with the secondary endpoints.
Sure.
Because-
Yeah
When you really think about it, the secondary endpoints are what will make this drug commercially successful if approved. You may recall, and others may recall, we've been saying this for years, when you talk to retina specialists and you ask them, what if you have a new drug that's safe and effective and can be delivered every six months, but results in a one- or two-letter decrease in vision, or even three-letter decrease in vision? The majority of retina specialists say, "Well, I don't care about that because the patients don't notice it, and it's such a small amount." We need to hit, obviously, the primary endpoint to get approval, but when we look at the secondaries, we're really confident that this drug is going to be commercially successful. First of all, safety. We were never worried about safety.
We never had a safety issue in the prior trials, and preclinically, the safety looked excellent, and as expected, there were really no safety issues. Things like cataract, elevated intraocular pressure, no difference between the groups. I mentioned retinal detachment, no difference between the two groups. There were no cases of severe inflammation, and in fact, any intraocular inflammation. There was only one in our arm and one in the control arm. IOI was very low as expected. No insert migration, no anterior chamber opacities. There was a higher rate of floaters, but none of the floaters really were, they were mild to moderate, and they didn't affect the vision.
Patients, obviously, they didn't ask for the inserts to be removed or anything like that. That was really the only mismatch worth noting. From a safety perspective, it's great. What we really did well also is anatomically. Anatomically is the measurement of CST on OCT, and that's what the doctors really look at to see if things are controlled. A normal foveal thickness is, let's say, 300 microns - 325 microns. At the end of the trial, we were 4 microns different than the control arm. Virtually identical to the control arm. We had really good wet AMD control. That's the other thing, that if you looked at the secondary endpoints first, you'd say, "Boy, this drug worked great. Should've hit the primary." Why didn't we? We believe we didn't hit the primary not because we didn't control wet AMD well. We did.
What about other things like supplements? Supplement-free rate up to six months was over 70%, at one year over 50%, and one supplement, almost 80% of the patients at the end of the year had zero one supplements. The other really great statistic here is 88% of the DURAVYU eyes got between two and four injections for the year, which means theoretically, one could control almost 90% of the wet AMD population using four injections or less. If you really ask the question, well, how many eyes get treatment extended to every three months now, even with the new second-generation drugs we have, it's not 88%. It may be a third of that. Again, that really gives us confidence that we can be commercially successful. It does actually amount to approximately two fewer injections per year, hopefully in the real world.
All of those secondary endpoints gives us confidence in LUCIA. We did another ad hoc analysis where we said, "Okay, let's just assume for a second that all the 16 eyes in the LUGANO trial that lost 15 or more letters, let's just assume the randomization didn't work, and let's randomly assign them 8 and 8 to each arm." We did that 10 times, and every time we did that, we were non-inferior to the control, and the difference was between 1.3 and 2.0 letters.
Sort of like re-jackknifing the LUGANO.
Yeah, exactly. If you make the assumption that this just, we got really bad luck, and let's divide the 15-letter losers in half. Remember, in DAVIO 2, there were more 15-letter losers in the control arm. But divide them in half, and we were non-inferior.
And you said the point estimate was in the 1 point-
1.3.
Okay.
To 2.0.
1.3 is essentially what you had at DAVIO-
Little bit more. But well in the non-inferiority range.
Right.
That type of analysis, if one assumes that the control group is going to revert back to what it usually does, and the actual numbers of letters lost is about the same for all these 15-letter losers, then we are quite confident we will be non-inferior. I have to say also, we have taken a look at the 10 letter- 1 5 letter losers. There were six in each arm. So there was not a big mismatch in the milder levels of visual acuity. It was only in the 15-letter losers.
Okay. Then did you want to mention the injection burden reduction too, just the number?
Yeah, it was.
40
A 42% reduction.
Right
in treatment burden, which again, better than expected. That again, is equivalent to about two fewer injections per year.
Right. Okay. As far as the LUCIA data, the expectation is in a few weeks, I gather, or mid-
I would say it. Yeah, still we are saying publicly Q4, but again, if you add two months onto when we posted the LUGANO data, we will be in that ballpark.
Okay.
Obviously, we need to accomplish the database lock and then do the statistical analyses, but we expect it relatively soon.
Okay. As far as the, like you mentioned, the historical standards, what you saw here in terms of this asymmetry was just very, very unexpected.
Unheard of.
Yeah.
Yeah. Between 6 times and 10 times less 15-letter losers than had ever been seen in a 2 mg EYLEA-
Yeah
Phase III.
Since you just mentioned that it was balanced in the 10-15, that really
Yeah
that sort of further illustrates that perhaps there was just
Well, that's all we've got right now.
Yeah.
That's all we can come up with. Certainly the preclinical data doesn't point to any type of toxicity. In rabbits, we dosed about 10 times the dose equivalent that we've had in humans, and we never found the maximally tolerated dose. So it doesn't, again, appear to be related to the inserts or the drugs per se. We had, again, six cases of GA that progressed into or near the fovea that caused some visual loss. But when you look at the overall group, no higher percentage of new GA. The rate of progression of the GA was less than what's been reported on average and essentially the same as the fellow eye.
The LUCIA study is sort of identical in most, if not all, respects to LUGANO, except there's a little bit of an ex-U.S. component.
Correct.
That's the only.
We did end up recruiting more patients, 475 in the end. Dropout rate again, Ramiro, similar, correct? Yeah. So similar dropout at this point. The LUGANO trial was 100% recruited in the U.S. LUCIA was 80% U.S., 20% international.
Okay. Assuming you have a good result on LUCIA, which would be the expectation, how are you thinking about the submission? You had one positive study, one study that there was a lot of mitigants, although it didn't hit the primary. What is the ophthalmology division going to do with that situation?
Ramiro, why don't you take that question?
Yeah. As soon as we would have the results from LUCIA, of course, we would engage with the agency for a discussion about our package. In ophthalmology, there are several precedents of submissions and drugs that got approved with one study being negative and then the other study being positive. Recent examples are Apellis with SYFOVRE and Outlook Therapeutics with a drug for wet AMD, so the same indication. What is important for us is that the secondary endpoints go in the same direction. The safety profile is pretty good. If we combine that with the LUCIA positive results, I think we are very confident that the agency would see this with good eyes, right? Because they don't look at this as a black-and-white thing. They look at the totality of the data. What is the unmet need for the patient?
I think they do appreciate the unmet need of treatment burden, a therapy that brings similar visual outcomes with less injection. I think they're going to see with good eyes our overall package.
Should we be anticipating any sort of combined/pooled work on LUCIA plus LUGANO, or that's just something that you do later? How do you feel about that?
For every single submission that you have two studies, it is very common to pool the two studies together. You do not necessarily need to show positive results on the pooled data, right? You have to show that the trend is similar. We are going to pool for the submission, but we do not think that necessarily needs to be positive.
Okay.
If I could add, we do not know exactly where the numbers of difference in visual acuity would have to be for the pooled data to hit because we do not know the standard deviation obviously at this point. But the ad hoc analysis, which I mentioned, where we randomly assigned the-
The re-randomization.
Yeah. The - 1.3- 2.0. I can tell you if we are in that range, we will hit the combined almost definitely.
I would agree, considering I ran some of these simulations myself.
You get very similar.
Very similar. Okay. As far as LUCIA later in the fall and the filing timelines, can you just reiterate what those are, please?
Yeah. I think, of course, with the results from LUGANO, we're going to have to discuss with the agency the overall package. Of course, now it becomes a little bit more just in terms of detail, the amount of data we have to include in the submission takes a little bit longer, but we're confident that we should be able to submit by first half of next year.
Okay. You mentioned, of course, the very important data in terms of the treatment burden reduction, two to four, essentially quarterly at worst, potentially even less. Just talk a bit about how you're going to position this to the retina community in terms of the ways in which because everyone's going to use it in a different way. There's going to be different strategies. People do different things as we know in the retina world. Just maybe enumerate how you see this playing out commercially in terms of the use cases or the types of patients that could be treated and just go through that.
Sure. Of course, there was some, obviously whatever the label says is going to provide some limitation.
Yeah
In the sense that if we're approved, we would expect to label something like wet AMD previously treated or treatment naive after three injections of an anti-VEGF every six months. I don't think doctors will use it more frequently than every six months because it won't be on the label, and therefore it'd be unlikely the payers would cover it. On the other hand, I think the way that doctors generally approach wet AMD schedule of injections, there's three ways we've done it for years. On label or let's call it on a schedule, which is rare nowadays, and has been from the start frankly. We've individualized therapy from the start. PRN or as needed, that means you see the patient quite frequently, but you only inject when they're active. That's really fallen by the wayside. There's not a lot of people who do that.
And then of course, treat and extend, which means you treat every visit, but you extend the time between the visits till you find what we refer to as the fluid-free interval, and then generally you stick to it. I think those three methods will probably be the way doctors adopt it, at least at the beginning. But it may be given that 88% figure of four injections per year or less, it wouldn't surprise me if doctors say, "Look, I want to take advantage of two MOAs, an extracellular ligand blocker and an intracellular receptor blocker, and just put patients on a schedule." That may be the way it's done.
On the other hand, one could see an easy type of treat and extend where you're using both drugs and extending out the visits and injecting it as you would in any kind of treat and extend type of program till you get out to six months between the DURAVYUs. If the patient's doing well, some of those patients may stay at every six months. Of course, most retina specialists will tell you they like to see the patients more frequently. That certainly at the beginning, I think there won't be a lot of doctors who are going to be comfortable going every six months right out of the bat. I think they're going to want to see how the patient responds and get used to it.
I think at the beginning, the surveys we've done and others have done suggest that, most retina specialists are saying about a third of their patients they would use a sustained-release tyrosine kinase inhibitor on, and that probably will grow as they get confidence in the safety and how it works. The low-hanging fruit, let's say, would be these patients who are being treated every four or five weeks. About 20% of the wet AMD population needs monthly treatment regardless of the drug. So one could see even under those situations where you have a monthly patient 12 times a year and you put them on DURAVYU, and at month three or month four there's a recurrence and you give them another anti-VEGF at that point. Well, all of a sudden you've gone from 12 injections a year to four.
Even though that you might call it a supplement or rescue whatever an additional injection is given or two might be given, that's a big win for everybody. So that whole idea that supplementation means failure, not at all in the real world because what doctors and patients want is confidence in their vision and their anatomic control, but with fewer injections and visits. We believe DURAVYU, if approved, can provide that.
Let's make sure we talk a little bit about DME too.
Yeah.
Tell us about why. DME is obviously a very different disease versus wet AMD. Maybe just discuss how the TKI works in DME, when we expect that data, how those trials are structured. Also, a question that we have gotten quite a lot, which I think is pretty straightforward to answer, is the risk related to DME, given what we saw with LUGANO, but I imagine that is easily answered.
Well, let me start with, first of all, DME is a multifactorial disease, and we have known that for years. If a disease is purely VEGF-mediated, like for example, proliferative diabetic retinopathy, if you give an anti-VEGF and you see the patient back in 48 hours, that disease has stopped. With DME, it can take multiple anti-VEGF injections to get it under control. Of course, corticosteroids work in DME. They do not really work in wet AMD, another indirect evidence that it is multifactorial. So we believe that the JAK1 inhibition that our drug provides will provide additional benefit, especially in DME, where IL-6, which is activated through the JAK1 receptor. The IL-6 pathway has been implicated in DME by quite a bit of evidence, both indirect and direct at this point. So we think that potential advantage is really going to help our DME results.
In fact, if you look at our phase II DME trial, the VERONA trial, at day one, the eyes were randomized, and they either received an EYLEA or they received an EYLEA, and 30 minutes later, they received our drug. At week four, at the first visit, the eyes that received our drug were already about 4 letters- 5 letters better than the EYLEA eyes, and they were about 40 microns - 50 microns drier. So that is our drug. Whether you say, "Well, it was a combination with the EYLEA," that is fine. I do not think doctors or patients will care. But if we can show that to be true in the phase III, I think we are going to be able to capture quite a bit of the DME market, which is currently about a $3 billion market in the U.S.
So our view is that the DME is further de-risked by the LUGANO results, and basically because two things, safety and anatomic control. Things like geographic atrophy obviously does not occur in DME. Sure, eyes can get a detached retina from any type of injection, but other things that can happen in the DME population, remember, cataracts are twice as likely. We had no increase in the rate of cataracts and the safety of the wet AMD trials. So that the safety and the anatomic results give us increased confidence in the DME result, and we think that the market penetration in DME, if we are approved in both indications, is probably going to be higher.
Can you just sort of review the endpoints there for both?
Ramiro, you want to talk about the DME trials?
Yeah. We have two phase III studies fully enrolled for DME, COMO and CAPRI. They are also identical studies, U.S. and international. The design is that on the control arm, patients get the five loading dose, which is only aflibercept, and then every eight weeks. For DURAVYU, they get DURAVYU at day one, because we want to show the benefit of JAK1 inhibition starting day one, and then they get three loading dose, and then DURAVYU every 24 weeks. The primary endpoint is change from baseline in BCVA to week 52 and 56, which is the same time point for wet AMD study. The studies are fully enrolled. We expect the results sometime end of next year.
Just so everyone's clear, this is also a non-inferiority trial or?
Right. So non-inferiority with the same non-inferiority margin of 4.5.
4.5.
That is right, 4.5 letters.
Okay. Is it also going to be, I do not know if you have discussed this yet, but is it going to be a staggered readout, COMO and CAPRI? Are they coming together?
The enrollment was done simultaneously.
Okay
The results will likely come together.
Okay. By the way, going back to what you were saying before about some of these downstream analyses you were doing on LUGANO, should we be expecting to see those before we see the LUCIA results? Or just so we have some sense of timing, do we know?
Yes.
Yes. Okay.
Shortly.
Yeah. Shortly. Okay. Very good. Why don't we spend a little bit of time talking about your manufacturing? Because obviously that's an important piece of the puzzle, a very important piece of the puzzle, and you're in very good shape there. So it would be good to understand how that's going along.
Yeah, we are. George, do you want to talk about the rationale behind us doing our own manufacturing and where we stand with it?
Sure. Just to remind everyone, we actually had our state-of-the-art facility built to our spec several years ago. We've been in the facility for almost two years now. If you look at NDA submission as an example, everyone's focused on clinical data, obviously. CMC is just as important, and I'll remind the audience that most CRLs happen with CMC, and we know that our team knows that. That facility has been in motion, in process of scale-up, to support a potential commercial launch. We have production batches on stability, registration batches on stability, and the team is now gearing up for the commercial production at scale. It's been quite a remarkable process, and just from a competitive and an IP perspective.
While, yes, we do have very good IP around DURAVYU, know-how is just as important, and we were very focused on making sure we control the production of those inserts, and that's all done at our facility in Massachusetts.
Okay.
I will add, obviously the scale is really, really important. If you need to make 10,000 inserts a year for a phase III trial, you need to make 20 or 30 times that for commercial success. In order to do that, our manufacturing operations and our R&D folks have been working really, really hard to essentially automate and semi-automate the entire process. The example I've talked about is in the past we hand inspected every insert. A human picked them up with tweezers under a microscope, weighed them, measured them, visually inspected them. It's all done by robot now. Much faster, much less breakage. The lot size have increased tremendously, and the yields have increased tremendously.
We are quite confident that should we be approved, we, even under the best circumstances, are going to be able to supply DURAVYU with this facility, at least for several years after launch. Although we have got plans again, if launch is successful, we will probably in a few years need to build a second facility.
Okay. Just in the last few minutes, just give us the quick snapshot on the cash, please. Also, we are asking all the companies this now with the world of AI, just to what extent are you using AI tools in your processes, either with regard to FDA submissions or analyzing data or any ways that you are more efficient internally? Just any kind of brief perspective there would be helpful.
Yeah. From a cash perspective, our cash guidance has actually been unchanged for probably a year now.
Which is into Q4 of 2027. When we ended June with $180 million of cash. Obviously, that gets us through all four data readouts. But clearly we will have a number of catalysts between now and then to potentially top off our cash. From an AI perspective, I think we are, like most companies our size, looking at it cautiously. I think we are very protective to ensure that none of our data gets into the public domain to train any of these programs.
Yeah.
But I think from an efficiency perspective, it's being used in certain corners, but not wholesale yet as a company.
Okay. All right, Jay, so any closing comments just overall in terms of how you want to position things and just reiterate the confidence ahead as we look to the LUCIA results?
Yeah. So again, while obviously we were disappointed in missing the primary endpoint, the ad hoc analyses all point to strong activity, safety, and durability of our drug, along with the secondary endpoints. And so again, looking at the control arm and the likelihood that we should revert back to the average rate of 15 letter loss in the control arm, strongly suggests to us that LUCIA is likely to be positive. And if it is positive, we believe we have a clear pathway to approval and then a very clear pathway to commercial success.
All right. Well, that's great. I'm looking forward to seeing those results maybe around Halloween, something like that. All right. Okay. Well, thank you very much.
Thank you.