Okay. Welcome, everyone. Apologies for the few-minute-late start. There have been elevator issues at the hotel. It's a problem I guess you have when so many people attend your conference I have to say. But really grateful to start this next session. I'm Steve Seedhouse with the Cantor Biotech team, and I'm, of course, joined by EyePoint. It's a pleasure to host Chief Executive Officer Jay Duker, Chief Financial Officer George Elston, and Chief Marketing Officer Ramiro Ribeiro for this conversation. Obviously, a really topical moment for the company, and we'll just get right into it. Let's start with the debrief on what you learned from LUGANO and, as you sit here today, the outlook for your next Phase III LUCIA, as a result of the analysis and the discussions that you've been having since.
Sure. First of all, thanks for having us. Yeah, so, as many of you are aware, we talked about our top-line results from LUGANO recently. I'd like to start out with the positives because I'm a positive guy. We did really well on the secondary endpoints. First of all, from a safety perspective, we were never concerned about the safety of the medication, but this was the first study in which we re-injected it, and so the safety looked really good even with re-injection. The second really positive aspect of the secondary endpoints was the anatomic control of the wet AMD. As you are probably aware, ophthalmologists, retina specialists, we look at OCTs to say, “Is there active fluid? Is there not fluid?”
At the primary endpoint of the trial, there was virtually no difference in the fluid in the central subfield thickness between our drug and on-label aflibercept, which was the control. So we had really terrific anatomic control. A reduction in treatment burden over 40%, after the load, which would correspond to approximately two fewer injections a year in the real world, which was excellent. When you looked at supplement-free rates, 76% up to six months, 54% up to a year. So over half the patients were successfully treated with our drug alone. When you look at other things like zero to one supplements at the end of the year, there was almost 80% of the patients in the DURAVYU arm had either zero or only one supplement. Remember, in the real world, if we're approved, a supplement's not a failure.
What doctors and patients want is a reliable way to reduce the treatment and the visit burden, keeping the anatomy and the vision stable. When you take another step and say, how about two supplements a year? Well, two supplements and two DURAVYUs is four injections a year, so every three months. We had almost 90% of the patients who were under control at four injections or less per year, which would, again, in the real world, would be tremendously better than anything we have right now. The elephant in the room, obviously, is we missed the primary endpoint. The primary endpoint, again, is non-inferiority change in visual acuity, measured from day one to an average of week 50 to week 56.
While we had excellent anatomic control and fewer supplements than expected, suggesting that the wet AMD control was really good, there was a real mismatch in the 15 letter losers between the DURAVYU arm and the control arm. What do I mean? Historically, if you look at 2 mg EYLEA, whether it's VIEW 1, VIEW 2, way at the beginning, or any time 2 mg EYLEA was used as a control, one would expect between 4%-5% of eyes losing 15 or more letters. In our DAVIO 2 phase II, 7.7% did. But in the phase III, it was only a half a percent, which is really unprecedented. When you look at the 4%-5% who lose 15 or more letters, some of those are eyes in which EYLEA just doesn't work.
There are eyes that have wet AMD, which even the monthly treatment, you can't control them. That's probably 1% or 2%. The other is eyes that lose vision for other reasons. These are elderly patients, and so they're going to get cataracts, they're going to get detached retina, they're going to get glaucoma, they're going to get geographic atrophy. The amazing part is in our control arm, there were none of those that lost 15 or more letters. That's really the mismatch. We've done some post-hoc analyses looking at that. One of them, what we did is we took all the 15 letter losers, and we said, you know what, let's just randomly divide them in half. What would have happened if they were randomly put into each group?
We did that 10 times, and each time we were non-inferior with the change in vision from -1.3 to -2.0 letters. What we're hopeful, looking ahead at the next trial, LUCIA, is if there is a regression to the mean in the control arm, then we are confident that we would be non-inferior. Moreover, we looked at these nine patients who lost vision in our arm, but primarily didn't lose it from wet AMD. Their wet AMD control was good. Why did they lose vision? One was a retinal detachment. There was only one retinal detachment in our arm. There was one in the control arm. The big difference is they were both successfully repaired with one operation, but the eye on our arm had never had cataract surgery. So that eye got a significant cataract right after the surgery, which is typical.
The eye in the control arm did have cataract surgery and wasn't going to get another cataract. That's the type of mismatch that unfortunately we saw. The other thing we saw was six eyes that lost vision due to geographic atrophy primarily. We've now, as of today, we've published some new findings on geographic atrophy, which I'd like to review.
First of all, when you looked at 15 letter losers in the DAVIO 2 trial, which had about 100 patients treated with our drug, none of them lost 15 or more letters from geographic atrophy. In fact, there was double the rate of 15 letter losers in EYLEA versus us in the phase II. Looking at the phase III, geographic atrophy was allowed in the trial to a certain degree. About 10% of the EYLEA eyes had it at the beginning, and about 8% of our eyes had it at the beginning. At the end of the trial at one year, it was 29% EYLEA versus 24% us. That's not statistically different, but numerically higher in the EYLEA arm.
Suggesting that our drug does not cause new GA. The next question is how about speed the growth of GA? We looked at that as well, and historically, if you look at the studies on geographic atrophy, it's typically about 2 - 2.2 millimeters squared per year. Our rate was 1.7 millimeters squared per year, so under what is typical. The other way to really do it is to look at the fellow eye because it's almost like a self-control. If you've got a patient that has geographic atrophy in both eyes, they tend to progress at the same rate, and that's what happened in our trial. It was 1.7 in the DURAVYU eyes, 1.6 in the fellow eye. So there's no evidence from that analysis that our drug sped the growth of preexisting geographic atrophy.
What we did find was that the geographic atrophy started closer to the fovea, approximately 800 microns to the fovea at the start versus 1,000 microns in the control arm. It progressed at the same rate, but it started closer. We think that that's the primary reason for the mismatch.
I see.
We've gone back, and again, looked at a lot of different aspects of the safety and the efficacy in the data from LUGANO. The best we can come up for is that the control arm really, with respect to the 15 letter losers, did something that was historically unheard of. By the way, we've looked at the 10 - 14 letter losers to see if that was a mismatch. That's kind of the next level up. There wasn't. Those were evenly distributed.
Okay. The good news here, because I like good news as well is you have an opportunity to test these hypotheses, right in your next phase III.
We do.
That is immediately forthcoming. But the critical question is what differences, if any, are there in that study? What are you able to look at on a blinded basis in that study to get? Could you go back and look at on a blinded basis GA lesions and how close they are to the fovea and sort of suspect whether there is any likelihood that you would have a similar bad luck stroke in LUCIA? What can you say ?
Yeah. First of all, when you look at masked data, you can make certain assumptions about it. Your assumptions might be wrong. Let's go back to LUGANO. Masked data for 15 letter losers, there were 4% of the eyes masked in totality that had 15 letter losers. Historically, EYLEA is 4%-5%.
Got you.
You look at the masked data and we said, great, it's going to be 4% or 5% EYLEA and probably 4% or 5% in us, and it's even. But it wasn't. You can look at masked data like that, but it doesn't necessarily going to play out the way it might historically. Now, we believe there'll be a regression to the mean, and we believe that it's more likely than not that the EYLEA arm will perform as it usually does historically. But things like geographic atrophy, that's read by the reading center, and the reading center looks at it at week 56. You need that last visit which has occurred, but then the reading center needs some time to actually do the readings and then statistically needs to analyze it. You really couldn't even in a masked fashion get any knowledge about that at this point.
I see.
We did or will very soon make some changes in the statistical analysis package from some learnings.
Okay.
Ramiro, you want to comment on those changes?
Yeah. We want to bring the learnings from LUGANO into the LUCIA readout. Again, the studies are identical. The difference that LUCIA has a little bit more patients, 475 versus 432, so give that a little bit more power to the study. But the analysis that we did on LUGANO on a post-hoc fashion, looking at patients that lost 15 letters for reasons not related to wet AMD, we're going to be doing that on a pre-specified manner with LUCIA. So as soon as we have all the data, imaging, clinical data from the reading center, we're going to be assessing if the patient lost vision because of wet AMD or not, and then have that as a sensitivity analysis for LUCIA. The primary endpoint doesn't change. It's still the same. We're going to use all data to inform policy, but we're going to have this supportive analysis.
Okay. Terrific. Focusing on the exclusion of that imbalanced cohort of patients and the ad hoc analysis where you had the 2.4 letter differential in the DURAVYU arm, it was non-inferior statistically at a 42% treatment burden reduction as you mentioned. Before you flip the data card, is that a data profile that based on your market research and based on your expectations that you're comfortable would be competitive in the marketplace?
Oh, yeah.
Yeah.
When you talk to retina specialists, and again, I think some of you really remember even back before with the phase II data where we would talk to KOLs and investors would talk to KOLs, and they will tell you one letter or two letters or even up to three letters difference is not noticeable by a patient, and the doctors really do not care. We need to be non-inferior in LUCIA for approval, obviously. But once we are approved, it is the secondary endpoints that cause your ability to be commercially successful. Our secondary endpoints were terrific.
I mean, if the doctors look at the secondary endpoints and make the assumption that we are approved, I think they are going to be very enthusiastic about the use of our drug. One thing that we have kind of been really pushing on really since the beginning of our program is the concept that in the real world, a supplement injection is not a failure. You cannot think of it as our drug needs to control 100% of the population with just our drug. That is not true at all. The example we give over and over is if I have got a patient I am treating monthly, let us say with EYLEA, and I cannot get them even to six weeks, and I have got DURAVYU now approved for every six months and I use DURAVYU, and over the next year I give them two DURAVYUs, but I also give them two EYLEAs.
Two supplements. It is great. Gone from 12 injections a year to four injections a year. Everybody is happy. We need to get over obviously the big hurdle, which is approval. But once we do that, hopefully with LUCIA, then we believe we have a very clear pathway for commercial success with our secondary endpoints.
Yep. One thing that is and I agree, the 42% treatment burden reduction I think maybe exceeded the bar that The Street at least, and then probably correspondingly investigators and everyone were expecting would be meaningful. One of the features of the phase III study, as you know, of course, is the criteria for rescue was a little bit more stringent.
in phase III, you didn't have the sort of physician's discretion. I've just wondered if that feature of the study allowed for less rescues, had they been sort of more liberal with rescues, maybe you retain a couple points of letters of vision. You hit the primary endpoint, but maybe the treatment burden reduction isn't 42%, maybe it's 35% or 32%. What do you make of that?
It wouldn't have been a couple of points. I mean, can we imagine that there was a handful of patients out there who, if they had received an earlier rescue, might have retained or improved a little more vision? Sure. Theoretically that's true. What goes against that argument are two things. First of all, look at the OCT control. If we were allowing too much active wet AMD, we wouldn't have gotten that tight of OCT control.
It's true.
The second thing to look at is the eyes that were supplement-free, and we've shown those graphs. The supplement-free eyes 54%, they had a 1.4 letter difference to the supplement-free EYLEA eyes. They were obviously that was non-inferior. But again, the OCT control was excellent. If you think we were letting some eyes slip. In the supplement-free, that would've been the canary in the coal mine because you would've seen the fluid increase, and you would've seen a bigger delta in the vision. All of that points to we had really good control of wet AMD in this study.
Yeah, I think especially in the second half of the study, after the second DURAVYU dose, it's pretty definitive evidence, even if you isolate those non-rescued patients of the drug being active. You're what? You're six months, nine months, 10 months out from the EYLEA induction, and you still have that good CST benefit. I'm just like why does the BCVA sort of slowly separate in the second half of the study despite good control of the 9 patients?
The 9 patients. Yeah. We have this slide in our deck where we show the visual acuity of the non-nine patients, and it's pretty flat.
Okay.
When you look at the visual acuity of the nine, those outliers who lost vision not due to wet AMD, first of all, they never gained vision during load, which is unusual. Even though their OCTs got dry. They had wet AMD. They dried. But the vision has never improved. They slowly decreased, and then they started to actually rapidly decrease towards the end. That's the difference. Those nine eyes, which was only 4% of the population, resulted in a significant difference in vision.
Okay. How satisfied are you with the safety and adverse event profile that you observed in LUGANO? Talking a lot about efficacy. I think the only events that were higher in the DURAVYU arm were floaters. Maybe you can talk about that. Neovascular AMD I think was numerically higher. Retinal hemorrhage. You have addressed this on your call, but maybe anything you have learned since. Is there any reason to suspect that any of those are sort of drug product or procedure related?
Yes and no. Let me talk about wet AMD hemorrhage, retinal edema, subretinal fluid. Those are all pieces of evidence of active wet AMD. We had more eyes with active wet AMD than did EYLEA by definition because we had more supplements. When an eye gets supplemented, it means it has active wet AMD. What does that mean? It may have a retinal hemorrhage, it may have fluid. Those are all tied up in the higher rate of neovascular AMD. On the other hand, the floaters, yeah, we were about 2.5 times control. They were mild to moderate, according to the patients. None of them caused decreased vision, none of them resulted in the patient wanting the inserts out or anything like that. Based on the timing, we suspect that many of them was the patient seeing the insert.
As you know, they are heavier than water, and so when they go in initially, as they are injected, sometimes the patient probably can see them, and then they sink down to the bottom of the vitreous cavity, and they pretty much stay there. There might be occasional patients who catch the edge of the insert. For example, we have had an FDA-approved insert for years with a steroid in it, the OZURDEX insert. Doctors see them occasionally. Patients see them occasionally. I have used them extensively. Never had a patient tell me they would not have another one. Yes, I think that mismatch was in floaters and pretty much everything else, it was really less than 5%, and there was no big difference. The important things, intraocular inflammation.
We had one case- treated with a mild steroid topically. We did not expect it. There is no reason our matrix should cause it. There is no reason vorolanib should cause it. But it is good to know that we did not. I want to remind everybody, these AE tables, that is from a re-injection also. We are the first TKI to show with reinjection AE tables. Yeah, so we are very happy with the safety, although I have said it already, pre-clinically, we have never really seen a safety issue, never found a maximally tolerated dose of vorolanib. In the prior studies, we did not really have any issues either, so no surprise.
What is the current guidance you are providing on the timing of the LUCIA data readout?
The guidance is Q4.
Okay.
Again, the LUCIA started approximately two months after LUGANO, so we are expecting the top line approximately two months.
Okay.
Which puts us no earlier, it is not going to be the first or second week of October. It could be sometime later in October or perhaps even early November.
Okay, great. Looking ahead, let's assume LUCIA is positive and maybe replicates the ad hoc analysis, excluding those nine patients, replicates that data profile we talked about. What will be your approach with FDA? What's the case you'll make for approval here on the, call it mixed data readout?
Well, I think, f irst of all, there are precedents in ophthalmology for approval with a one trial positive, one trial negative outlook at Apellis. There's precedent. The second thing is that we have, if we hit LUCIA with primary endpoint, we have a really good explanation as to why LUGANO didn't hit. All of the sub-analysis that we've done have pointed to the over-performance at the bottom, at the 15 letter losers of the control arm. If that doesn't happen, and that control arm reverts back to what's expected, I think there's a very easy explanation for the top line miss. We believe that the FDA recognizes the importance of patients to have sustained anti-VEGF activity for long-term, and the advantage to patients to be able to have fewer injections. Therefore, we're confident that the agency would, with a positive LUCIA readout, look favorably at our application.
How does, I would imagine, and correct me if you think otherwise, but with the DME studies also ongoing here, they're kind of lost in the discussion here, and maybe shouldn't be because would we get data from those studies before there would be any regulatory decision from FDA anyway on wet AMD, no matter how quickly you could put together
Yeah, I think that's likely.
Okay, how should we think about that? Are you doing anything in those studies amending the SAP or otherwise as a result of some of these quirks that you now see?
We're not amending the SAP. We thought a lot about that, but there really is nothing that would make sense at this point. DME has been lost in the shuffle for a long time, for a lot of people. DME is a $3 billion business. We are the only TKI in DME. We have two fully enrolled phase III trials, and we think that the results of LUGANO de-risked DME even more. Why? First of all, the safety. Diabetics have a higher risk of cataracts. They have higher risk of inflammation. They have higher risk of glaucoma. We had none of those problems. I think the safety is a read-through to DME. I think the anatomy is a read-through to DME. Both are exudative diseases with fluid in the retina, and we were able to control that fluid really well.
Remember, the VERONA trial, our phase II DME trial, showed that at month one, week four, the first visit after the drugs went in, the eyes that got DURAVYU were already seeing 4 - 5 letters better than the EYLEA eyes, and they were already 40 to 50 microns drier. We have set up the DME trials with a secondary endpoint to look at week four to see if that is replicated. If we are non-inferior in DME, but we can show that we get there faster and get dryness faster, I think we are going to be used in the majority of DME patients. If you are a patient, why would not you want to see better faster with fewer injections? Everybody would.
Yeah.
Remember, that is a population that has a hard time coming in for the necessary injections. While the actual market is smaller than wet AMD, we think potentially the penetration of the market for us would be even larger. It is a real opportunity.
Interesting. Just back to wet AMD, given the result of LUGANO, I know you have had conviction in your alignment on the 4.5 letter non-inferiority margin for your phase III program before then, but what would you say to address the question, is the FDA likely to maybe make the result more stringent for LUCIA because of the ambiguous result in LUGANO? Is there a higher hurdle in LUCIA statistically .
We have no indication that they would do that, no. Again, we have every indication that a positive LUCIA trial will be sufficient for us to file the NDA, and the confirmatory data will come from LUGANO. Again, remember, it's totality of data. It's benefit to the patient and the totality of data. So that's all the secondaries that we've talked about that were very positive. I think the agency is going to recognize that.
Is notwithstanding wonderful conferences like the Cantor Global Healthcare Conference, is the next time we'll hear from you the LUCIA results? Are you planning, I know you, for example, indicated this GA analysis today. Is there any sort of further analyses or presentations you would add?
Yes. On our website, as of this morning, you can find those slides for the GA analysis. Among other, we have a nice fibrosis analysis that showed that it was higher in the control arm, too. At Retina Society we have two presentations coming up later this month. We have presentations at the American Academy of Ophthalmology in October, EURETINA in between in October, and then we are planning on publicly showing the LUCIA data for the first time at the FLORetina meeting, which is the first week of December.
Terrific. Okay, well, we will look forward to all those updates and appreciate obviously you being here and walking through the discussion.
Our pleasure.
Super helpful, and thanks to everyone in the audience.
Thanks for everyone's patience for being a few minutes late. Thanks, Steve.