All right. Good afternoon, everyone, and thanks for joining us for the EyePoint session. As you know, EyePoint is an implant technology developing TKIs for DME and hopefully maybe one day wet AMD. With us today are Jay Duker, the CEO of EyePoint, and George Elston, CFO. Thank you for joining us, and maybe you can sort of just level set us and tell us where EyePoint stands today after—
Sure.
—an unfortunate outcome for LUGANO and LUCIA.
Yeah. Thank you very much, Annabel, for having us, and thanks for the opportunity. So, quite simply, where we stand today is we're a DME company. Diabetic macular edema is currently a $5 billion market worldwide, with $3 billion in the U.S. We are currently in two identical simultaneous phase III trials in DME, COMO and CAPRI, that are both fully enrolled, and we should see the top line data in just under a year, at the beginning of Q4 of next year. We have a lot of reasons to be optimistic about the DME results, despite, as you know, our primary endpoint miss in wet AMD, and happy to expound and go into detail about those this afternoon.
Mm-hmm. Okay, sure. Why don't we go Well, before we get into that, maybe you can speak to how we should think about TKIs for retinal conditions, and is it something Does this tell us more about TKIs in general, or is it something specific to DURAVYU? Is it something specific to clinical trial design? Maybe you can expound on that a little bit.
That's a great question, and I'm not sure we can answer the question now or try to extrapolate about TKIs in general. What we know about our TKI is that it's quite active pre-clinically. Based on the PK data that we have in the phase III trials, we know that the inserts were releasing at levels that we can extrapolate were many times above IC50 in the retina and choroid. What we also know by the PK data systemically is, of course, we redosed at month eight during the wet AMD trials, and we actually saw higher levels of the drug, vorolanib, after the redosing. Yet it's hard to look at the clinical data and see that there was any difference in outcome after the second injection. If someone suggested that perhaps the higher dose might have worked, we don't see any evidence of that.
TKIs in preclinical models work quite well. In fact, there's a model that we ran of the chick chorioallantoic membrane model where our TKIs and others at IC50 worked better than a ligand blocker did in this model of neovascularization. The science is there. The drug was releasing. The drug was there in doses that should be effective. Again, I think that if you look at the primary endpoint, we believe that the evidence is clear that the drug was active. In both trials, we showed a gain of over two letters, and that shows activity.
Certainly, if you look at the subgroup of patients who were supplement-free after the load, in both trials, over 50% of the eyes went the rest of the trial up to week 56 with no additional treatment, just our drug. In the pooled analysis, those eyes were non-inferior to the unsupplemented EYLEA eyes. We know, and at least you could argue, if the lack of supplement shows the activity in those eyes, those eyes did great.
The question is that a large subgroup of patients that we can identify a priori, and therefore, if we were to run or someone was to run a third trial in wet AMD, would we be able, through inclusion and exclusion criteria, to identify that group of patients and have a more positive result? The other thing from both trials to note is the secondary endpoints.
We did quite well in the secondary endpoints. In particular, the one that really matters the most besides visual acuity is safety. We dosed the drug in over 440 patients for one year with redosing, and the safety looks really quite strong. When you look at the AE tables, the mismatch, none of the safety concerns were more than 8%, and most were around 2% or 3%. Any of the mismatch you saw was a mismatch for active wet AMD, meaning AEs that involved retinal hemorrhage or fluid were slightly greater in our arms than in the control arm because we had more eyes that had active wet AMD by definition, because we had about 30% more supplements in our arm than in the control arm. So, the safety really looks fine.
The other secondary endpoint, which we think is de-risking for DME, is the OCT data, which is a measurement of fluid in the retina. The CSTs in the pooled data from the two trials in the DURAVYU arms showed a difference of only 10 microns. That's tiny. That's a standard deviation. A normal CST is up to about 325 microns. So a 10-micron difference is really not clinically important at all. So between the safety and the control of the fluid, we think both of those are really important to de-risk the DME result.
Also, both trials showed the exact same reduction in treatment burden. On average, if approved in the real world, if this would hold, we would have on average of about two injections less than on-label EYLEA for wet AMD. Once again, showing that the drug was active and potent, just wasn't able to hold the vision quite enough to be non-inferior.
Right. Does that also maybe speak to on-label EYLEA being more potent than what we understand it to be real world?
I don't know if we can say that per se. On-label EYLEA performed in our study the way one would expect, gained slightly under six letters. That, I think that that's consistent with other trials. There's always a few exceptions. Trials that enroll nowadays where the patients who have wet AMD are usually now being seen at an earlier point in their disease and enrolling in trials earlier, this is what one would expect. Of course, in the real world, on average, patients get fewer than every other month EYLEA, not necessarily because that's what the doctor wants to give.
There's real-world issues around patients coming in. That's again why sustained release makes a lot of sense. Because if there are missed visits or patients just can't c ome in for visits, or the treatment burden after years becomes too much, if you have an insert that releases for six months or longer with a single injection, that obviously can take care of those patients. Of course, 40% of wet AMD patients drop out from treatment after a year and a half, which is an amazingly high number. Many of those probably do it because they just can't keep up with the visits. Again, a sustained release offering would really make a difference to those number of patients who drop out. Back to the question about how the EYLEA arm performed, I think it performed as expected, as one in a non-inferiority trial.
Okay. I guess a couple more questions on wet AMD, and then we're going to switch over to DME. I guess, how important was patient selection for these trials? I know that your phase III trials had a sort of an all-comers, much more broad than what you had in phase II, and maybe a little bit more selective than some other TKI trials or maybe some other even anti-VEGF trials. How do you think about it now?
I would say other than the inclusion of 25% previously treated patients, our inclusion and exclusion criteria was very similar to the previous five or so wet AMD studies that have been done. We really didn't change that much. There was a difference though, again, between the phase II and the phase III trial. In the phase II trial, we enrolled all patients who were previously treated, and in that trial, they had to receive an anti-VEGF two weeks to five weeks prior to screening. The reason we designed it that way was we didn't want to enroll patients who were non-responders.
This was an opportunity to test over a short period of time if they were actually responding. We had an OCT cutoff for fluid of 350 microns, but we also enrolled patients who saw up to 85 letters, which is 20/20. Now, in a phase III wet AMD trial, the FDA is pretty adamant that you need to enroll patients who have active wet AMD.
Makes sense. What has that defined? The agency defines it as 78 letters or better, so you can't enroll patients with 20/20 vision. They have to have fluid on OCT that is due to wet AMD as adjudicated by a reading center. As a result, we enrolled patients in the phase III by regulatory mandate, all of whom were active. In the phase II, if you define, use that strict definition, they're probably over 50% of the patients, because of that injection that occurred as a run-in at the time of screening, were not active. I think that's the big difference. It really wasn't inclusion/exclusion criteria per se. Again, if you look at the results, fortunately or unfortunately, however you want to look at it, the naive patients performed the same way as the previously treated.
In other words, in the analysis of each, neither of those groups were non-inferior. In the second trial, the LUCIA trial, the previously treated eyes were numerically close to the EYLEA eyes, the DURAVYU eyes were, but they were still not non-inferior due to the relatively low N in that subgroup.
Mm-hmm. Okay. I see. Last question on wet AMD. What are the sub-analyses that you're going to do to see if there is any specific group that responds better?
We're involved in that right now. We haven't yet been confident enough to say we found the smoking gun, meaning, I mentioned the patients who were supplement-free, and it would be great if there was a delineating factor or some type of biomarker where we could say, "This is what sets them apart," and know that ahead of time. Obviously, if we were to do that and run the same trial using just those supplement-free patients, we would have had a great result. We're looking for that, and we're looking at a lot of different potential biomarkers and demographic factors to see if we can find something that suggests who does well with the drug. The other side is, if there's a subgroup of patients who clearly do appear to be non-responders, then obviously, if we were to go forward with another trial, we wouldn't want to enroll them.
So that activity is ongoing, and there are several layers to this analysis. For example, one can look at things like cutoff of fluid, the patients who entered with over 400 microns or under 400 microns. That is one way of doing it. Of course, it is better to do it on a sliding scale and see if there is a tipping point. In other words, look at the various levels of OCT at the beginning and see if there is a certain cutoff of fluid where all of a sudden we get more non-responders. Those type analyses take a little bit longer, but that is what we are doing right now.
Okay. Got it. All right, I want to move over to DME because that is obviously the focus right now. Maybe you can talk about the pathophysiology of DME and how it is different than wet AMD, and what gives you some hope that this is active in DME.
Sure. We have quite a bit of hope, actually. First of all, we can start with pathogenesis. The essence of wet AMD is choroidal neovascularization. It is new blood vessel growth from the choroidal layer under the retina. DME is a disease of retinal blood vessels, and the normal retinal blood vessels become impaired and leak. VEGF plays a role in both, but wet AMD appears to be a primarily VEGF-mediated disease, where DME is VEGF plus inflammatory factors. We have actually known that for a while. Intravitreal corticosteroids work really well in DME. In fact, if you do a study on patients who are pseudophakic, meaning they have had their cataract out already, corticosteroids work just as well as anti-VEGF ligand blockers do. We do not tend to use it as much in the real world because of the potential side effects, which is cataract and glaucoma.
We also have just recently seen data from EyeBio Merck that a Wnt agonist, which has no direct anti-VEGF activity, can be non-inferior to an on-label anti-VEGF. I think that is more evidence that VEGF alone is not the whole story in DME. Of course, the other pieces of information, I already mentioned the safety that we saw in the wet AMD population, the fluid control, and the reduction in treatment burden, all of which I think have some read-through to DME. Other pieces of evidence, though, really come from our phase II VERONA trial. The VERONA trial enrolled active DME patients. It was actually the first trial we ran that required all the patients to come in with activity, meaning decreased vision and fluid due to diabetic macular edema. There was a washout period of eight weeks, so they could not have had any anti-VEGF for eight weeks.
It was almost the opposite of what we did in the phase II wet AMD trial. The reason for that is DME is a different disease. It actually has a beginning and a middle and an end. Most patients with DME do not need lifetime therapy. If one were to design a trial and enrolled patients who had no fluid and good vision in DME, the argument could be, well, they didn't need treatment at all. We couldn't make that argument in VERONA. They clearly all needed treatment. They were all active. As early as four weeks, the DURAVYU-treated eyes had better vision and drier retinas than the EYLEA-treated eyes. In the end of the trial, at 24 weeks, the high dose, the 2.7 mg dose, which is our go-forward dose in DURAVYU, had the same visual acuity as the EYLEA arm did.
But it had fewer rescues and also had significantly less fluid on OCT. The other things we were able to show in that trial is in a dose-dependent fashion, because we also tested a dose of 1.3 mg of DURAVYU, and in a dose-dependent fashion, there was less leakage on fluorescein angiography in the DURAVYU eyes than in the EYLEA eyes. So that suggests that it potential for an additional anti-inflammatory effect from the IL-6 pathway.
We know that our drug has activity against the JAK1 receptor, and that is one of the primary activators of the IL-6 pathway. IL-6 has been implicated now as one of the major inflammatory proteins in DME in many studies, including vitrectomy studies, where eyes that were undergoing vitrectomy that had DME had significantly higher levels of IL-6 than eyes that were diabetic but didn't have DME. There are other companies out there pursuing IL-6 plus anti-VEGF as a pathway forward to improve DME therapy, and there's some clinical evidence for that as well. So that fluorescein angiography, while indirect evidence, is suggestive that it's not just a VEGF effect that we were seeing with our drug. In addition, in both doses of DURAVYU in the VERONA trial, 50% of the eyes at the end of the trial had no DME.
All of the EYLEA eyes still had some evidence of DME. Finally, we had two patients that we've shown the case reports, and they're on decks, of eyes that had been on VABYSMO prior to enrolling in the trial, entered the trial with fluid and with our 2.7 mg dose and one shot of EYLEA at the beginning of the trial, which they all got, and no other treatment. They actually, at the end of 24 weeks, had improved vision and less fluid than they had had prior on the VABYSMO. So that's the kind of finding that really gives us considerable optimism that we will have a different result in DME than we did in wet AMD.
Just to clarify, in DME, the dose response, you believe that the dose response that you saw in DME was more related to the IL-6 targeting versus the, I guess, the less of a dose response that you saw in wet AMD—
Yeah.
—which has no inflammatory component.
Yes, except it wasn't a dose response in wet AMD because we only tested one dose.
Oh, okay.
2.7 mg in the phase III. We did test two doses.
But in two doses in the phase II, you had.
We did.
Yeah.
There was no significant—
Right.
—difference. Whereas in the DME trial t here was a difference between the two doses, and that's one of the reasons we'd actually already selected the 2.7 mg dose. But it gave us relief that we had picked the right dose, the higher dose, because of the DME.
So you think it's the actual targeting, the JAK targeting, and not really the.
We can't say for sure, and I don't want to go out on a limb and claim that to be true, but that is consistent with the other findings.
Okay.
And of course, we've done some preclinical work too. There's models of leakage, and in those models of leakage, if you give the They're created by adding VEGF and IL-6 to cells, and if you give an anti-VEGF in that model, you get a reduction in the leakage. If you then add the anti-VEGF and IL-6 blocker, you get more reduction. When we added vorolanib at the doses that we're getting in humans, we got the same reduction in leakage that you saw with the combination of IL-6 and anti-VEGF, and we presented that data as well. So again, it's preclinical but it is, we think, compelling to show that we can accomplish in that model the same reduction in leakage that an anti-VEGF, anti-IL-6 did.
I guess that goes to my question. How do you know if that IL-6 targeting is strong enough? Can't know. I mean, you were known as a pan-VEGF product or-
You don't. You don't.
Right.
But in the end, retinal specialists who would be able to offer their patients a non-inferiority product that lasted six months or longer and had a safety profile, I'm not sure that they're going to care how much of that effect is anti-VEGF versus anti-IL-6.
Right. Just on the population for DME, are they as variable as they are for wet AMD? I mean, how should we think about the variability of the patient population for DME?
I'm not sure that it's easy to compare variability because you have to define what variability means. One of the things that's true in the DME population is it can take many more injections of an anti-VEGF before you achieve dryness and improved vision. That's when you look at the labels, for example, for two milligram EYLEA, it's a five-injection load, five injections monthly, where in wet AMD, it's three injections. There is variability of response, and we've attributed that to the fact that there's this inflammatory component that you may not initially, with the VEGF being, you're not addressing necessarily.
Okay. I'm going to say something controversial now, and you can tell me whether this assumption is correct or not correct. Speaking to KOLs, they've commented that if something doesn't work in wet AMD, it's not going to work in DME or DR. Is that too broad an assumption?
Corticosteroids.
Is that what they are going by?
No, that is the answer. Corticosteroids work very well in DME and have virtually no effect.
No effect.
in wet AMD.
That's too broad a statement—
It's too broad a statement.
—on their part.
I think while there are certainly associations, because most of what we're using for those diseases are ligand-blocking biologics. That's not what DURAVYU is. While there is certainly a connection through VEGF, we remain optimistic that we're going to see again, if we have the same type of secondary endpoints but we're non-inferior, it's going to be a grand slam.
The non-inferiority margin was the issue that we faced. It wasn't the safety, and it wasn't the reduction in treatment burden. It wasn't the fluid control. I don't think it's a controversial take, but I don't think that it's backed up by the reality of corticosteroids, which clearly work in DME and don't work in wet AMD.
Got it. I guess another question I would have is, do you need an anti-VEGF for that drying capacity?
Well, we don't know.
Initially.
We actually don't know because we've never tested our drug without at least one dose of EYLEA.
Now, remember, in wet AMD phase II and phase III, we did a full load in everybody, three injections of EYLEA at the beginning. In the VERONA trial, we only did one EYLEA. We didn't do the full load, yet we had really good results, and in fact, in that study, better than the EYLEA arm did. I don't think that we can say that you need to have an anti-VEGF alongside it. But I will also say, we talk about supplement injections and the way these studies are designed. I've said this over and over. In the real world, supplementation is not a failure. Doctors don't mind injecting. They just want to get better results in the long term.
I think the patients mind the injections.
Yeah, we all want to get better results in the long term, and we all want to be able to control the intervals better. The idea of using two MOAs for a chronic disease is something we do all the time in medicine. You could argue we're starting to do it with VABYSMO in retinal disease. I think if in the real world, if DURAVYU were to be approved for either or both those indications of wet AMD and DME, I think doctors would often choose to use it with a ligand blocker because the possibility that there's synergy, and again, you can certainly reduce the treatment burden, even if you're giving an occasional ligand blocker along within every six months DURAVYU.
Okay. Have you talked to the FDA since the LUGANO and LUCIA trial and thought about any changes that you need to make to the DME trial to COMO and CAPRI?
We've had no formal discussions—
Okay.
—with the FDA. We have not discussed with the FDA any changes, significant changes in the protocol.
Significant changes?
Yes.
As opposed to some changes?
Yes.
Okay.
Well, some changes.
There's some changes.
We don't have to discuss. So actually, what we've done is we've made the rescue criteria in the DME trials, five letters and 50 microns of fluid off best in study.
And what was it before?
Ten letters and 50 microns.
That's a change that we've made. That, of course, the rescue criteria is up to the company to determine. It shouldn't affect many patients because the first time a patient could be rescued in the trial would be week 20, I believe.
Because we didn't allow rescues before the full load was done. The change shouldn't really affect many patients because there were very few rescues that had been given before that change was made.
Okay. All right. I'm going to put George in the spotlight now. Obviously, there have been some strategic changes made at the company. Can you talk about those and what kind of runway that gives you right now to complete all of this?
Yeah, sure. On the heels of LUCIA, we made the strategic decision to about a 53% reduction in the size of the organization. We ended September with about $118 million of cash, and we're actively working through scenarios to make sure that our cash gets us to and hopefully through the DME data, which is early Q4 next year. We're continuing to do that, and that's really looking at it and based on our current situation, without any other influences, be it a financing partnering that type of thing. It's unfortunate to lose really great colleagues. It's a strategic decision we had to make. That annualizes about a $30 million cost savings, but we really won't start seeing that until middle of next year, due to severance and WARN notice and things like that.
Were there reductions on the clinical side at all? Like, is there a difference between the wet AMD program and the DME program and the people who are involved?
I would say reductions were really across the board. We've kept our core capabilities, certainly in clinical. If you look at EyePoint as a company, our spend is its people and its clinical trials. So we've really looked to moderate that to make sure that we've got the right people in the right places to make sure we continue to execute on DME.
Continue to enable our ability to make the product in our Northbridge facility to support an NDA. Which, by the way, is also why, even though we've announced that we're concluding year two of the wet AMD trials, we've extended it to some degree because we wanted to make sure that we had enough patients through year two to support the safety end of 300 patients for a DME NDA filing. There's a couple of moving parts there that we're still working through, all with an eye toward managing capital, getting to key inflection points, and managing our burn.
Okay. I guess last big picture question. I mean, as you think about wanting to approach the FDA with some of your findings and knowing how FDA has been flexible with some other programs, what do you think that you could potentially negotiate with them?
For wet AMD or DME?
For wet AMD. Yeah, for DME, for any of this.
Yeah.
Do you think the FDA is going to be workable with you?
Oh, absolutely. They've been great partners with us so far. We believe that if, in fact, we can find scientific evidence that there is a biomarker or subgroup that does well with our drug, we look forward to having formal discussions with the FDA about the design of a trial and size and length and things like that. No, they've been a terrific partner for us.
They've obviously shown a tremendous amount of flexibility with this new class, given other players out there.
I'm not sure I would say that.
Or no, maybe not. Maybe that's a controversial statement.
No, I'm not sure I would agree that they've shown a lot of flexibility. I think that they've been consistent.
That's really what you want. There are guidelines around the design of these trials. I mentioned, for example, having to enroll patients that actually have active wet AMD. Those are the kind of things that I don't believe they're going to change.
Okay.
Issues around safety and things like that. My point is, they have been very open and honest, but I do believe that they are following the rules. That's great because we know exactly where we stand.
Okay. Fantastic. All right. Well, thank you very much for your time.
Thank you.
Thank you.
Thanks, everybody.