The following transcript has been edited for clarity.

Charles C. Wykoff, MD, PhD: Hi, I’m Charles Wykoff from Houston, Texas. Great to be here with you to talk about some fascinating new data.

Carl Regillo, MD, FACS: And I'm Carl Regillo. I’m from Philadelphia, Mid-Atlantic Retina/Wills Eye Hospital.

Dr. Wykoff: So we’re at Retina Society 2026 here in Los Angeles, and we’re discussing the LUGANO top-line data, primary and secondary endpoints. Carl, tell us what you observed.

Dr. Regillo: LUGANO was 1 of the 2 phase 3 studies for EYP-1901 (Duravyu; EyePoint) in the treatment of neovascular age-related macular degeneration (AMD). We have the top-line results, and we’re presenting it for the first time at this meeting. So what did it show? Well, it was a head-to-head study of EYP-1901 dosed every 6 months after loading phase compared to on-label aflibercept 2 mg (Eylea; Regeneron) dosed every 8 weeks, again after a loading phase. And in this study, of course, both arms could be supplemented as needed depending on certain criteria. So that’s the basic study design. Top-line or primary endpoint was mean best-corrected visual acuity (BCVA) change from baseline, compared to the control arm, Eylea, at weeks 52 and 56 with a standard 4.5-letter noninferiority margin.

The groups were well balanced. It was a mix of mostly treatment-naïve wet AMD, some previously treated. They behaved as expected more or less, meaning we saw BCVA increase, it was well maintained in the control group, but it dipped down a little bit in the EYP-1901 arm, and it did not meet the primary noninferiority endpoint.

But, what we saw with the optical coherence tomography (OCT) data was that the mean central subfield thickness (CST) reduction was actually very well maintained, very comparable to the control arm. So we had an unexpected finding here. We didn’t get the visual acuity held as well as we were expecting, but we got disease control as expected.

A deeper dive into the data revealed that there was this disproportionate group of outliers in the EYP-1901 arm that lost 3 or more lines of vision, mostly from non–wet AMD causes: foveal geographic atrophy (GA), some glaucoma progression, and 1 retinal detachment case that lost a lot of vision. So collectively this group brought the mean down. When we ran the data excluding those 9 outliers, the visual results actually turned out pretty good, within the noninferiority margin of just a couple letters. So it didn’t meet the primary endpoint—the visual acuity curves were not noninferior—but there was good evidence that in the full data set that the drug did control disease adequately in the wet AMD setting here.

So we have these unusual outliers, something unexpected, that seemed to have affected the visual outcomes. And you’re presenting the secondary outcomes, and that sheds more light and helps to complete the picture of LUGANO. So now I’ll put it back in your hands.

Dr. Wykoff: Great summary, Carl. And as you look at the secondary endpoints, the key secondary prespecified endpoint was treatment burden. But to get to the treatment burden calculation, we’ve got to look at the supplemental injections first. And as you pointed out, both arms received 3 monthly Eylea injections and the EYP-1901 arm received every-6-month EYP-1901. The control arm was a true on-label Eylea control arm—every-8-week fixed dosing. And the fascinating angle about the supplementation is that both arms were eligible for supplementation. And criteria are always super important to dig into. So the criteria for supplementation were a greater than 5 letter loss with 75 µm increase in fluid or vision-threatening hemorrhage. And again, both arms could receive supplements.

The question to me going into this was, what percent of the Eylea arm is going to get supplement? Because most trials don’t supplement the Eylea control arm when it’s a fixed Q8.

If you look at the supplementation, in the EYP-1901 arm the majority of patients did not need supplements. Seventy-nine percent received zero or 1 supplement through 56 weeks. Only about 7% of patients received supplement for hemorrhage. That was a question going into it. About 3% received supplement for non–prespecified criteria. They didn’t meet the criteria. So the vast majority, about 90% received supplementation for that BCVA and CST change.

About 10% of the Eylea control arm received supplementation. So using those numbers and calculating the treatment burden, it was [3.1 vs 5.3] for the number of injections after the loading phase in the EYP-1901 vs the aflibercept control arm, which was a 42% reduction in treatment burden with EYP-1901 with an annualized mean of 1.1 rescues per year in the EYP-1901 arm. So it showed that in a subpopulation of patients, the majority of patients, 54% of patients [were] supplement free, and you had very good visual acuity outcomes. Actually the curves were overlapping through the first 32 weeks. And then by the end, there was a slight difference in visual acuity, but still very good control in both from an anatomic perspective.

And then finally from a safety perspective, when you dive into the ocular adverse events, it’s interesting because when you inject Eylea, you have a clear substance vs injecting EYP-1901 or any other drug that is a bioerodible, biodegradable implant, patients are going to see that implant. And so we actually saw an imbalance in vitreous floaters, which was due to seeing those implants, not a fragmented implant, but the implant itself. And that was about 9% vs about 3%, I think. There was a slight imbalance also about 7% vs 2% for wet AMD. And that was due to some of the eyes that needed rescue. The investigator called an adverse event, quite reasonably so.

So again, I think that echoes the point that for a large proportion of patients, we are seeing good control. One of the questions that comes out now looking at the data is how do we find out who those patients are? And then the last point about safety is just to dive into all the intraocular inflammation (IOI) and serious ocular adverse events. There were 2 retinal detachments, 1 in each arm. Actually the eye with EYP-1901 was phakic and they ended up losing about 40 letters—to your point about the imbalance at the end of the trial because of cataract—vs the pseudophakic aflibercept patient actually gained a letter. And there were 2 IOI cases, one actually in each arm. The aflibercept patient had iritis and vitritis that was moderate and the EYP 1901 patient had moderate anterior-chamber cell. It was actually 2+ cell at week 8. Both of them were treated with topical steroids, both resolved, and both continued on-label treatment.

Overall, from a treatment burden perspective, we did see significant reductions in treatment burden. And the challenge is that when we focus on this subpopulation of supplement-free patients, they did great, but in the larger population, we did not hit the primary endpoint.

What are you looking forward to next?

Dr. Regillo: So to me, I see all this data and I say, we’ve got a product that works pretty well across the board and is particularly good in a large percentage of the wet AMD patients. Good safety profile, does what it’s supposed to do, which is to decrease the treatment burden. It’s maintenance-phase sustained delivery of an anti-VEGF–like tyrosine kinase inhibitor (TKI)in this case.

To complete the picture, we need the second phase 3 study. These outliers that I’ve mentioned obviously did affect the mean BCVA and did draw it down enough that it did not reach the primary endpoint. But I still think for the second identical LUCIA study that’s going to have data by the end of this year that there’s a good chance that will hit, and then we'll have the complete data set, pool data, and so forth to work with.

Dr. Wykoff: I agree. I very much look forward to LUCIA data. We’ll see if these outliers are real or if this was truly bad luck. It’ll be very interesting to see what happens with the data. RP