The following transcript has been edited for clarity.

Hi, I'm Roger Goldberg, MD, MBA, from Bay Area Retina Associates in Walnut Creek, California. It’s great to be down here in Southern California at Retina Society 2026. I was pleased to present today data on functional optical coherence tomography (OCT-F) from the OAKS and DERBY studies, which looked at pegcetacoplan (Syfovre; Apellis/Biogen) for the treatment of geographic atrophy (GA).

We know when we try to look at a functional endpoint in GA, that microperimetry is probably our best modality, but a standard 68-point MAIA grid has several limitations, including the fact that the spots themselves, these 68 spots, only cover about 3% of the macula and each spot is about 2° apart. So there’s lots of missing spots of vision. This is particularly problematic for a disease like GA, which only progresses about 100 µm to 300 µm per year. So microperimetry as a functional outcome, although it’s very sensitive, is very difficult to use in the real world.

When we think about those 68 points, the central 4, which represent the central 3° of vision, are the most critical. And we actually saw a great correlation between the number of scotomatous points of those central 4 and the baseline visual acuity in the OAKS study. And we also saw, among the patients who did not have an absolute scotoma in those central 4 points, if they progressed to having all 4 points being scotomatous, they lost significantly more vision than if they were able to maintain at least some retinal sensitivity in those 4 points. It’s been previously shown that pegcetacoplan, whether dosed monthly or every other month, led to about a 33% risk reduction in developing a scotoma of those 4 points.

To overcome some of the limitations of microperimetry testing—including that it can take up to 8 to 9 minutes per eye in clinical trials, so it’s really not a practical test for routine clinical practice—this study looked at using something called functional OCT, or OCT-F, where you take the spectral-domain OCT images and you take the microperimetry data and you can actually correlate each specific A-scan on the OCT with a microperimetry result. And you put all of this through a deep learning model. It converts the volumetric scan of an OCT into a 2-dimensional image, an en face image, with the inferred retinal sensitivity. There’s actually great correlation between these results and the microperimetry results obtained during the OAKS study.

So what happens then instead of just looking at those 4 points in the central subfield, what if we look at the inferred retinal sensitivity in the central 1 mm, the central subfield, on an OCT? And here you can look at data from both DERBY and OAKS. Even though DERBY didn’t have actual microperimetry, both studies acquired these OCT scans over the course of the study. And again, we see a protective effect of pegcetacoplan at reducing the risk, by about one-third, of developing an absolute scotoma in the central subfield on OCT.

Does this benefit vision? It sure does. Amongst the patients who develop an absolute scotoma in the central subfield of inferred retinal sensitivity of basically zero, those patients lost about 15 letters of vision over the 2 years of the study. And if they’re able to maintain at least some retinal sensitivity in that central subfield, they lose about 44% less vision, so about 8 letters of vision loss.

So this is another way to confirm and look at a benefit of pegcetacoplan treatment, but instead of using microperimetry, using the inferred retinal sensitivity on OCT-F.

Thanks so much for listening. RP