Geographic atrophy progression in routine clinical practice: before and after pegcetacoplan treatment

Abstract

Purpose : Pegcetacoplan (Syfovre, Apellis) was the first FDA-approved therapy for the treatment of geographic atrophy (GA), a leading cause of blindness. There is a need for data regarding efficacy with pegcetacoplan in routine clinical practice.

Methods : Retrospective IRB-approved analysis of eyes receiving ≥ 3 pegcetacoplan injections with ≥ 5 years of data prior to dosing, from July 2017 to September 2023. Total GA lesion area was calculated from optical coherence tomography (Spectralis, Heidelberg Engineering) using RetinAI Discovery, an artificial intelligence software platform (Ikerian AG, formerly RetinAI). GA growth rates were determined for baseline to first pegcetacoplan (P1) & first pegcetacoplan to most recent clinical visit (T1). Paired t-tests used to assess significance.

Results : 152 patients (231 eyes) included. Baseline mean age (SD) was 77.2 (8.8) & 63 (41%) were male. Mean best recorded VA was 20/60 Snellen, 72 (31%) were phakic, & 132 (57%) had concurrent neovascular AMD. During T1, mean (SD) follow-up interval was 4.5 (0.9) months & mean number of pegcetacoplan injections was 3.6 (0.9), with mean interval of 7.5 (0.7) weeks between injections. Among all eyes with measurable GA at baseline, there was no significant difference between P1 & T2; mean (SD) annual GA area growth rates were 0.98 (0.8) mm², 95% confidence interval (CI) [0.9, 1.1] vs 0.98 (5.0) mm², 95% CI [0.3, 1.7]. Data from eyes with GA area between 2.5 & 17.5 mm² (n = 66, 35%) were similar.

Conclusions : These preliminary data reflect short-term pegcetacoplan treatment in routine clinical practice from a mixed population with & without sub-foveal GA, & with more than half having concurrent neovascular AMD. Nevertheless, the annual GA growth rate as measured over a preceding 5-year interval without treatment was approximately 1 mm², consistent with data from natural history reports, supporting the validity of the current version of this AI modeling. Looking forward, longer-term data with pegcetacoplan treatment will be analyzed and several algorithms are under development, including a projection algorithm to estimate theoretical GA growth with & without pegcetacoplan treatment, as well as additional segmentation abilities to assess longitudinal changes in additional retina layers compared to previous reports from the DERBY, OAKS & GALE studies of pegcetacoplan.

This abstract was presented at the 2024 ARVO Annual Meeting, held in Seattle, WA, May 5-9, 2024.