A recent study funded by Apellis Pharmaceuticals evaluated changes in artificial intelligence (AI)-derived optical coherence tomography (OCT) biomarkers in routine clinical practice before and after pegcetacoplan (SYFOVRE) treatment initiation in twopatient categories: those diagnosed with geographic atrophy (GA) secondary to age-related macular degeneration (AMD) as well as concurrent neovascular AMD (nAMD).
These findings were presented at the American Society of Retinal Specialists (ASRS) 2026 meeting in Montréal, Canada.
Give me some background on SYFOVRE.
SYFOVRE (pegcetacoplan injection 15 mg/0.1 mL) is a C3 complement inhibitor approved by the FDA in 2023 the treatment of GA secondary to AMD.The basis for its approval: Two phase 3 pivotal clinical trials (OAKS [NCT03525613] and DERBY [NCT03525600]) comparing the efficacy and safety of SYFOVRE in monthly or bimonthly dosing with the use of sham injections.
- The findings: Demonstrated a clinically meaningful reduction in GA lesion growth—36% in DERBY and 29% in OAKS—when compared to the sham injections from months 18 to 24.
However: Patients with GA and concurrent nAMD at baseline were excluded from enrollment.
Why does this matter?
The effect of pegcetacoplan observed in GA and nAMD is clinically meaningful because this subpopulation is common in the real-world but has been excluded from enrollment in all prior clinical trials.
As such: This data could be used by clinicians to help inform their use of pegcetacoplan in patients with GA and concurrent nAMD.
Backtrack for a moment—wasn’t Apellis in the news recently?
Indeed; a couple times over, in fact. In early April, Biogen Inc. and Apellis initiated a definitive merger agreement for Biogen to purchase the latter at an estimated cost of $5.6 billion, (that deal was finalized a few months later).
Regarding clinical data: Apellis released long-term (5-year) data on SYFOVRE in late 2025—while also announcing plans to submit a new application to the FDA for a pre-filled syringe (PFS) version of the therapeutic in H1 2026.
- Take note: There have been no confirmed submissions as of July 2026.
Now talk about the study.
This was a retrospective analysis of data from the Retina Consultants of America (RCA) network.
What to know: Investigators analyzed changes in AI-derived OCT biomarkers (photoreceptor [PR] and retinal pigment epithelium [RPE] loss) in the GA and GA + nAMD subgroups across natural history (NH) and treatment phases.
Some notes on the design:
- GA with concurrent nAMD was defined as ≥1 anti-vascular endothelial growth factor (VEGF) intravitreal injection during NH
- Anti-VEGF injection frequency was collected before and after pegcetacoplan initiation.
- OCT imaging was processed with the RetinAI Discovery (FDA-cleared platform) and its AI algorithms (GA module, layer and fluid segmentation modules) to segment PR and RPE loss (layer thickness <5 µm).
Tell me about the patients comprising the study cohort.
Eligible eyes had:
- ≥12 months natural history (NH) prior to pegcetacoplan treatment
- ≥3 OCT acquisitions in NH (Heidelberg Spectralis)
- No lower GA lesion size limit
Eyes from the NH phase were followed into the treatment phase with ≥12 months pegcetacoplan treatment (and ≥5 injections), ≥3 OCT acquisitions, and GA lesion size 1.5-15 mm2 at treatment baseline.
And the findings?
The data was presented by Hasenin Al-khersan, MD, a vitreoretinal specialist at the Retina Consultants of Texas in Houston, Texas.
Based on this reporting: The NH analysis was conducted on a total of 497 eyes from 375 patients (GA: 240 eyes; GA + nAMD: 257 eyes) with a mean NH follow-up of 2.5±0.6 years.
Baseline biomarker magnitudes and NH growth rates were numerically similar between GA and GA + nAMD.
What did OCT biomarker changes look like?
During the NH phase analysis, they were as follows:
- RPE loss rate (p=0.125):
- GA: 1.49 mm2/year
- GA + nAMD: 1.35 mm2/year
- PR loss rate (p=0.026):
- GA: 1.79 mm2/year
- GA + nAMD: 1.55 mm2/year
Moving on to the treatment phase…
In total: 240 eyes from 199 patients met the inclusion criteria (GA: 98 eyes; GA + nAMD: 142 eyes).
Pegcetacoplan was administered approximately every 6.1±4.9 weeks and GA + nAMD eyes received concurrent anti-VEGF at mean of 5.7±4.5 week intervals.
For GA patients:
- PR loss rate decreased by 41% and RPE loss rate decreased by 23% (p<0.001 for both)
- Pegcetaplan delayed projected progression by ~5 months for PR and ~2.8 months for RPE
For those with GA + nAMD:
- PR loss rate decreased by 58% and RPE loss rate decreased by ~40% (p<0.001 for both)
- This represented a numerically larger effect than GA
- SYFOVRE delayed projected progression by ~7 months for PR and ~4.8 months for RPE
Anything else?
Across cohorts, treatment interval patterns were comparable—and biomarker responses showed that faster NH progressors remained faster during treatment, albeit at reduced rates.
Visual acuity and intraocular pressure were similar between the GA and GA + nAMD eyes at NH and treatment baselines.
Looking beyond this: Investigators reported that thefindings align with previous reports from routine clinical practice and also trial-based AI OCT analyses.
And lastly: the take home.
The AI-derived OCT biomarkers demonstrated that pegcetacoplan was associated with meaningful slowing of PR and RPE loss over 12 months in both advanced AMD phenotypes involving GA and GA with concurrent nAMD.
The study authors concluded that the data support the utility of OCT AI analytics for monitoring disease progression and treatment response.