AI-assisted analysis of early fluid dynamics following aflibercept 8 mg in treatment-naïve neovascular AMD

Purpose To evaluate early morphological changes following intravitreal aflibercept 8 mg in treatment-naïve neovascular age-related macular degeneration (nAMD) using artificial intelligence (AI)-assisted optical coherence tomography (OCT) segmentation. Methods Thirty eyes with treatment-naïve subfoveal nAMD received three monthly aflibercept 8 mg injections. OCT scans were obtained at baseline and days 1, 7, 14, 30, 60, and 90. AI-based segmentation software (Fluid Monitor, RetInSight GmbH) quantified intraretinal fluid (IRF), subretinal fluid (SRF), and pigment epithelial detachment (PED) volumes in the central 6-mm area. Morphologic changes were correlated with best-corrected visual acuity (BCVA). Results IRF, SRF, and PED showed distinct reduction patterns. In eyes with IRF ( n = 22), mean volume decreased by 65% at day 1, 94% at day 14, and 96% at day 90 ( p < 0.05). Among eyes with SRF ( n = 28), SRF volume decreased by 18%, 84%, and 99% at days 1, 14, and 90, respectively ( p ≤ 0.002). PED volume declined more gradually among affected eyes ( n = 23), with mean reductions of 7% at day 1, 34% by day 14, and 53% by day 90 ( p ≤ 0.023). BCVA improved from 51.8 ± 14.4 to 62.6 ± 14.4 ETDRS letters by day 90 (mean gain + 10.8; p < 0.001). Multivariate regression identified a significant association between IRF reduction at day 14 and BCVA gain at day 90 ( β = −0.018, p = 0.036). Conclusions Aflibercept 8 mg induced rapid, compartment-specific morphologic responses in treatment-naïve nAMD, with early IRF resolution and significant BCVA improvement. AI-assisted analysis provided detailed pharmacodynamic insights and may support individualised treatment strategies during the loading phase.