Longitudinal Changes in Inner Retinal Layers in Treatment-Naïve Neovascular AMD During Intravitreal Anti-VEGF Therapy.

PURPOSE: To investigate longitudinal changes in retinal nerve fiber layer (RNFL) and ganglion cell- inner plexiform layer (GC-IPL) thickness in treatment-naïve neovascular age-related macular degeneration (nAMD) under intravitreal anti-VEGF therapy (IVT), compared with untreated fellow eyes with dry AMD (dAMD). DESIGN: Retrospective single-center cohort study. SUBJECTS: Consecutive patients with treatment-naïve nAMD receiving anti-VEGF therapy under a treat-and-extend regimen and a minimum follow-up of 24 months were included. Untreated fellow eyes with dAMD served as controls. METHODS: RNFL and GC-IPL thickness were assessed longitudinally using spectral-domain optical coherence tomography and RetinAI analysis. Best-corrected visual acuity (BCVA) and structural retinal changes were evaluated over time. Linear mixed-effects models were used to assess factors associated with retinal layer changes. MAIN OUTCOME MEASURES: Longitudinal changes in RNFL thickness, GC-IPL thickness, and BCVA. RESULTS: In nAMD eyes, median RNFL thickness in the 3 mm ring zone decreased from 32.3 μm (30.1-35.4) at baseline to 30.9 μm (29.0-33.4; adjusted P < .001) after the loading phase and remained stable thereafter (all adjusted P > .05), with a similar pattern observed in fellow dAMD eyes. In contrast, GC-IPL thickness progressively declined in both groups. In nAMD eyes, median GC-IPL thickness decreased from 89.1 μm (81.9-94.9) at baseline to 85.1 μm (77.7-90.2) after loading (adjusted P < .001), followed by continued thinning throughout follow-up. BCVA in nAMD eyes improved after loading with a median gain of 4.8 ETDRS letters (0-10.2; adjusted P < .001), remained stable for up to 3 years, and gradually declined thereafter. Mixed-effects modeling showed RNFL thickness was associated with injection number, whereas GC-IPL loss was associated with age and follow-up duration, independent of injection number or sex. CONCLUSIONS: The regression in RNFL and GC-IPL thickness during the loading phase in eyes with nAMD may reflect the resolution of pre-treatment swelling. RNFL thickness remained stable thereafter, while GC-IPL thickness and BCVA declined after loading similarly to fellow eyes with dAMD. According to the mixed-effects model, GC-IPL thinning appears linked to progression of the underlying degenerative process, whereas RNFL thickness appears comparatively preserved.