AI-Driven OCT Biomarkers for Choroidal Neovascularization Assessment in Punctate Inner Choroidopathy and Multifocal Choroiditis.

Purpose To assess potential optical coherence tomography (OCT) biomarkers associated with the presence and future development of secondary choroidal neovascularization (CNV) in inflammatory lesions using artificial intelligence (AI)-based segmentation in patients with punctate inner choroidopathy (PIC) and multifocal choroiditis (MFC). Design Multicenter, retrospective cross-sectional and cohort study. Subjects The study included 208 eyes from 156 patients, comprising 112 patients with PIC (143 eyes) and 44 patients with MFC (65 eyes). Methods Sequential OCT scans from patients with MFC and PIC were analyzed using an AI software tool (Discovery® OCT Biomarker Detector, RetinAI AG, Switzerland). The software evaluated morphological biomarkers, fluid and lesion compartment volumes, and retinal layer thickness. Manual corrections were performed on AI-based segmentations when necessary. Data were compared between MFC and PIC, and between eyes with and without CNV presence and development. Linear mixed-effects modeling was employed to identify predictive biomarkers for CNV development. Main Outcome Measures The primary outcomes were baseline lesion volume and presence of CNV and the secondary outcome was best-corrected visual acuity (BCVA). Results At baseline, CNV was present in 72.6% of the eyes (75.5% in PIC and 66.1% in MFC). During follow-up, 20 eyes developed CNV from inflammatory lesions (16 with PIC and 4 with MFC). CNV-positive eyes had significantly lower BCVA at baseline (0.28 ± 0.26 LogMAR vs. 0.14 ± 0.21 LogMAR; p < 0.0001). Baseline lesion volume was significantly greater in CNV-positive eyes (33.80 nL ± 63.71) compared to CNV-negative eyes (5.44 nL ± 12.44; p < 0.0001). Larger lesion volume (β = 24.29, 95% CI: 9.77-38.80; p = 0.001) and longer disease duration (β = 0.17, 95% CI: 0.07-0.27; p = 0.001) were associated with CNV presence. Conclusions Larger baseline lesion volume and longer disease duration are associated with secondary CNV in PIC and MFC patients. AI-driven OCT analysis has potential to identify key biomarkers, supporting personalized treatment strategies and improving patient outcomes.