Prognostic Ability of Automated RPE Sinuosity for Late AMD Supersedes Qualitative Outer Retinal Band Assessment.

Purpose: Outer retinal band integrity strongly predicts late age-related macular degeneration (AMD), however, it is often assessed subjectively and vulnerable to intergrader variability. This study quantifies retinal pigment epithelium (RPE) sinuosity from automated segmentations of a commercially available optical coherence tomography (OCT) and its ability to predict late AMD. Methods: Consecutive patients with intermediate AMD who progressed to late disease (n = 69) or remained stable (n = 198) were recruited. Outer retinal band integrity was measured quantitatively using RPE sinuosity, calculated as the ratio of RPE:RPE-fit line length across macular B-scans. RPE sinuosity was measured in the foveal B-scan alongside the mean, maximum, and median values across the five central B-scans. The RPE and ellipsoid zone (EZ) continuity was graded qualitatively, as well as other AMD biomarkers currently used for prognostication. Adjusted odds ratios for progression were calculated for each variable. Results: Mean follow-up time was 3.15 years (1.93 years). Pigmentary abnormalities (P = 0.001), subretinal drusenoid deposits (SDDs; P > 0.0001), drusen volume (P = 0.01), and late AMD in the fellow eye (P = 0.001) were present in a higher proportion of progressor than stable eyes at baseline. Median RPE sinuosity was the strongest predictor of progression within 5 years (2.52 [1.14-5.55]) and across the study period (2.35 [1.25-4.43]). No qualitative measures could predict late AMD within any time frame. Conclusions: RPE sinuosity outperformed qualitative outer retinal band integrity assessments for late AMD prediction and is a reliable, repeatable evaluation method. Verifying these results using different OCT devices in a prospective cohort is needed to determine whether the feature can enhance late AMD risk models and improve clinical management.