Abstract
PURPOSE: The purpose of this study was to determine which baseline factors, including ultra-widefield fluorescein angiography (UWF-FA) nonperfusion and macular optical coherence tomography angiography (OCTA) metrics, predict 2-year progression of retinal nonperfusion on UWF-FA in diabetic retinopathy (DR). METHODS: This prospective observational study included 73 patients (110 eyes) with diabetes across the spectrum of DR severity, followed for 2 years. Retinal nonperfusion was manually quantified as ischemic index, and progression was calculated as the 2-year minus baseline ischemic index for gradable retina. For receiver operating characteristic (ROC) analyses, progression was further defined as a binary event when the 2-year ischemic index increase exceeded two standard deviations (SDs) above the mean progression observed in non-referable DR. Linear mixed-effects models evaluated baseline predictors, and ROC analyses assessed the discriminative performance of significant predictors. RESULTS: Mean 2-year ischemic index progression was 0.45 ± 0.90%, with greater progression in moderate and severe nonproliferative DR than in non-referable DR. In multivariate models, greater baseline ischemic index (standardized β = 0.27, P = 0.002), higher deep capillary plexus (DCP) geometric perfusion deficit (β = 0.36, P = 0.006), and lower DCP vessel density (β = -0.22, P = 0.012) independently predicted total nonperfusion progression. Baseline ischemic index, DCP geometric perfusion deficit, and DCP vessel density yielded areas under the curve of 0.84, 0.77, and 0.70, respectively, for detecting binary nonperfusion progression. CONCLUSIONS: In this cohort, baseline UWF-FA nonperfusion was the strongest predictor of retinal nonperfusion progression, whereas macular DCP OCTA metrics provided independent, noninvasive prognostic information. Macular DCP OCTA parameters may serve as adjunctive biomarkers to refine DR risk stratification, particularly when UWF-FA is unavailable.