Agreement Between Color, Fluorescein Angiography, and Spectral Domain Optical Coherence Tomography in the Detection of Macular Fibrosis in Neovascular Age-Related Macular Degeneration.

Abstract

To evaluate the agreement between fibrosis on color imaging-, fluorescein angiography (FA)-, and spectral domain optical coherence tomography (SD-OCT)-detected hyperreflective material (HRM) and assess their clinical relevance. Clinical cohort and diagnostic accuracy study. Multimodal fundus images (color, FA, and SD-OCT) of 130 eyes with neovascular age-related macular degeneration, collected 18 months after the initiation of anti-vascular endothelial growth factor treatment (anti-VEGF) as part of the Early Detection of Neovascular AMD (EDNA) study, were regraded for fibrosis and HRM. HRM location was assigned as subretinal (SR) and/or subretinal pigment epithelial (SPE). Agreement between detection methods was assessed with the kappa statistic, and regression analysis with best corrected visual acuity (BCVA) as outcome variable was used to evaluate clinical relevance. Kappa was 0.56 for FA and 0.40 for HRM on SD-OCT, using color as the reference. Regression against BCVA showed low R² values for all tests (R² < 0.09). In the HRM-OCT model, with no HRM as the reference and location and dimensions as covariates, the R² increased to 0.301. Letter loss was 21.1 (P < .0001) for SR and 8.1 (P = .045) for SPE. When HRM on SD-OCT was the reference, the sensitivity for color and FA combination was 87.5% but was lower at 33.3% for SPE only. Well-defined HRM detected by SD-OCT during the maintenance phase of anti-VEGF therapy explains BCVA variance, positioning OCT as a superior standard for detecting fibrosis.