In Vivo Quantification of Bruch's Membrane Changes in Angioid Streaks and Neovascular Age-related Macular...

To investigate Bruch's membrane (BrM) morphology in angioid streak (AS)-associated macular neovascularization (MNV) and neovascular age-related macular degeneration (nAMD) using multimodal imaging.

Retrospective clinical cohort study.

Treatment-naïve eyes with MNV secondary to AS or nAMD that achieved a dry macula after 3 loading intravitreal antivascular endothelial growth factor (VEGF) injections.

Baseline indocyanine green angiography (ICGA) images obtained before antiVEGF treatment were used to measure axial length-corrected MNV area and identify late-phase speckled hypofluorescent areas (SHA). Optical coherence tomography (OCT) images after dry macula achievement were used to assess central foveal thickness (CFT), subfoveal choroidal thickness (SCT), fibrovascular pigment epithelial detachment (PED) thickness, BrM undulation, and a newly defined BrM reflectivity index (BRI) calculated from intensity-normalized images. Group comparisons assessed BrM-related biomarker differences, and receiver operating characteristic (ROC) analyses evaluated their diagnostic performance.

Differences in ICGA- and OCT-derived findings between AS-associated MNV and nAMD, including SHA, BrM undulation, and BRI, and the diagnostic performance of BRI and BrM undulation in differentiating AS-associated MNV from nAMD.

Twenty-five eyes with AS and 32 eyes with nAMD were analyzed. Among eyes with available ICGA, MNV area was larger in AS than in nAMD (5.34 ± 3.29 mm² vs 3.45 ± 1.22 mm², P = .005), and late-phase SHA was more frequent in AS (86.4% vs 18.8%, P < .001). CFT and fibrovascular PED thickness did not differ significantly, whereas SCT was lower in AS-associated MNV than in nAMD (166.0 ± 67.1 µm vs 217.2 ± 112.0 µm; P = .049). BrM undulation was more common in the AS group (24.0% vs 3.1%, P = .036). BRI was higher in AS than in nAMD (0.905 ± 0.148 vs 0.663 ± 0.169, P < .001). The ROC curve for BRI showed an area under the curve of 0.850 (95% confidence interval, 0.754-0.946) with an optimal cutoff of 0.749, yielding 0.880 sensitivity and 0.656 specificity.

Compared with nAMD, AS-associated MNV showed distinct ICGA and OCT features, including larger MNV area, more frequent late-phase SHA and BrM undulation, and higher BRI after antiVEGF loading. These findings may reflect BrM alterations associated with calcification and mechanical instability. BrM-derived ICGA and OCT biomarkers may help characterize AS pathophysiology and differentiate AS-associated MNV from nAMD.