Abstract
PURPOSE: To quantify choriocapillaris (CC) flow deficits using projection-resolved optical coherence tomographic angiography (PR-OCTA) and to evaluate whether they are correlated with diabetic retinopathy (DR). METHODS: In this retrospective study, OCTA scans covering a range of DR severities were acquired. Shadowing artifacts caused by hard exudates, large inner retinal vessels, and vitreous floaters were detected, along with the retinal fluid area. The CC deficit ratio (percentage of deficit area) was measured and compared with shadowing artifacts and area beneath retinal fluid excluded to assess the effect of these confounding features. The Spearman correlation between the CC deficit ratio and disease severity was calculated, as well as the diagnostic performance (area under the receiver operating characteristic curve). RESULTS: One hundred eighty-three participants were enrolled in this study. We found a significant positive correlation between age and CC deficit ratio in normal controls (Pearson correlation, P = 0.02). There was no significant difference in deficit ratio between measurements excluding only shadows and those excluding both shadows and the area beneath the fluid (Mann-Whitney U-test, P = 0.81). The CC deficit ratio was positively correlated with DR severity (Spearman correlation, P < 0.01). It was significantly higher in nonproliferative DR and PDR patients compared to normal controls (Tukey HSD, P < 0.01) but not in those with diabetes without DR. The proposed CC deficit segmentation method achieved high repeatability with an intraclass correlation coefficient of 0.92. CONCLUSIONS: The CC deficit ratio measured by the proposed method correlates with the presence and severity of DR.