Layer-specific optical coherence tomography angiography biomarkers in diabetic retinopathy: Insights from a small cohort and the road ahead

Dear Editor, Diabetic retinopathy (DR) is a leading cause of preventable blindness worldwide, driven by progressive microvascular damage involving both the superficial and deep retinal capillary plexuses. Chronic hyperglycemia induces endothelial dysfunction, pericyte loss, and capillary dropout, leading to nonperfusion, foveal avascular zone (FAZ) enlargement, and vessel remodeling.[1] Early detection and monitoring of these microvascular changes are essential for timely intervention and preservation of visual function. Optical coherence tomography angiography (OCTA) has transformed DR assessment by enabling non-invasive, depth-resolved visualization of retinal microvasculature. Unlike fluorescein angiography, OCTA provides dye-free, high-resolution maps of perfused vessels in the superficial vascular plexus (SVP), intermediate vascular plexus (IVP), and deep vascular plexus (DVP). These data images from the OCTA machine were imported into OCTAVA (Open Source Software) and analyzed. Quantitative OCTA metrics — such as vessel density, FAZ area, vessel diameter, and tortuosity — have been shown to correlate with DR severity, macular edema, and functional outcomes.[2,3] In our prospective cohort study of 31 eyes (mean age 54 years, 67.7% male), we observed significant associations between diabetes duration and increased vessel diameter and tortuosity in the SVP and IVP, suggesting progressive vascular remodeling with chronic disease. Clinically significant macular edema (CSME) demonstrated robust correlations with tortuosity across all plexuses (SVP, P = 0.007; IVP, P = 0.031; DVP, P = 0.002), implicating widespread microvascular instability in macular edema pathogenesis. Anti-VEGF therapy was associated with a significant reduction in SVP tortuosity (P = 0.036), consistent with vascular stabilization effects reported in other studies.[3] Of particular interest, changes in the DVP — specifically reduced vessel diameter and tortuosity — were strongly associated with worse best-corrected visual acuity (BCVA), suggesting that deeper capillary plexus alterations may disproportionately contribute to visual decline in DR. This finding reinforces the emerging concept that DVP integrity may serve as a sensitive biomarker for visual prognosis [Fig. 1].Figure 1: Study methodThe integration of artificial intelligence (AI) and machine learning (ML) into OCTA analysis has further enhanced its diagnostic and predictive capabilities. For example, Elsharkawy et al.[4] utilized multiple instance learning to classify DR from weakly labeled en face OCTA images, reducing subjectivity and inter-grader variability. Such automated approaches may allow scalable screening and longitudinal monitoring in both clinical and community settings. However, challenges remain in standardizing OCTA acquisition protocols, segmentation algorithms, and analysis pipelines. Ong et al.[5] emphasized that variability across devices and software can hinder cross-study comparability and limit multicenter trial applications. Furthermore, OCTA’s relatively small field of view compared to ultra-widefield fluorescein angiography may overlook peripheral ischemia, though emerging montage and widefield OCTA techniques aim to overcome this limitation. Our small sample size warrants cautious interpretation of these findings, yet they underscore the value of OCTA-derived metrics in elucidating DR pathophysiology and guiding individualized management. Larger, multicenter studies are needed to validate layer-specific biomarkers, establish normative databases, and integrate OCTA into therapeutic decision-making and clinical trial endpoints. Financial support and sponsorship: Nil. Conflicts of interest: There are no conflicts of interest.