Diagnostic Accuracy of Optical Coherence Tomography Angiography versus Fluorescein Angiography in the Detection of Diabetic Macular Ischaemia: A Cross-Sectional Comparative Study

Abstract

Background: Diabetic macular ischaemia (DMI) is a vision-threatening manifestation of diabetic retinopathy (DR) characterised by enlargement of the foveal avascular zone (FAZ) and capillary dropout, traditionally identified by fundus fluorescein angiography (FFA). Optical coherence tomography angiography (OCT-A) is a non-invasive, dye-free imaging modality that provides high-resolution depth-resolved maps of the retinal microvasculature and may obviate the risks and limitations of FFA. The present study evaluated the diagnostic accuracy of OCT-A compared with FFA for the detection of DMI in patients with diabetic retinopathy. Methods: A prospective cross-sectional comparative diagnostic accuracy study was conducted over 12 months in a tertiary care vitreoretinal unit. One hundred eyes of 80 consecutive patients with type 2 diabetes mellitus and any stage of DR were imaged using FFA and 3 × 3 mm macular OCT-A on the same day. DMI on FFA was defined according to the Early Treatment Diabetic Retinopathy Study (ETDRS) classification. On OCT-A, FAZ area, FAZ perimeter, circularity, vessel density of the superficial and deep capillary plexuses, and qualitative features were measured. Sensitivity, specificity, positive and negative predictive values, area under the receiver operating characteristic (ROC) curve and Cohen's kappa for agreement were calculated, with FFA as the reference standard. Results: DMI was identified by FFA in 35 of 100 eyes (35.0%). The mean FAZ area on OCT-A was significantly larger in eyes with DMI (0.46 ± 0.14 mm²) than in those without (0.26 ± 0.07 mm²; p < 0.001). The area under the ROC curve for OCT-A FAZ area was 0.912 (95% CI 0.852–0.972). At an optimal cut-off of 0.36 mm², OCT-A showed sensitivity 85.7%, specificity 87.8%, positive predictive value 81.1% and negative predictive value 91.5%. Inter-modality agreement was substantial (Cohen's kappa = 0.78; p < 0.001). Vessel density of both capillary plexuses was significantly reduced with progressive DR severity. Conclusion: OCT-A demonstrated high diagnostic accuracy for DMI compared with FFA, with substantial inter-modality agreement, and offers the additional advantages of being noninvasive, rapid, and capable of depth-resolved quantification. These findings support the routine use of OCT-A as a complementary or alternative tool to FFA in the assessment of macular ischaemia in patients with diabetic retinopathy.