Abstract
PURPOSE: The purpose of this study was to identify novel features on spectral-domain optical coherence tomography angiography (SD-OCTA) that improve detection of polypoidal lesions in polypoidal choroidal vasculopathy (PCV). METHODS: In this prospective study, 22 eyes from 20 patients with indocyanine green angiography (ICGA) confirmed PCV underwent same-day imaging with both SD-OCTA and swept-source OCTA (SS-OCTA). Images were independently reviewed, and blinded to ICGA and each other. Morphological features of polypoidal lesions were compared to establish combined diagnostic criteria to improve SD-OCTA detection, including hyper- or hypo-flow signals on angiographic en face images, circular hypo-reflective signals on structural en face images, and corresponding flow signals within retinal pigment epithelial detachments (PEDs) on B-scans. Hyporeflective signals on structural en face images were defined as well-circumscribed and low-reflectivity areas. Diagnostic accuracy was assessed using sensitivity, specificity, positive predictive value, negative predictive value, and area under the curve (AUC). RESULTS: ICGA identified 52 polypoidal lesions. SS-OCTA showed superior detection (98%) compared to SD-OCTA (69%) after manual segmentation correction. Hyporeflective signals on SD-OCTA structural en face images were observed in 73% of lesions, higher than SS-OCTA (52%). Applying combined criteria to SD-OCTA enabled detection of 8 additional lesions, raising the detection rate to 85% and diagnostic accuracy to 95%. CONCLUSIONS: This combined approach significantly improves SD-OCTA diagnostic performance and is recommended in clinical practice, especially when standard retinal pigment epithelium (RPE) and Bruch's membrane (BM) segmentations were used. TRANSLATIONAL RELEVANCE: Identifying hyporeflective signals on structural en face SD-OCTA improves polypoidal lesion detection, enhancing the clinical utility of SD-OCTA in diagnosing PCV.