Work

SS-OCTA versus ICGA: a comparison of detection efficiency for large choroidal vein patterns in AMD patients

Year: 2025

Type: article

Abstract: PURPOSE: To compare the efficacy of swept-source optical coherence tomography angiography (SS-OCTA) and indocyanine green angiography (ICGA) in visualizing and classifying large choroidal vein topography, and to evaluate the association between specific venous patterns and pachychoroid spectrum disease (PSD)-associated macular neovascularization (MNV). METHODS: This retrospective cross-sectional study included 222 eyes from 160 patients with neovascular age-related macular degeneration (nAMD). All participants underwent both SS-OCTA and ICGA imaging. Choroidal venous patterns were classified into four types based on ICGA venous phase images: watershed, anastomotic, upper thicker, and lower thicker. SS-OCTA images were processed and analyzed using Fiji software. Diagnostic agreement between modalities was assessed using Cohen's Kappa. Logistic regression was used to evaluate the association between venous patterns and disease subtypes. RESULTS: SS-OCTA and ICGA showed almost perfect agreement in classifying choroidal venous patterns in discernible cases (unweighted Kappa = 0.978). Among 77 eyes where ICGA failed to provide a clear classification, SS-OCTA achieved a definitive diagnosis in 87.0 % (P < 0.001), with the anastomotic type being the most common. Furthermore, the watershed pattern was significantly more prevalent in non-PCV nAMD compared to PCV (29.8 % vs. 12.1 %, p = 0.002), suggesting distinct choroidal venous backgrounds between these subtypes. CONCLUSION: SS-OCTA provides a reliable, non-invasive alternative to ICGA for visualizing and classifying large choroidal veins, with superior performance in cases obscured by hemorrhage or leakage. The distinct distribution of venous patterns across nAMD subtypes highlights the potential role of choroidal venous architecture in disease phenotyping and pathogenesis.

Source: Photodiagnosis and Photodynamic Therapy

DOI: 10.1016/j.pdpdt.2025.105303