Association between retinal AV ratio and coronary artery disease severity in acute coronary syndrome and chronic coronary syndrome patients: A prospective study

Abstract

Background

Early, non-invasive detection of coronary artery disease (CAD) is a significant challenge. Given the anatomical and pathophysiological parallels between retinal and coronary microvasculature, retinal arteriovenous (AV) ratio may serve as a surrogate marker of CAD. This study aimed to evaluate correlation between retinal AV ratio and CAD severity as assessed by SYNTAX score in patients with acute coronary syndrome (ACS) and chronic coronary syndrome (CCS), using artificial intelligence (AI)–based retinal image analysis.

Methods

In this prospective study, 332 participants were enrolled: 110 with ACS, 120 with CCS, and 102 angiographically normal controls. Retinal fundus imaging was analyzed using AI models (VC-Net and SegFormer) to compute AV ratio. Coronary angiography was performed, and SYNTAX scores were calculated. Correlations between AV ratio and SYNTAX score were assessed, and a machine learning model (RETFound + RBF Kernel Ridge Regression) was developed to predict SYNTAX categories from retinal data.

Results

Mean AV ratios were similar across CCS (0.624 ± 0.11), ACS (0.620 ± 0.12), and controls (0.650 ± 0.10; p = 0.153). In CCS patients, a significant inverse correlation was observed between AV ratio and SYNTAX score (r = −0.344; p < 0.001), which remained after adjustment (r = −0.300; p = 0.002). AI model accurately classified patients into SYNTAX risk categories (94.1 % accuracy).

Conclusion

Retinal AV ratio is significantly associated with CAD severity in CCS patients. An AI-based tool can automatically derive AV Ratio and provide rapid, non-invasive estimate of coronary disease burden, showing promise for risk stratification. This approach warrants further validation in larger cohorts.