This study, conducted at a single center in Beijing, evaluated whether quantitative retinal vascular parameters relate to presence of coronary artery disease (CAD) and whether a retinal phenotype–based model can aid early non-invasive screening. The analysis included 417 adults with suspected angina who underwent their first coronary angiography; high-quality fundus images were required within 24 hours prior to angiography.
CAD was defined as ≥50% stenosis on angiography. Compared with non-CAD controls (n=190), CAD patients (n=227) showed higher prevalence of hypertension, dyslipidemia, and diabetes, and greater fasting glucose, lipoprotein(a), triglycerides, and HbA1c.
The primary aim was to quantify associations between retinal vascular metrics and CAD status. A fully automated artificial intelligence pipeline measured multiple retinal features, including fractal dimension (FD) and vessel diameter (VD) among others.
Three predictive models were assessed: retinal parameters alone, traditional risk factors alone, and a combined model incorporating both retinal and clinical data. The combined model achieved favorable diagnostic performance with high interpretability, suggesting potential utility for early CAD screening.
BMJ Open published a clinical update in Research Highlights on 02 Apr 2026.
The item focuses on Retinal vascular phenotyping for early detection of coronary artery disease: quantitative assessment and diagnostic modelling.
Review the original article for the full source wording and details.