Seeing Heart Risk: Retinal AI Could Guide Statin Decisions
A multicenter prospective trial shows a retinal-imaging AI (CLAiR AI) accurately estimates 10-year ASCVD risk and could help identify patients who may benefit from lipid-lowering therapy. Among 847 participants without ASCVD, about 26% had a 10-year risk ≥7.5% (the older threshold), while the 2026 ACC/AHA PREVENT guideline lowers the threshold to 5%, potentially expanding statin eligibility. The AI achieved 91.1% sensitivity, 86.2% specificity, and ~0.96 AUC, using a standard slit-lamp camera that requires no dilation or blood draw and could be used in primary care or optometry. A larger validation is needed, and future outputs should ideally estimate actual disease risk (e.g., coronary plaque, heart attacks) beyond risk scoring.