CLAiR System: Can AI Reliably Assess CV Risk From Retinal Images?

A new artificial‑intelligence (AI) system called CLAiR, which analyzes retinal images to estimate cardiovascular risk, showed strong agreement with standard atherosclerotic cardiovascular disease (ASCVD) risk‑score calculators, according to findings presented during an Investigative Horizons session at ACC.26. 

The prospective U.S. study enrolled 874 participants aged 40–75 without known atherosclerosis or lipid‑lowering therapy at 10 eye care and primary‑care sites.

Using standard retinal‑camera images, the CLAiR AI model assessed cardiovascular risk in 94% of captured images and achieved 91.1% sensitivity and 86.2% specificity compared with conventional 10‑year ASCVD risk estimates; about 26% of participants had an ASCVD risk ≥7.5%. 

The system is not intended to replace formal risk assessment, but may serve as a noninvasive, opportunistic screen during routine eye exams to flag patients who would benefit from preventive cardiology evaluation.

Investigators emphasize that for CLAiR to have real‑world impact, health systems will need structured referral pathways linking elevated AI‑predicted risk from the eye clinic to primary‑care and cardiovascular‑prevention visits, ideally with guideline‑directed therapy for those at higher risk. 

Retinal imaging is already widely available in U.S. eye clinics, though coverage and access vary, and CLAiR is not designed for use in pregnant individuals or those with advanced ocular disease.