AI Eye Scans Could Help Identify Heart Disease and Stroke Risk

By HospiMedica International staff writers
Posted on 02 Apr 2025

Image: Associate Professor Lisa Zhu and Wenyi Hu investigated the use of an AI eye test in GP clinics (Photo courtesy of CERA)

New research has explored the advantages of utilizing artificial intelligence (AI) retinal imaging for screening cardiovascular diseases in general practice (GP) clinics and highlighted areas where improvements can be made.

The study, conducted at the Centre For Eye Research Australia (CERA, Melbourne, VIC, Australia), marks the first time AI-powered eye scans have been tested in GP clinics for cardiovascular disease screening. The research, published in npj Digital Medicine, found that the fast and non-invasive AI retinal scan could be seamlessly incorporated into primary care settings for detecting the risk of heart attacks and strokes. The study involved 361 participants, aged 45-70, who were patients at two GP clinics and had undergone a full or partial cardiovascular risk assessment—such as blood pressure or cholesterol testing—within the last six months. Each participant had a retinal scan taken using a desktop retinal camera, which captures images of the blood vessels at the back of the eye, followed by an AI-generated, real-time cardiovascular disease risk report.

Participants were also assessed using the World Health Organization’s (WHO) CVD risk chart, a widely recognized clinical tool that considers age, sex, smoking status, blood pressure, diabetes, and total cholesterol. The results from the retinal scans were compared to the WHO CVD risk scores to assess the correlation between the two methods. Both scoring systems were then validated using data from over 27,500 records from the UK Biobank. The comparison revealed that the retinal scan-based risk scores provided similar accuracy to the WHO scores in predicting the likelihood of heart attack or stroke over the next decade. The study showed a moderate correlation between the retinal scan results and the WHO risk scores—67.4% of results matched, while 17.1% were overestimated and 19.5% were underestimated by the retinal scan.

The retinal scan and WHO method displayed comparable abilities in predicting 10-year risks for coronary heart disease or stroke when matched against data from the UK Biobank. The study also found a 93.9% imaging success rate, with the majority of patients receiving a retinal scan that could be graded for cardiovascular risk. A survey of patients and eight participating GPs assessed their satisfaction with the technology and its potential for future use. Results showed that 92.5% of patients and 87.5% of GPs were satisfied with the AI-powered retinal scan. The researchers believe retinal scanning holds significant potential in expanding cardiovascular risk assessments in Australia. In the future, this technology could serve as an early triage tool, alerting GPs to patients who may require further investigation or testing.

“In the future I would see patients receiving an SMS, reminding them to have an eye scan which would send a report on their risk to their doctor who could then follow up with further tests if needed,” said Dr. Malcolm Clark, a Camberwell general practitioner whose clinic took part in the research and a co-author on the paper. “This could become part of regular health screenings like a cervical screening test or a fecal occult blood test.”

Related Links:
CERA