Longevity Briefs: AI Could Predict Your Risk Of Heart Disease Just By Looking At Your Eye

Longevity briefs provides a short summary of novel research in biology, medicine, or biotechnology that caught the attention of our researchers in Oxford, due to its potential to improve our health, wellbeing, and longevity.

Why is this research important: Cardiovascular disease is the World’s leading cause of death by far. Research suggests that up to 75% of strokes and ‘cardiovascular events’ (such as heart attacks) can be prevented if signs of cardiovascular disease are detected early on. Identifying high-risk people generally requires blood pressure measurements and blood tests, which need to be done in a clinical setting. This makes screening healthy people for early signs of heart disease challenging. Some research suggests that underlying cardiovascular problems, such as narrowing of the coronary artery in the heart, can lead to narrowing of the tiny arteries in the retina – the part of the eye that converts light into electrical signals that are sent to the brain. Because of this, scientists hope that imaging the retina could be used as a quick and non-invasive way to calculate a person’s risk of cardiovascular disease.

What did the researchers do: Here, researchers developed an AI program called QUARTZ (QUantitative Analysis of Retinal vessels Topology and siZe, for those wondering) that takes retinal images and computes the width, area, and ‘curviness’ of blood vessels. They then applied QUARTZ to over 88 000 retinal images from participants of the UK Biobank, which is a large research database of medical information from over half a million participants. By comparing these images (alongside age, sex, smoking status and medical history) to known risk of cardiovascular disease for these individuals, they were able to develop prediction models for cardiovascular risk.

Key takeaway(s) from this research: Researchers then applied these prediction models to a new group of 7411 individuals aged 48-92. Participants were followed up for 7-9 years to see how many would go on to die from cardiovascular diseases.

In terms of calculating cardiovascular risk, QUARTZ performed similarly to the current ‘gold standard’ the Framingham risk score, which uses age, cholesterol, smoking, blood pressure and medical history to estimate cardiac events risk over the subsequent 10-years. The Framingham risk score was better at predicting heart attacks in those most at risk, while QUARTZ was slightly better at predicting stroke in men, and slightly worse at predicting stroke in women.

The authors suggest that since retinal imaging is already commonplace in many countries including the US and UK, it could be used as a quick and non-invasive cardiovascular health check to identify those at medium to high risk, who could then get further more detailed risk assessment.