AI Can Now Predict Risk Of Heart Attack & Disease Just From A Regular Eye Check

A routine eye scan is necessary to detail vision-related matters, of course, but there is a possibility that the addition of artificial intelligence will be able to give it yet another function: predicting the risk of a heart attack.

Developed by the University of Leeds, this deep-learning system is touted to read retinal scans to identify the telltale signs of a heart attack with up to 80% accuracy, making it a suitable “second referral” for cardiovascular checkups. 

The AI system identifies “associations between pathology in the retina and changes in the patient’s heart,” according to a statement from the university. Once image patterns were learned, the size and pumping efficiency of the left ventricle could be derived by the technology based on just that scan.

Data on these two aspects of the heart are, currently, only available through diagnostic tests like echocardiography or MRI scans. These can be expensive, as well as hard to access. 

The eye scan, on the other hand, is easier for most to undergo. 

“Retinal scans are comparatively cheap and routinely used in many optician practices,” explains Professor Alex Frangi, supervisor of the study.

“As a result of automated screening, patients who are at high risk of becoming ill could be referred for specialist cardiac services. The system could also be used to track early signs of heart disease.” 

An enlarged ventricle is linked to a higher risk of heart disease. Combining this information with other demographic data, such as a patient’s age, would allow the AI to predict the approximate risk of them suffering a heart attack in the next 12 months.

This system is detailed in a paper published in the journal Nature Machine Intelligence.

[via NewsNation and the University of Leeds, image via the University of Leeds / UK Biobank]