Eye scans offer heart disease hope

Scientists have trained an artificial intelligence (AI) system to analyse retinal scans and identify whether a patient was likely to have a heart attack within a year.

Researchers, led by scientists at the University of Leeds, have identified that minute changes to blood vessels in the retina are indicators of broader vascular disease, including heart complications.

In the Nature Machine Intelligence journal, these researchers explained that the system was 70-80% accurate and could be useful as a triage tool.

The researchers added that the new AI system could revolutionise the way patients are screened for signs of heart disease, as pumping efficiency of the heart’s left ventricle can currently only be measured via diagnostic tests such as echocardiography or magnetic resonance imaging.