Software developed using machine learning could be used to predict someone’s risk of heart disease in less than a minute by analyzing the veins and arteries in the eyes.
The new research, published in the British Journal of Ophthalmology, paves the way for the development of rapid and inexpensive cardiovascular screenings if the findings are confirmed in future clinical trials. These scans can let individuals know their risk of stroke and heart attack without the need for blood tests or even blood pressure measurements.
“This AI tool can let someone know their risk level in 60 seconds or less,” Alicja Rudnicka, the study’s lead author, told The Guardian. Initial tests show that the predictions made by the software are as accurate as those made by the current tests.
The researchers compared the results from their software called QUARTZ (Retinal vessels Topology and size) to 10-year risk estimates produced by the standard Framingham Risk Score test (FRS). They found that the two methods had “comparable performance.”
The software works by analyzing the network of blood vessels found in the retina of the eye. The total area occupied by these arteries and veins, their width, and their “bend” are measured. Because all of these factors are affected by a person’s heart health, the software can make predictions about a person’s risk of developing heart disease simply by looking at a non-invasive snapshot of their eyes.
“The study adds to a growing body of knowledge that the eye can be used as a window to the rest of the body,” said Pearse Keane, ophthalmology and AI analysis researcher, in an interview with The Verge. They know you might see signs of blood pressure, but the problem was that it was difficult to do a manual assessment.” Keane says using machine learning can overcome this challenge.
The use of AI in disease diagnosis from eye scans is one of the fastest growing areas of machine learning in medicine. The first AI diagnostic device approved by the FDA was used to screen for eye diseases, and research shows that this is how AI can detect conditions ranging from diabetic retinopathy to Alzheimer’s (Keane’s own field of research). Tools that implement these findings are currently in various stages of development, but questions remain about the reliability and universality of their diagnostics.
28708
