Software developed using machine learning can be used to predict a person’s risk of heart disease in less than a minute by analyzing the veins and arteries in their eye.
The new research, published in the British Journal of Ophthalmology, paves the way for the development of rapid and low-cost cardiovascular screenings, if the findings are validated in future clinical trials. These screenings would let individuals know their risk of stroke and heart attack without the need for blood tests or even blood pressure readings.
“This AI tool can let someone know the level of risk in 60 seconds or less,” said study lead author Alicja Rudnicka, told The Guardian. The study found that the predictions were just as accurate as those of the current tests.
“The eye can be used as a window to the rest of the body.”
The software works by analyzing the web of blood vessels in the retina of the eye. It measures the total area covered by these arteries and veins, as well as their width and “tortuousness” (how pliable they are). All of these factors are influenced by a person’s heart health, allowing the software to make predictions about a person’s risk of heart disease simply by looking at a non-invasive snapshot of their eye.
“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, a researcher in ophthalmology and AI analysis who was not affiliated with the study. The edge. “Doctors have known for over a hundred years that you can look into the eye and see signs of diabetes and high blood pressure. But the problem was the manual assessment: the manual delineation of the ships by human experts.” Using machine learning, Keane says, can overcome this challenge.
Using AI to diagnose diseases based on eye scans has proven to be one of the fastest developing areas of machine learning medicine. The first-ever FDA-approved AI diagnostic device was used to screen for eye disease, and research suggests that AI can detect a range of conditions in this way, from diabetic retinopathy to Alzheimer’s disease (Keane’s own area of research). Instruments applying these findings are at various stages of development, but questions remain about the reliability and universality of their diagnoses.
This recent study, conducted by a team from St George’s, University of London, has only tested the eye scans of, for example, white patients. The team pulled their test data from the UK Biobank, a database that happens to be 94.6 percent white (reflecting the UK’s own demographics in the age range of patients included in the BioBank). Such biases should be weighed up in the future to ensure that each diagnostic tool is equally accurate for different ethnicities.
The researchers compared the results of their software, called QUARTZ (an inventive acronym derived from the phrase “QUantitative Analysis of Retinal Vessel Topology and siZe”) with 10-year risk predictions produced by the standard Framingham Risk Score test (FRS). They found that the two methods had “comparable performance.”
The big challenge, Keane says, is taking this kind of work from “code to clinic.” Who can turn this kind of research into a diagnostic tool, he wonders; would it be the UK National Health Service (NHS) or a company emerging from the university? And what level of performance will regulators need before approving the use of the software? “At what point do we say ‘let’s put a fork in it, we’re done’, and turn it into a commercial product?”
