The method-detailed in a paper published on Monday in the Nature journal Biomedical Engineering-involves analyzing blood vessels in an area of the eye called the retinal fundus. This test can predict the risk of suffering a major cardiac disease such as heart attack.
The algorithm used scans of a patient's eye to build a picture of their general health, a technique which is already used in medical research.
The researchers trained deep learning algorithms on data from thousands of patients recorded in a massive United Kingdom study, which was used with retinal scans to produce a program that can identify risk factors from the scan information alone.
Krumholz, however, cautioned that an eye scan isn't ready to replace more conventional approaches. The algorithm had roughly the same accuracy as current leading methods.
The researchers concluded: "The opportunity to one day readily understand the health of a patient's blood vessels, key to cardiovascular health, with a simple retinal image could lower the barrier to engage in critical conversations on preventive measures to protect against a cardiovascular event". Medical professionals today look for signs of heart disease by using a device to inspect the retina or drawing the patient's blood.
Health outcomes such as hospitalisation, mortality and cause of death were also logged.
Scanning someone's eye to know if they'll have heart disease may sound like a odd and random thing to do, but there's sufficient research that backs this up. The green highlights are the parts the algorithm found most helpful in predicting blood pressure.
This kind of technological solution could produce fast, cheap and noninvasive tests that could be administered in a range of settings. The work needs to be validated and repeated on more people before it gains broader acceptance, several outside physicians said. Predicting the factors that put a person at risk of a heart attack or stroke was an offshoot of the original research. Observing and quantifying associations with medical images is a challenge because of the wide variety of features present in real images.
Further tests are required before this latest method can be used within a clinical setting.