Google AI Can Predict Heart Disease With Eye Scans

Imagine going to the doctor’s office, getting a simple eye scan, and finding out if you’re at risk of having a heart attack. Google scientists and the research company Verily have developed an algorithm that can predict a person’s risk for heart disease just by looking at the individual’s eye.

The company’s software analyzes scans that output data such as a patient’s age, blood pressure, and whether he or she is a smoker. This information is then used to determine if the person is at risk for experiencing a cardiac event. The process is as effective as current leading methods reports The Verge.

The screening does not require a blood test and may be quicker and easier for doctors to assess a person’s cardiovascular risk. However, more testing is required before it’s put into mainstream use.

Google and Verily scientists analyzed medical data and eye scans from nearly 300,000 patients to train its algorithm. They searched for patterns that would indicate cardiovascular risk. The rear interior wall of the eye, known as the fundus, has blood vessels that are a map of a person’s overall health. Using cameras and microscopes, doctors can look deep into the eyes and determine blood pressure, age, and smoking status, all of which are related to cardiovascular health.

By looking at retinal scans, Google’s algorithm was able to tell the difference between two patients, one who had a cardiovascular event within five years and one who did not, 70 percent of the time. This percentage is only slightly less accurate as the current method of predicting cardiovascular risk, SCORE, which necessitates a blood test.

Does this mean artificial intelligence will soon replace humans in the doctor’s office? According to Luke Oakden-Rayner, a medical researcher at the University of Adelaide who specializes in machine learning analysis, technology such as this will simply enhance a medical practitioner’s work. He told The Verge: “They’re taking data that’s been captured for one clinical reason and getting more out of it than we currently do. Rather than replacing doctors, it’s trying to extend what we can actually do.”

The paper, “Prediction of cardiovascular risk factors from retinal fundus photographs via deep learning,” was published in the Nature journal Biomedical Engineering.

This is a test