Google builds AI that can predict heart disease

Ray Weaver
February 21, 2018

After analyzing data from over a quarter million patients, the neural network can predict the patient's age (within a 4-year range), gender, smoking status, blood pressure, body mass index, and risk of cardiovascular disease. The findings, published Monday in the journal Nature Biomedical Engineering, explain that Google's AI makes its predictions by examining images of the back of a patient's eye in order to develop a profile of the patient, including several characteristics that could determine cardiovascular risk. And while doctors can usually distinguish between the retinal images of patients with severe high blood pressure and those without, the algorithm could go further, and predict the systolic blood pressure for all patients.

The company's latest approach with AI predicting heart disease could build on doctors' current abilities by providing a tool that people could one day use to quickly and easily screen themselves for health risks that can contribute to heart disease, which is the leading cause of death worldwide.

However, despite the success of their experiment, Google said that more research has to be done.

"They're taking data that's been captured for one clinical reason and getting more out of it than we now do", Luke Oakden-Rayner, a medical researcher from the University of Adelaide, told The Verge.

The research, one of an increasing number of conceptual health-technology studies, was conducted by Google and Verily Life Sciences, a subsidiary of Google's parent Alphabet. In the testing phases, the algorithm was able to identify heart conditions 70% of the time, which is a slightly lower rate of success than the longer SCORE process which is correct around 72% of the time. "However, we don't precisely know in a particular individual how these factors add up, so in some patients, we may perform sophisticated tests ... to help better stratify an individual's risk for having a cardiovascular event such as a heart attack or stroke", declared study co-author Dr. Michael McConnell, a medical researcher at Verily.

"They're taking data that's been captured for one clinical reason and getting more out of it than we now do", he said. "But we need to validate".

Google's identification method is unique because the AI trained itself to find the pattern used to identify heart risk in the images. For example, the algorithm paid more attention to blood vessels for making predictions about blood pressure, as shown in the image above.

While Peng said that it might take years for their newly developed AI technology to be used in real-world medical applications, she's optimistic that artificial intelligence can be applied in many areas of scientific discovery.

Other reports by Insurance News

Discuss This Article