Our eyeballs apparently contain information that could revolutionize cardiovascular medicine.
Artificial intelligence software developed by Google in conjunction with its biotech subsidiary company Verily can scan retinal images to predict heart disease at nearly the same accuracy rate as a traditional blood test, United Press International reports. 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.
From the retinal images, Google's AI can determine within impressive degrees of accuracy a patient's age, gender, blood pressure, and smoking status, as well as even the past occurrence of major cardiovascular events, The Verge explains. The program taught itself how to analyze eyeballs after using machine learning techniques to pore over more than 284,000 retinal images; while studying, the AI used what UPI describes as a visual "heatmap" to learn which parts of the eye's anatomy contained certain predictive factors. The AI eventually learned, for example, that to analyze a patient's blood pressure, it was prudent to examine the blood vessels in the eye.
To test its capabilities, researchers sicced the AI on two patient pools, totaling more than 13,000 patients. The AI made correct predictions on the future risk of heart disease in 70 percent of cases — nearly the same accuracy rate as the blood-test method doctors traditionally use, which has a 72 percent accuracy rate.
Harlan M. Krumholz, the director of Yale's Center for Outcomes Research and Evaluation, predicted that the findings of Google's AI show that machine learning and artificial intelligence will "more precisely hone our understanding of disease and individuals," helping physicians "understand these processes and diagnoses in ways that we haven't been able to before." Read the full study here.