Scientists from Google and its health-tech subsidiary Verily have discovered a new way to assess a person’s risk of heart disease. Google says it was able to accurately predict which patient would experience a heart attack or other major cardiovascular event within five years 70 percent of the time. Google recently presented its findings in the online medical journal Nature Biomedical Engineering. The research was also shared before peer review last September.
Researchers at Google worked to apply artificial intelligence to predict the likelihood that a patient will suffer a heart attack or stroke. Google used models based on data from 284,335 patients. This information included eye scans as well as general medical data. Neural networks were used to mine the data for patterns and associate telltale signs in the eye scans with the metrics needed to predict cardiovascular risk.
The researchers found that by examining images of the patient’s retina, the software was able to predict a patient’s risk of suffering a major cardiac event. The algorithm had roughly the same accuracy as current leading methods. The results were validated on two independent data sets of 12,026 and 999 patients.
Medical researchers have previously shown a correlation between blood vessels in the retina and the risk of a patient having a major cardiovascular episode. The rear interior wall of the eye is full of blood vessels that reflect the body’s overall health. Medical professionals today look for signs of heart disease by using a device to inspect the retina or drawing the patient’s blood. Heart disease is currently the leading cause of death worldwide.
Lily Peng, a doctor and lead researcher on the project, says Google was surprised by the results. Predicting the factors that put a person at risk of a heart attack or stroke was an offshoot of the original research. The team was originally working on predicting eye disease. The exercise was then expanded to predicting whether the person was a smoker or what their blood pressure was. The research naturally progressed from there.
This kind of technological solution could produce fast, cheap and noninvasive tests that could be administered in a range of settings. In time, physicians might study such retinal images as part of routine health check-ups to help assess and manage patients’ health risks. Peng is optimistic that artificial intelligence can be applied to other areas of scientific discovery as well, like cancer research.
However, Google cautions that more research needs to be done. Observing and quantifying associations with medical images is a challenge because of the wide variety of features present in real images. The method will need to be tested more thoroughly before it can be used in a clinical setting.