Verily Life Sciences, Alphabet Inc research organization developed deep-learning algorithms to predict cardiovascular disorder by analyzing visual images of retina
The study was published in Nature Biomedical Engineering on February 18, 2018. Google’s algorithm accurately detected high blood pressure level leading to cardiovascular risk by analyzing retina patterns of eye. This new development has applications in healthcare industry, which is based on artificial intelligence. The retinal images can also be used to predict blood pressure, age, gender, cardiovascular health history, and smoking status of person.
“Using deep-learning algorithms trained on data from 284,335 patients, we were able to predict cardiovascular risk factors from retinal images with surprisingly high accuracy for patients from two independent datasets of 12,026 and 999 patients,” said Lily Peng, Google Brain Team Product Manager. As a part of study, the team has collected data from around 48,101 patients from the UK Biobank database and 236,234 patients from EyePACS database. Retinal images offer real-time detection of health-associated risks, which is cost-effective as well as non-invasive. The algorithms are programmed to predict onset of major cardiovascular events such as heart attack within five years.
“Our algorithm could pick out the patient who had the CV event 70 percent of the time. This performance approaches the accuracy of other CV risk calculators that require a blood draw to measure cholesterol,” wrote Peng. Furthermore, to check the accuracy of the system, researchers used attention maps to observe algorithm predictions. They examined the factor focused by the system, such as whether blood vessels are observed to predict age, smoking status, and blood pressure. Peng believes that opening the black box to explain way of prediction will give doctors more confidence in the algorithm. However, further research is required before using this technique in clinical settings.