Can Google Predict Cardiovascular Risk from an Eye Scan?
Wednesday, February 28 2018 | 00 h 00 min | Vision Science
Verily Life Sciences, formerly Google Life Sciences, say they have developed a method to assess cardiovascular risk using machine learning.
Details about the software published in Biomedical Engineering say that the deep-learning models can determine a host of cardiovascular risk factors for an individual; including age, blood pressure and whether the individual smokes, with nothing more than a retinal fundus image.
Google’s AI was able to tell with 70% accuracy which patients had recently had strokes by analyzing the appearance and diameter of the retinal blood vessels.
Though use of artificial intelligence to make diagnoses is still “decades” away, Google is using studies like Project Baseline, where they exhaustively document the medical status of 10,000 volunteers, to determine how to apply Big Data and deep-learning algorithms to health care diagnoses.