Google’s AI could detect heart diseases from retina scans

Have you ever wondered why doctors examine your eyes even when the ailment is seemingly somewhere else? One’s eyes, or rather the retina, doctors say, can reveal a lot about your general well being.

It turns out what doctors can, Google can do better.

On Tuesday, Google published an official blog saying its deep learning algorithm can accurately predict a person’s cardiovascular health by evaluating retinal images. In other words, Google’s AI can tell if your heart is in a good state.

The discovery was put forward by scientists Ryan Poplin, Avinash V. Varadarajan, Katy Blumer, Yun Liu, Michael V. McConnell, and Greg S. Corrado in a research paper titled ‘Prediction of cardiovascular risk factors from retinal fundus photographs via deep learning’, published in Nature Biomedical Engineering. The scientists are part of Verily Life Sciences, Alphabet's R&D team dedicated to innovation in health. Alphabet is Google’s parent company.

In the blog, Google says by using deep learning algorithms trained on data from 284,335 patients, the team was able to predict cardiovascular risk factors from retinal images with surprisingly high accuracy for patients from two independent data-sets of 12,026 and 999 patients. The algorithm could distinguish the retinal images of a smoker from that of a non-smoker 71% of the time. While doctors can typically distinguish between the retinal images of patients with severe high blood pressure and normal patients, Google says its algorithm could go further to predict the systolic blood pressure within 11 millimeters of mercury (mmHg) on average for patients overall, including those with and without high blood pressure.

There’s more. Google says the algorithm could also fairly accurately predict the risk of a cardiovascular event such as a stroke or heart attack directly.

Google says its approach uses deep learning to draw connections between changes in the human anatomy and disease, similar to how doctors learn to associate signs and symptoms with the diagnosis of a new disease. This, the company says, could help scientists generate more targeted hypotheses and drive a wide range of future research.

The discovery is crucial as it opens the possibility of early detection of heart diseases without having to undergo a series of tests, or even a visit to a doctor. Imagine a medically-approved retina scanner that is connected to the Internet—and thus Google’s AI brain—updating you every morning on the state of your heart. It could be something like a blood sugar monitor millions of patients use everyday, albeit with a far deeper info about your heart. Of course, this, for now, is only a possibility—but at least there is one now.

Google isn’t the only company to research on deploying AI to assess health or in assisting doctors with diagnoses. IBM, with its Watson AI, and a clutch of ambitious startups in Silicon Valley and Israel are working on various aspects of AI in healthcare.