New Google AI Can Detect Cardiovascular Diseases Using Retinal Scans


A new algorithm developed by Google and its sibling-company Verily Life Sciences can assess a person's risk of cardiovascular diseases. The AI analyzes retinal images.

The AI scans images of a person's retina and accurately predicts the risk of major cardiac events such as heart attack or stroke. The robotic AI uses standard retinal scans to analyze the blood vessels in the retina.

Deep Learning Algorithm

The study led by researchers from Google, Verily Life Sciences, and the Stanford School of Medicine used deep-learning algorithms to extract new knowledge from retinal images.

The AI scanned the retinal images and identified cardiovascular risk factors based on age, gender, blood pressure, and smoking status, among others.

"Our algorithm used the entire image to quantify the association between the image and the risk of heart attack or stroke," says Lily Peng MD PhD, product manager of Google Brain Team and one of the study's lead authors.

The deep-learning models were configurated to study the anatomical features of each retinal scan to generate specific prediction.

Data from 284,335 patients were studied by the AI algorithm and validated the analyses on two independent datasets of 12,026 and 999 patients.

Looking Through The Eyes

Retinal scan or retinal imaging is already being for used for the diagnosis and management of age-related eye diseases and diabetic retinopathy. It could also be an important source of biomarkers for the diagnosis of chronic and long-term illnesses. The Google AI only scans images of the retina to make the prediction.

If given a retinal image of a patient who experienced a major cardiovascular event and the retinal image of another patient who did not, the algorithm could pick out the patient who had a cardiovascular event with 70 percent accuracy.

"While doctors can typically distinguish between the retinal images of patients with severe high blood pressure and normal patients, our algorithm could go further to predict the systolic blood pressure within 11 mmHg on average for patients overall, including those with and without high blood pressure," Peng said.

The AI's approach is comparable to the accuracy of other forms of cardiovascular risk assessment such as blood tests.

The technique that uses deep neural networks generated a heatmap or graphical representation of data that reveals which pixels in an image are the most important in predicting risk factors.

Each cardiovascular risk factor uses a distinct pattern, such as the blood vessels for blood pressure and optic discs for other predictions.

"Pattern recognition and making use of images is one of the best areas for AI right now," says Harlan M. Krumholz, a professor of medicine and director of Yale's Center for Outcomes Research and Evaluation.

Google's new research could represent a new method of scientific discovery, including cancer research. The use of photographs, scans, and sensors will improve physical examinations of patients and determining specific treatments.

The study "Prediction of cardiovascular risk factors from retinal fundus photographs via deep learning" is published in the journal Nature Biomedical Engineering.

ⓒ 2018 All rights reserved. Do not reproduce without permission.
Real Time Analytics