Researchers at Cedars-Sinai have created an artificial intelligence-enabled tool that can more easily predict whether a person will have a heart attack.
The tool, in The Lancet Digital Healthaccurately predicted which patients would develop a heart attack within five years, based on the amount and composition of plaque in the arteries that supply the heart.
Plaque buildup can cause arteries to narrow, making it harder for blood to reach the heart, increasing the likelihood of a heart attack. A medical test called coronary computed tomography angiography (CTA) takes 3D images of the heart and arteries and allows doctors to estimate how narrow a patient’s arteries are. However, until now, there has not been a simple, automated and fast method to measure visible plaques in CTA images.
“Coronary plaques are often not measured because there isn’t a fully automated method,” said Damini Dey, PhD, director of the Quantitative Image Analysis Laboratory at the Cedars-Sinai Institute for Biomedical Imaging Research and senior author of the study. It takes at least 25 to 30 minutes, but now we can use this procedure to quantify plaques from CTA images in 5 to 6 seconds.”
Dey and colleagues analyzed CTA images from 1,196 individuals who underwent coronary CTA at 11 sites in Australia, Germany, Japan, Scotland, and the United States. The researchers trained an AI algorithm to measure plaque by letting it learn from coronary CTA images of 921 people, which had been analyzed by trained doctors.
The algorithm first outlines the coronary arteries in 3D images and then identifies blood and plaque deposits within the coronary arteries. The researchers found that the tool’s measurements corresponded to the amount of plaque in the coronary CTA. They also matched the results to images from two invasive tests considered highly accurate in assessing coronary plaque and stenosis: intravascular ultrasound and catheter-based coronary angiography.
In the end, the researchers found that the AI algorithm’s measurements based on the CTA images accurately predicted heart attack risk over a five-year period in 1,611 people who participated in a multicenter trial called the SCOT-HEART trial.
“More research is needed, but it’s possible that we’ll be able to predict if and how long a person is likely to have a heart attack based on the number and composition of plaques imaged using this standard test,” said Dey, who is also a co-author of Cedars-Sinai Biosciences. Professor of Medical Sciences.
Dey and colleagues are continuing to study how their AI algorithm can quantify plaque deposition in patients undergoing coronary CTA.
Funding: This research was funded by the National Heart, Lung and Blood Institute under award number 1R01HL148787-01A1.