Researchers at Cedars-Sinai say they can now predict a person’s chances for a heart attack five years in advance with a new AI solution designed to measure the amount and assess the composition of plaque buildup.
Whereas standard testing utilizes computed tomography angiography (CTA) to estimate a patient’s chance for a heart attack based on the narrowing of their arteries, coronary plaques are often excluded from this estimate because there is no simple, automated way to measure them. And manually measuring them takes experts at least 25 to 30 minutes.
The algorithm at Cedars-Sinai outlines the coronary arteries in the 3D images produced with CTA, and identifies the blood and quantifies plaque deposits within the coronary arteries, in five to six seconds. "Total plaque volume and high-risk plaque volume can predict risk of heart attack. Plaque measured by the AI tool can allow more precise identification of patients for whom appropriate treatment could be prescribed, to reduce their risk of heart attack," Damini Dey, director of the quantitative image analysis lab in the Biomedical Imaging Research Institute at Cedars-Sinai and senior author of the study, told HCB News.
Dey says more studies are needed but that it may be possible to predict if and how soon a person is likely to have a heart attack through plaque analysis. She and her colleagues trained the algorithm on coronary CTA scans from 921 people at 11 sites in Australia, Germany, Japan, Scotland and the U.S. Scans were already assessed by trained doctors. The solution’s measurements corresponded with plaque amounts seen in coronary arteries. They were also compared to images taken with intravascular ultrasound and catheter-based coronary angiography, both of which are considered to be highly accurate in assessing coronary artery plaque and narrowing.
They then applied it to CTA images for 1,611 people enrolled in the multicenter SCOT-HEART trial and were accurately able to predict the risk for heart attack within all of them over the next five years.
They plan to continue to study how well the algorithm quantifies plaque deposits in such cases.
The study was funded by the National Heart, Lung, and Blood Institute.
The findings were published in The Lancet Digital Health.
Cedars Sinai previously released another study in January that said that combining results from coronary 18F-NaF uptake on PET and quantitative coronary plaque characteristics on CT angiography and assessing them with AI enhances
heart attack prediction accuracy.
The authors of that study assessed nearly 300 patients with hybrid coronary 18F-NaF PET and contrast CT coronary angiography and then calculated a joint score for heart attack risk with AI. The technique showed substantial improvement in prediction of heart attack over clinical data alone.