Researchers at Cedars-Sinai Medical Center have developed an advanced AI tool that detects heart attack risk up to five years in advance. This innovation comes at a critical time, as cardiovascular diseases remain the world's leading cause of death.
According to the World Health Organization, cardiovascular diseases claim nearly 18 million lives annually—31% of all global deaths, surpassing any other cause. Ongoing research aims to sharpen risk prediction. In February 2022, University of Leeds scientists unveiled an AI that analyzes retinal scans for heart attack risk.
Now, in a forthcoming study in The Lancet Digital Health, experts from the Cedars-Sinai Biomedical Imaging Research Institute (BIRI) in Los Angeles introduce their own AI model. Unlike retinal analysis, this system examines CT angiography results—a specialized imaging technique that visualizes otherwise invisible blood vessels using contrast-enhanced X-rays.
The AI predicts heart attack risk five years ahead with high accuracy (r=0.922, p<0.0001). Trained on CT angiography images from 921 patients, it expertly identifies atheroma plaques—buildups of fat, cholesterol, fibrous tissue, or calcium that narrow arteries and restrict blood flow, heightening infarction risk.
The team established plaque thresholds to classify patients into low- or high-risk groups for events within five years. Results were rigorously validated on 1,611 patients from the 2018 Scot-Heart study. While larger trials are needed to fully affirm its clinical use, the researchers are confident this AI, integrated with CT angiography, offers highly reliable risk prediction with minimal error.
