AI has Learned to Find Hidden Heart Risks In CT Scans

By examining CT scans, researchers from Mass General Brigham and the U.S. Department of Veterans Affairs have created an artificial intelligence (AI) tool called AI-CAC that can identify hidden heart disease risk. With almost 90% accuracy, it can detect coronary artery calcium (CAC), a significant factor in heart attack risk and long-term mortality. After being trained on a sizable dataset of CT scans from veterans, the tool discovered a strong correlation between high CAC scores and an elevated risk of death over a ten-year period. AI-CAC may assist physicians in identifying at-risk patients early and directing care toward prevention by repurposing existing scans that were initially obtained for other purposes, such as lung cancer screening.