Heart Health

AI Predicts Fatal Heart Attacks 89% Accurately Using Hidden MRI Patterns

Johns Hopkins researchers developed an AI system called MAARS that predicts sudden cardiac death with 89% accuracy by analyzing heart MRI scans and medical records. The system identifies hidden scarring patterns in patients with hypertrophic cardiomyopathy, a common inherited heart condition affecting 1 in 200-500 people worldwide. Current clinical guidelines used by doctors are only 50% accurate at identifying high-risk patients. MAARS was particularly effective for patients aged 40-60, achieving 93% accuracy in this highest-risk group. The breakthrough could save lives by better identifying who needs protective defibrillators while sparing others from unnecessary procedures.