AI-Powered ECG Assessments May Improve Heart Transplant Rejection Warnings

Associate Professor Lior Jankelson and colleagues at NYU Langone Health developed an AI model combining electrocardiograms and blood tests to identify heart transplant rejection and potentially reduce unnecessary biopsies. Trained using more than 5,300 ECG readings from 389 transplant recipients, the model correctly identified 97% of patients without rejection in a 38-patient test group. "Our results highlight that electrocardiograms contain an abundance of physiological information," Jankelson said, noting their potential to improve detection and enable earlier treatment. The team plans to validate the approach in larger, multicenter studies.