Pan-Can Analysis of Electrocardiogram Waveform Patterns Using Artificial Intelligence
Electrocardiogram (ECG) are routine medical images that capture information about cardiac electromechanical health. ECGs may be leveraged to inform treatment decisions and/or surveillance for cardiac complications after cancer chemotherapy. In the context of this proposal, we aim to develop artificial intelligence leveraged ECG interpretation (AI-ECG) as a predictive biomarker for intensive cancer therapy. Leveraging a large institutional repository of ECG images, we will develop cancer-specific models to predict toxicity from treatment and test these models in a multicenter setting. Further, we aim to evaluate AI-ECG as a diagnostic tool for infiltrative cardiac disorders such as cardiac amyloidosis. The project involves collaboration with multiple centers.