Abstract Personalized computational models of cardiac electrophysiology (CEP), referred to as cardiac digital twins (CDTs), are considered a key technology for realizing precision cardiology by tailoring therapies to individual patients. However, the high computational demands of physically detailed simulations of CEP in human hearts prohibit broader adoption of CDTs in practical applications. Here, we aim to address this challenge with a lightweight Physics-Informed Eikonal (PIE) model that replicates the full spectrum of biophysical phenomena embedded within high fidelity CEP models, including the initiation and maintenance mechanisms of arrhythmias. The PIE model is able to simulate CEP in the human heart along with clinically observable electro-cardiograms in near real-time, rendering like-for-like comparisons to clinical data feasible. Validity of the PIE model is shown by comparing against a high-fidelity gold-standard CEP model in complex arrhythmia simulations, including a scar-mediated ventricular tachycardia in a human whole-heart model. The PIE model delivers near-real time performance and retains full physical fidelity as required for the calibration of CDTs, thus providing a fundamental core technology for leveraging CDTs in industrial and clinical applications.
Plank et al. (Fri,) studied this question.