Spatiotemporal modeling and optimization for personalized cardiac simulation

Abstract

Computational modeling of the heart has contributed tremendously in quantitatively understanding the cardiac functions, showing great potential to assist medical doctors in heart-disease diagnosis. However, cardiac simulation is generally subject to uncertainties and variabilities among different individuals. Traditional “one-size-fits-all” simulation is limited in providing individualized optimal diagnosis and treatment for patients with heart disease. Realizing the full potential of cardiac computational modeling in clinical practice requires effective and efficient model personalization. In this paper, we develop a spatiotemporal modeling and optimization framework for cardiac model calibration. The proposed calibration framework not only effectively quantifies the spatiotemporal discrepancy between the simulation model and the physical cardiac system, but also increases the computational efficiency in personalized modeling of cardiac electrophysiology. The model performance is validated and evaluated in the 3D cardiac simulation. Numerical experiments demonstrate that the proposed framework significantly outperforms traditional approaches in calibrating the cardiac simulation.