4 - Data-driven reduction of cardiac models

Abstract

A common challenge in physiological modeling is the trade-off between the fidelity of the model and its complexity. Reliable models of cardiac physiology and pathology could have a tremendous value in supporting clinical decisions, by enabling advanced diagnosis, precise risk stratification and personalized therapy planning. However, for this technology to express its potential it is crucial that the data processing time is compatible with the fast-paced clinical workflow. In the ideal case, detailed and patient-specific physics phenomena should be reproduced and simulated with processing times enabling real-time interaction with the model solution. In many cases, real-time processing cannot be achieved if not at the expense of the descriptive power of the model. This chapter describes data-driven approaches for the reduction of physiological models, with the aim of retaining the ability of the model to describe complex physical phenomena, while at the same time drastically reducing the computational complexity.