Implementation of a patient-specific cardiac model

This chapter describes a possible implementation of a comprehensive model of heart function. The presentation follows the same organization of the previous chapter. We first illustrate how the anatomical model can be build starting from medical images, and enriched with semantic information based on shape analysis. We then discuss efficient computational techniques for the electrophysiology model, based on Lattice–Boltzmann method for the approximation of the monodomain problem and on a graph-based solution of the Eikonal equation. The computation of surface electrical potential and ECG signals on the torso is also addressed, in particular based on the boundary element method. For the solution of the electromechanics problem we consider an efficient GPU-based implementation of the TLED method, discussing how to couple the biomechanics solver with the electrophysiology solver through the active force term, computed as a function of the electrical potential. Heart function depends critically on the interaction between heart tissue and blood: we dedicate specific attention to the implementation of boundary conditions for the electromechanics problem, including the coupling with models of the valves, pericardium bag and circulatory system. Finally, a section is devoted to computational models for hemodynamics, focusing in particular on techniques to solve the coupled fluid-structure interaction problem in the cardiac chambers.