A personalized multiscale model of ventricular cardiac mechanics*

Numerical simulations of the heart are invaluable for investigating cardiovascular disease mechanisms and treatments. Modeling the circulatory system in these simulations is essential to capture physiological cardiac behavior. To reduce computational cost, blood circulation is typically represented with a zero-dimensional (0D) lumped-parameter network (LPN). Previously, we described a novel scheme to couple three-dimensional (3D) cardiac mechanics models with 0D circulatory models using an idealized left ventricle. Here, we extend this coupling scheme to an image-derived biomechanical model of the ventricles, coupled to a closed-loop LPN. The model parameters are personalized to match clinical measurements, including phase-resolved (3D+t) CT images, cuff blood pressures, and electrocardiogram waveforms. We personalize the model by first tuning LPN parameters, then determining the stress-free configuration and passive material parameters of the myocardium, and finally optimizing the active contraction model parameters. Preliminary results show physiological cardiovascular behavior, including the atrioventricular plane motion and reasonable pressure-volume loops. The isovolumic phases of the cardiac cycle are captured without ad hoc treatment. This model can investigate congenital heart defects and surgical interventions and may be extended to four-chamber heart models to simulate whole-heart mechanics.