Construction of a four-chambered, fluid-structure interaction model of the heart with validation studies of physiologic left heart performance

In this work we introduce a fluid-structure interaction (FSI) model of the whole heart with fully three-dimensional descriptions of the valves and subvalvular complexes. The function of the heart has been a subject of research for millenia from early dissection studies of anatomy to modern disease and device investigations that capture multiple scales and organ systems.Large animal models and imaging modalities are the pinnacle of current organ scale research efforts with animal models being used for in vivo research that cannot be captured by imaging studies on humans.However, there are inherent costs, limitations, and risks associated with imaging studies and animal models.Computational modeling provides and opportunity to augment cardiac modeling studies by querying aspects unavailable to imaging and animal models while reducing some of the inherent risk and cost of working with living subjects. FSI methods, such as the immersed finite element/difference method, can leverage the inclusion of both fluid dynamics and structural mechanics to capture the complex interplay of the blood and tissues of the heart. Previous FSI whole heart models have included two-dimensional descriptions of the valves which are less suited for studies that intend on understanding the stresses and strains of the leaflets during the cardiac cycle in a diseased state or under intervention.The proposed model provides a platform for atrioventricular studies focused on valves such as mitral regurgitation and the various treatment methods.Through a number of simulations we demonstrate that the model can generate physiologic pressure, volume, and flow rates in the left heart while also dynamically responding to stimuli such as by capturing the Frank-Starling mechanism.The future of physics informed heart modeling is set to expand as the U.S. Food and Drug Administration released guidelines for the inclusion of computational models in clinical device studies in November of 2023, and we are excited for the potential applications of the model to this effort.