Simulating Blood Flow in Left Atrium using a Personalized Multiscale Electromechanics Modeling Framework

We innovated a workflow to simulate blood flow in the left atrium (LA) by coupling it to a personalized multiscale electromechanics model. The workflow begins by extracting the patient's clinical data, including time-dependent computed tomographic (CT) images and electrocardiogram (ECG) data, and feeding it to a semi-automatic tuning framework for personalizing the multiscale electromechanics model. In the tuning process, we developed a novel inverse finite element analysis framework to determine myocardial material parameters (passive expansion and active contraction) by applying realistic boundary conditions and physiological pressures, so that the simulated deformation matches clinical data, including four-chamber min/max volumes, cuff-based pressures, and LA volumes throughout the cardiac cycle. We then perform multiscale fluid-structure interaction (FSI) simulations in the LA using validated stabilized finite element methods and patient-specific tissue parameters, thereby integrating biophysics-based myocyte activation, tissue contraction, and blood flow coupled to a 0D lumped parameter network (LPN)-based circulatory system model. This integrated model overcomes the limitations of simulating LA blood dynamics by imposing image-based wall motion, enabling us to simulate realistic tissue contraction patterns and blood flow under pathological conditions, such as atrial fibrillation, and predict the biomechanical response to treatment and remodeling.