PROJECT SUMMARY Congenital heart disease (CHD) affects one in one hundred babies and is the leading cause of infant mortality in the US1. Single ventricle (SV) physiology is among the most complex and high-risk CHD diagnoses, in which patients are born with only one functional pumping chamber in the heart. These patients are typically palliated with three open heart surgeries culminating with a Fontan procedure, subjecting patients to a lifetime of elevated venous pressures and high rates of morbidity and mortality. As many as half of Fontan patients degenerate into heart failure and required a transplant by the age of 25. However, for a subset of patients with borderline SV physiology, drastic improvements in outcomes can be achieved if a bi-ventricular circulation can be restored. Despite these advantages, current reconstruction procedures require “on the fly” surgical planning in which the surgeon must customize a baffle design for each individual patient in the operating room. Computational modeling is well positioned to address these needs by providing surgical teams with predictive simulations. However, current simulation capabilities are hindered by several key factors: 1) anatomic models are laborious to construct, 2) there is a current lack of data characterizing CHD hearts, including material properties, fiber orientations, and Purkinje system structure and these quantities are critically needed for accurate simulations, and 3) current solvers do not combine all the relevant physics for whole-heart simulations. We aim to address these needs by developing a pediatric cardiac simulator to support surgical planning in complex congenital heart disease and to deploy it in a prospective clinical study. To accomplish these goals, we propose the following three specific aims: 1) To Enable Rapid Patient Specific Model Construction of Topologically Unique CHD Hearts With Machine Learning Methods Based On Signed Distance Fields. 2) To Characterize Mechanics and Microstructure in CHD Hearts Using Ex Vivo MR Acquisition and Finite-element Based Inverse Modeling. 3) To Prospectively Demonstrate and Validate a Novel Multi-Physics Cardiac Solver for Biventricular Reconstructive Surgical Planning in CHD Patients. Our proposed study will tightly integrate image processing and machine learning, advanced experimental magnetic resonance image acquisition methods, and development of state-of-the-art multi-physics finite element solvers (combining fluid and solid mechanics, active contraction, valves, and electrophysiology) to address an immediate clinical need in a high-risk and understudied patient population. A primary goal is to demonstrate improvements in short-term clinical outcomes before the end of the project. This proposal brings together an interdisciplinary team comprising experts in computational modeling of cardiovascular biomechanics, advanced MRI methods, microstructural and mechanical tissue characterization, and pediatric cardiac surgery. O...