Computational fluid dynamics framework for large-scale simulation in pediatric cardiology

There is a high demand for patient specific cardiovascular therapeutics, especially in pediatric cardiology which is confronted with complex and rather unique congenital diseases. Current predictors for disease severity and treatment selection have been proven to be suboptimal creating profound burden of premature morbidity and mortality. Over the past decade, the influence of blood hemodynamics has become increasingly acknowledged, especially in the context of congenital diseases of the aortic arch. MRI-based 2D and 3D flow measurements are nowadays possible, although restricted by cumbersome acquisition protocols and limited acquisition resolution. Computational fluid dynamics (CFD) offers a valuable alternative that also enables treatment outcome prediction. However, the current methods rely on a sequence of complicated manual steps that lack the scalability required within clinical settings. We propose a computation framework for large-scale hemodynamics simulations in pediatric cardiology to aid diagnostic and therapy decision making in patients affected by congenital disease of the aortic valve (AV) and the aorta. Our method provides a deterministic and streamlined processing pipeline to perform CFD simulations based on patient-specific boundary conditions. Thus, blood flow simulations are performed using an embedded boundary method within a level-set formulation with boundary conditions provided by patient-specific anatomical and hemodynamical models. The capabilities of our framework are demonstrated by performing blood flow analysis on patients selected from an FDA-sponsored multicenter clinical trial.