Reduced order coronary hemodynamics modeling

Image-based computational fluid dynamic simulations have been widely employed to characterize the global and local blood flow features in the cardiovascular system. Although these 3D simulations enable comprehensive analysis of hemodynamics, they come at a high computational cost. Reduced-order modeling (ROM) of blood flow provides the ability to study hemodynamics of large cardiovascular networks with comparatively negligible computational cost. This is useful when multiple simulations are desirable, such as for data assimilation, optimization and uncertainty analysis. The goal of this study is to develop ROM to capture vessel-level pressure and flow dynamics in the coronary arteries. Here we developed a fully lumped parameter model to predict pressure and flow throughout patient-specific models of the coronary arteries. We test the ability of our ROM to compute fractional flow reserve (FFR) in stenotic coronary arteries, since such information is highly relevant to clinical diagnosis of coronary disease. Finally, we evaluated the robustness of the computed FFR by quantifying the uncertainties in our modeling. Our approach demonstrated strong agreement with fully 3D simulations for several patient-specific models including arteries with significant stenosis.