협착병변을 가진 관상동맥에 대한 혈류역학적 지표의 불확실성 정량화 및 민감도 분석

Abstract

In clinical, fractional flow reserve (FFR) and wall shear stress (WSS) are useful for assessing the severity of atherosclerosis. The computational fluid dynamics (CFD) method to access these indices is widely used, however, the uncertainty of the CFD prior parameters made it is not straightforward to predict, there are many sources of uncertainty, as the accuracy in vivo measurement. In this study, we conducted uncertainty quantification (UQ) and sensitivity analysis (SA) for an idealized stenotic coronary. By using the non-intrusive Point-Collocation method under helping of Chaospy software combined with a 3D-0D coupled P2P1 finite element solver. We applied these methods in terms of ideal conditions, with the assumption that blood vessels are rigid walls and blood flow is a Newtonian fluid. Two different models were considered: single stenosed vessel and tandem stenosed vessel with mean values for UQ and SA are taken from reliable sources. The UQ and SA results were expressed into expected and standard deviation values of the quantities of interest. These results show that UQ and SA are necessary when a small error in uncertain input (5% or 10%) can lead to a big change in output (can up to 20% in centerline velocity of the tandem stenosed vessel case), make the wrong decision. This study helps us understand the effect of different inputs on the hemodynamic indices of the stenotic coronary.