Hemodynamic wall shear stress in models of atherosclerotic plaques using phase contrast magnetic resonance velocimetry and computational fluid dynamics

Arterial wall shear stress (WSS) is thought to be a factor that determines locations in the vasculature where atherosclerotic plaques are formed. Providing patient specific WSS data may increase the potential for early detection and treatment before serious clinical complications occur. Phase contrast magnetic resonance imaging (PC-MRI) offers a non-invasive method for determining blood velocity profiles. However, the limited resolution of this technique restricts the accuracy of the near-wall velocity data that are needed to calculate WSS. The purpose of this research was to determine if improved WSS calculations from current magnetic resonance imaging technologies could be developed.; PC-MRI data were obtained for anatomically scaled phantoms representing blood vessels with and without symmetric stenoses under average and peak steady flow conditions. WSS values were calculated by two methods: (1) directly from PC-MRI velocity profiles and (2) from computational fluid dynamic (CFD) simulations with MRI defined geometries and inlet boundary conditions. The accuracy of both methods was determined by comparing the results to gold standard WSS data derived from CFD simulations using ideal geometries and boundary conditions.; Both methodologies resulted in data with large variations in WSS values between adjacent axial and angular locations. The direct calculation of WSS from PC-MRI data yielded large underestimations of the maximum WSS values present within the stenosis geometries. These errors were related to the low resolution of the velocity data obtained from PC-MRI as well as the inability to accurately detect the phantom wall location due to partial volume errors and low signal-to-noise ratios. The CFD simulations yielded moderate underestimations of WSS within the stenoses. The MRI derived computational geometries and the underestimation of the flow rates used for the inlet boundary conditions were found to cause the greatest errors for the average and peak flow rate simulations, respectively.; It was determined that WSS values derived from the CFD simulations are more accurate compared to the direct calculation of WSS from PC-MRI data. To improve upon these CFD methodologies, more studies are required to optimize geometry smoothing and reconstruction in order to reduce WSS errors while maintaining physiologically relevant geometries.