RESUMEN
Wall Shear Stress (WSS) has been demonstrated to be a biomarker of the development of atherosclerosis. In vivo assessment of WSS is still challenging, but 4D Flow MRI represents a promising tool to provide 3D velocity data from which WSS can be calculated. In this study, a system based on Laser Doppler Velocimetry (LDV) was developed to validate new improvements of 4D Flow MRI acquisitions and derived WSS computing. A hydraulic circuit was manufactured to allow both 4D Flow MRI and LDV velocity measurements. WSS profiles were calculated with one 2D and one 3D method. Results indicated an excellent agreement between MRI and LDV velocity data, and thus the set-up enabled the evaluation of the improved performances of 3D with respect to the 2D-WSS computation method. To provide a concrete example of the efficacy of this method, the influence of the spatial resolution of MRI data on derived 3D-WSS profiles was investigated. This investigation showed that, with acquisition times compatible with standard clinical conditions, a refined MRI resolution does not improve WSS assessment, if the impact of noise is unreduced. This study represents a reliable basis to validate with LDV WSS calculation methods based on 4D Flow MRI.
RESUMEN
Wall shear stress (WSS) is a relevant hemodynamic indicator of the local stress applied on the endothelium surface. More specifically, its spatiotemporal distribution reveals crucial in the evolution of many pathologies such as aneurysm, stenosis, and atherosclerosis. This paper introduces a new solution, called PaLMA, to quantify the WSS from 4D Flow MRI data. It relies on a two-step local parametric model, to accurately describe the vessel wall and the velocity-vector field in the neighborhood of a given point of interest. Extensive validations have been performed on synthetic 4D Flow MRI data, including four datasets generated from patient specific computational fluid dynamics simulations on carotids. The validation tests are focused on the impact of the noise component, of the resolution level, and of the segmentation accuracy concerning the vessel position in the context of complex flow patterns. In simulated cases aimed to reproduce clinical acquisition conditions, the WSS quantification performance reached by PaLMA is significantly higher (with a gain in RMSE of 12 to 27%) than the reference one obtained using the smoothing B-spline method proposed by Potters et al. (2015) method, while the computation time is equivalent for both WSS quantification methods.