Blood flow in cerebral aneurysms: comparison of phase contrast magnetic resonance and computational fluid dynamics--preliminary experience.
Academic Article
Overview
abstract
PURPOSE: Computational fluid dynamics (CFD) simulations are increasingly used to model cerebral aneurysm hemodynamics. We investigated the capability of phase contrast magnetic resonance imaging (pcMRI), guided by specialized software for optimal slice definition (NOVA, Vassol Inc.) as a non-invasive method to measure intra-aneurysmal blood flow patterns in-vivo. In a novel approach, these blood flow patterns measured with pcMRI were qualitatively compared to the ones calculated with CFD. MATERIALS AND METHODS: The volumetric inflow rates into three unruptured cerebral aneurysms and the temporal variations of the intra-aneurysmal blood flow patterns were recorded with pcMRI. Transient CFD simulations were performed on geometric models of these aneurysms derived from 3D digital subtraction angiograms. Calculated intra-aneurysmal blood flow patterns were compared at the times of maximum and minimum arterial inflow to the ones measured with pcMRI and the temporal variations of these patterns during the cardiac cycle were investigated. RESULTS: In all three aneurysms, the main features of intra-aneurysmal flow patterns obtained with pcMRI consisted of areas with positive velocities components and areas with negative velocities components. The measured velocities ranged from approx. +/- 60 to +/- 100 cm/sec. Comparison with calculated CFD simulations showed good correlation with regard to the spatial distribution of these areas, while differences in calculated magnitudes of velocities were found. CONCLUSION: CFD simulations using inflow boundary conditions measured with pcMRI yield main features of intra-aneurysmal velocity patterns corresponding to intra-aneurysmal measurements performed with pcMRI. Thus, pcMRI may become a valuable complementary technique to CFD simulations to obtain in-vivo reference data for the study of aneurysmal hemodynamics. More data is needed to compare and fully explore the capabilities of both methods.
