Comparison of patient-specific inlet boundary conditions in the numerical modelling of blood flow in abdominal aortic aneurysm disease.

Three inlet boundary condition datasets were derived from phase-contrast MRI: (i) centre line velocity data converted to two-dimensional (2D) velocity profile using Womersley equations (Womersley), (ii) 2D velocity profile with one axial component of velocity (1CV), (iii) 2D velocity profile with three components of velocity (3CV). Computational fluid dynamics was performed using a rigid wall approach with geometry data extracted from the computed tomography dataset. Helical flow was present in the 1CV and 3CV simulations, with more complex patterns for the 3CV case. The Womersley method produced simplified flow patterns with an absence of helical flow. Mean values of quantitative indices (helical flow index, mean wall shear stress, oscillatory index) were compared with the 3CV inlet data. These were lower for both the Womersley inlet data (28%, 71%, 56%) and the 1CV inlet data (9%, 24%, 69%). It was concluded that inlet methods based on centre line velocity, such as might be obtained from Doppler ultrasound, lead to significantly simplified abdominal aortic aneurysm haemodynamics and thus are not recommended. Single velocity component (axial) data from MRI might suffice when general flow characteristics and spatial wall shear stress are required. Ideally 2D MRI velocity profiles with 3-velocity component data are preferred to fully account for helical flow.