Computational modeling of cerebral aneurysms

Abstract

Wall shear (WSS) plays an important role in the genesis, progression and rupture of cerebral aneurysms. However the measurement of WSS in vivo is still difficult. In this PhD project we present a computational approach to evaluate the WSS and other flow parameters. This approach couples the flow solvers of different dimensions (0D, 1D and 3D) into a computational pipeline so that a balance is achieved between the modelling accuracy and computational efficiency. Furthermore, the vascular models are constructed directly from medical scanning images, from the large vasculature level to the small local vessel level. Using the 0D/1D part of the pipeline we at first simulate the blood flow in a patient-specific cerebral vasculature which contains an arterial tree and a venous tree. The computed velocity waveforms are supplied to a 3D flow solver for two intracranial aneurysms as their boundary conditions so that the WSS variation in the two aneurysms over a cardiac cycle can be realistically modelled. The computational results for the arterial and venous threes are validated via ultrasonography, and 3D flow models via Digital Subtraction Angiography (DSA) and phase contrast Magnetic Resonance Angiography (PC MRA). The computed WSS values are also compared with literature and our results are consistent with the published data of other research groups. The novel contributions of this project include: (1) a more compact numerical scheme (McCormack) to solve the expanded 1D flow systems (also applied to venous trees) is developed from scratch; (2) a haemodynamic model is developed for a patient-specific cerebral venous tree based on (1); and (3) the 3D flow model in an intracranial aneurysm is validated directly from DSA. This body of research has been presented as three papers, which have either been published or submitted to international peer-reviewed journals.