Computational modelling and haemodynamic investigation of intracranial aneurysms before and after flow-diversion treatment

Haemodynamic information is believed to be one of the most crucial factors affecting the initiation, development, growth, and rupture of intracranial aneurysms (IAs). Many studies of haemodynamic simulation contribute to the understanding of aneurysmal flow dynamics; however, clarification and justification of the validity of simulation results inconclusive remains. Besides, for most simulations of flow-diversion effect using a model flow-diverting (FD) stent, the properties of model FD stents failed to be calibrated to match the represented device, which would lead to inaccurate predictions of flow-diversion efficacy. Thus, this study aims to discover the factors that contribute to a more accurate and comprehensive simulation of aneurysmal haemodynamics and its flow-diversion treatment. Two main aspects were investigated in this thesis : 1) evaluation of the accuracy of computational fluid dynamics (CFD) predictions of aneurysmal haemodynamics before and after FD stent treatment, by comparing to in vivo and in vitro observation s; and 2 ) investigation of the practicality of using porous medium (PM) stent models, with calibrati on of stent model parameters to represent the commercially available FD stents, by derivation of parameters like permeability (k) that account for the flow resistance induced by the model stent. To meet these needs, haemodynamic investigations of IAs and their flow-diversion treatments have been performed using different approaches. In Chapter 3 and 4, the validity of CFD predictions of aneurysmal haemodynamics was evaluated, by comparing the resolved velocity field in patient-specific aneurysms with experimental methods, such as particle image velocimetry (PIV) and phase-contrast magnetic resonance imaging (PCMRI). By comparing against a physical stent model, the PM model stent was found to be a practical tool to assist flow-iversion simulation. In Chapter 5 and 6, the PM model thickness range that would help to retain the simulation benefits without compromising the accuracy was first confirmed; then PM model FD stents were respectively calibrated to reflect the flow resistance created by several treatment modes using commercially available FD stents. Overall, CFD prediction is proved to be able to accurately resolve the aneurysmal flow dynamics , and the investigation of calibrated PM stent modelling provides an individualised method that is more efficient than CFD simulations with a fully-resolved stent model yet retains simulation accuracy. These studies substantially contribute to an improved validity of aneurysmal haemodynamic simulation, thereby enhancing the clinical relevance of such aneurysmal haemodynamic studies in the future.