Haemodynamic simulation and optimisation for the flow-diversion treatment of intracranial aneurysms

Despite various complications reported, flow-diverting (FD) stent implantation has become a major mode of treatment for intracranial aneurysms. Review of literature suggests that the treatment outcomes are closely associated with the aneurysmal haemodynamics changed by the implanted stent. In this thesis, I sought to examine the haemodynamic changes following different strategies of FD treatment and provide practical solutions to improve the stent's flow-diversion efficacy. As the first accomplishment, I developed in Chapter 3 an automated optimisation method to be used to accommodate stent configuration to a specific aneurysm geometry. Adopting this method, wire structure of a stent can be modified to alter the local haemodynamics towards maximal reduction of the aneurysmal inflow. Furthermore, employing a virtual stent deployment technique and computational fluid dynamics analysis, I systematically investigated in Chapter 4 and 5 the aneurysmal haemodynamics following the flow-diversion treatment with a stent compaction technique applied or with dual-FD stents deployed - two of the most commonly adopted treatment strategies in the current interventional practice. These studies provided quantitative results illustrating stent wire characteristics and the subsequent aneurysmal haemodynamic changes in various flow-diversion scenarios. Finally, I examined in Chapter 6 the aneurysmal haemodynamics affected by incomplete stent expansion (IncSE) - a condition suspected to cause delayed aneurysm occlusion, for which various severities of IncSE occurring at different segments of the parent artery were modelled and examined. Results suggest that the effects of IncSE vary greatly with respect to the location where it occurs. Based upon this series of studies, a workflow was put forward to individualise the treatment plan corresponding to the haemodynamic characteristics of a specific patient, from the patient aneurysm model reconstruction, to computer-aided stent deployment rehearsal, and then to the haemodynamic outcome analysis to determine the most effective treatment plan prior to the real treatment. An example of implementing this workflow to predict the post-treatment haemodynamics thereby determining the most favourable treatment strategy was illustrated in detail in Chapter 7. Apart from this useful application, methodologies developed in the thesis, including the FD stent modelling technique, the stent structural optimisation method, the classification of stent compaction, and the in-house software for stent deployment simulation, would certainly contribute to future research and development of FD stents and interventional planning.