Computational Fluid Dynamics of Intracranial Aneurysms: Eulerian and Lagrangian Analysis of the Effect of Endovascular Treatment on Hemodynamics

Abstract

Intracranial aneurysms are dilated portions of an artery that supplies bllod to the brain. These abnormal dilatations of the arterial wall carry the risk of rupture, which represents a leading cause of subarachnoid hemorrhage and have very high mortatility and morbidity. Endovascular therapies are deployed by neurointerventionalists to treat intracranial aneurysms, reducing further growth and the risk of rupture. The success of these therapies involves the cessation of blood flow into the aneurysmal cavity, because of a clot fully occluding the aneurysm, allowing reendothelialization of the parent vessel. If there is any remnant flow, the treatment is considered unsuccessful and may warrant retreatment. While both coil embolization devices and flow-diverting stents (FDS) are proven endovascular therapies, there is no way to predict the treatment outcome in either of two treatment modalities, whether using pre- or post-operative information. The patient is therefore required to return for medical imaging to determine the outcome, which increases the burden on the healthcare system, as well as the procedural and rupture risk of the patient. The inability to predict treatment outcome can be addressed with the use of computational fluid dynamics (CFD). While many studies have been conducted with small patient populations, the use of CFD to understand the evolution of intracranial aneurysm has yet to identify thresholds for hemodynamics metrics to predict treatment outcomes, or even which metrics are physiologically relevant. Standard image-based patient-specific CFD simulations rely on a variety of models to account for the effect of treatment on hemodynamics. Unfortunately, a lack of standardization and automation, coupled with the uncertainty associated with many of the models used historically for their simplicity without a rigorous validation of their accuracy, has led to a lack of consensus on which hemodynamics metrics should be studied for their predictive potential. This dissertation investigates two main forms of endovascular therapy of cerebral aneurysms: coils and FDS, via a computational simulation framework that introduces novel models and validates them against gold-standard, experimentally-derived coil- and stent-resolved simulations. The overall goal is to contribute to the state-of-the-art simulation framework to move CFD towards becoming a clinical standard of care tool. The first question addressed is whether the two devices can be considered analogous when seeking metrics that are predictive of outcomes. The follow-up question is whether a new framework for studying the hemodynamics of intraaneurysmal flow and potential thrombosis, Lagrangian particle tracking, can shed light on the physiological processes that determine aneurysm embolization, and success post-treatment. As a side project, during my Fullbright stay at the Otto Von Guericke University in Magdeburg, Germany, I considered the differences between saccular and fusiform aneurysms and how the hemodynamics and FDS deployment differ for these two phenotypes. Finally, my last contribution is the evaluation of a new porous media model, used to model the coil mass comparing it to coil-resolved simulations to determine its efficacy in predicting Eulerian and Lagrangian metrics.