Computational and Experimental Investigation into the Hemodynamics of Endovascularly Treated Cerebral Aneurysms

Thesis (Ph.D.)--University of Washington, 2018This thesis investigates the influence of mechanical stresses and fluid motion on the efficacy of endovascular treatment methods for cerebral aneurysms. The two primary endovascular treatments, embolic coils and flow-diverting stents, are designed to reduce blood-flow and stress inside the aneurysm volume - creating an environment that enables the development of a stable intra-luminal thrombus. Successful embolization of an aneurysm eliminates the risk of aneurysm rupture. Unfortunately, endovascular aneurysm treatments have a high rate of failure, leaving the patient at risk of rupture and warranting further intervention. The hemodynamic environment inside of aneurysms, both before and after treatment, likely plays a crucial role in the healing process. In this thesis, we hope to improve the clinical outcomes of endovascularly treated aneurysms by characterizing the hemodynamic modifications caused by endovascular treatments and developing clinically accessible tools that are predictive of treatment outcome. The primary work of this thesis is the development of a computational framework that predicts the outcomes of endovascularly treated cerebral aneurysms based on computed hemodynamic metrics. In collaboration with Harboview Medical Center, we create patient-specific computational simulations of the hemodynamic environment, both before and after endovascular treatment. Changes in the aneurysmal hemodynamics are then compared to the outcomes of each treatment, assessed 18 months following endovascular repair. The model developed in this thesis, importantly, incorporates patient-specific velocity and pressure measurements which improves the accuracy of the model, and, thus, it's clinical applicability. The results from this study have shown that significant differences exist between the hemodynamic environments in successful and unsuccessful cerebral aneurysm treatments, suggesting that the model is predictive of treatment outcome. Aneurysms successfully treated with flow-diverting stents are found to have a larger reduction in aneurysmal flow-rate and aneurysm WSS following treatment than those aneurysm that did not fully heal. Coiled aneurysms that failed treatment are found to have a significantly larger increase in shear at the neck of the aneurysm following treatment than successfully coiled aneurysms. The results from this study improve our fundamental understanding of treatment induced aneurysmal hemodynamic changes and moves hemodynamic based prediction of treatment outcomes closer to the realm of patient care and treatment planning. In a second study, we characterize the hemodynamics of idealized cerebral aneurysms treated with flow-diverting stents. Using stereo particle image velocimetry (PIV), we measure the flow-field in silicone idealized aneurysm models following flow-diverting stent treatment. The parameter space explored in this study is the flow-rate and curvature of the parent-vessel, or parent-vessel Reynolds and Dean number. Flow-diverting stents are found to significantly reduce the velocity inside the aneurysm, creating an aneurysmal flow-environment that is viscous dominated, despite inertial dominated parent-vessel flow. Thus, at low values of Dean number, flow enters the aneurysm at the leading edge and remains attached to the aneurysm wall before exiting at the downstream edge, rotating with the direction of the parent-vessel flow, characteristics similar to Stokes flow over a cavity. As the Dean number is increased, the flow along the leading edge begins to separate, and the recirculation region grows, until, above a Dean number of 180, the flow inside the aneurysm is fully recirculating. The magnitude of flow entering the aneurysm is also found to increase as a function of parent-vessel Dean number. Flow-diverting stents are designed to achieve aneurysm occlusion by sufficiently reducing the amount of blood-flow entering the aneurysm so a thrombus can begin to form. The results from this study, however, show that aneurysms attached to vessel with a high Dean number are subjected to large values of aneurysmal flow, which may compromise the effectiveness of the flow-diverting stent treatments. In the final part of this thesis, we investigate the accuracy of a modeling technique commonly used in the simulation of blood flow in coiled cerebral aneurysms. Rather than resolve the complex structure of a deployed mass of coils inside an aneurysm, the porous media model approximates the resistance of the coil mass by numerically adding a momentum sink term to the volume of the aneurysm sac. The porous-media approach is the gold-standard for simulating blood-flow through embolic coils due to its numerical efficiency, however, its accuracy and limitations have never been assessed. In this study, we compared the accuracy of simulations that rely on the porous media model to simulations that resolve the coil-mass. Deployed coil geometries are obtained from high-resolution, synchrotron radiation based micro-CT scans of patient-specific silicon models of aneurysms "treated" with embolic coils. The porous media model is shown to both over- and under-estimate shear stress inside the aneurysm dome, and systematically over-estimate the flow-rate of blood entering the aneurysm. Using homogenization methods we analyzed the complex coil-structures in four treated aneurysm and developed an improved porous media model for each aneurysm that better accounts for the complex structure of the deployed coils. Simulations that incorporate the improved porous media model are shown to be more accurate than the standard approach, improving the clinical utility of the simulation results