Date of Award
Master of Science (MS)
LaDisa, Jr., John F.
Stent geometry influences local hemodynamic alterations (i.e. the forces moving blood through the cardiovascular system) associated with adverse clinical outcomes. Computational fluid dynamics (CFD) is frequently used to quantify stent-induced hemodynamic disturbances, but previous CFD studies have relied on simplified device or vascular representations. Additionally, efforts to minimize stent-induced hemodynamic disturbances using CFD models often only compare a small number of possible stent geometries. This thesis describes methods for modeling commercial stents in patient-specific vessels along with computational techniques for determining optimal stent geometries that address the limitations of previous studies.
An efficient and robust method was developed for virtually implanting stent models into patient-specific vascular geometries derived from medical imaging data. Models of commercial stent designs were parameterized to allow easy control over design features. Stent models were then virtually implanted into vessel geometries using a series of Boolean operations. This approach allowed stented vessel models to be automatically regenerated for rapid analysis of the contribution of design features to resulting hemodynamic alterations. The applicability of the method was demonstrated with patient-specific models of a stented coronary artery bifurcation and basilar trunk aneurysm to reveal how it can be used to investigate differences in hemodynamic performance in complex vascular beds for a variety of clinical scenarios.
To identify hemodynamically optimal stents designs, a computational framework was constructed to couple CFD with a derivative-free optimization algorithm. The optimization algorithm was fully-automated such that solid model construction, mesh generation, CFD simulation and time-averaged wall shear stress (TAWSS) quantification did not require user intervention. The method was applied to determine the optimal number of circumferentially repeating stent cells (NC) for a slotted-tube stents and various commercial stents. Optimal stent designs were defined as those minimizing the area of low TAWSS. It was determined the optimal value of NC is dependent on the intrastrut angle with respect to the primary flow direction. Additionally, the geometries of current commercial stents were found to generally incorporate a greater NC than is hemodynamically optimal.
The application of the virtual stent implantation and optimization methods may lead to stents with superior hemodynamic performance and the potential for improved clinical outcomes. Future in vivo studies are needed to validate the findings of the computational results obtained from the methods developed in this thesis.