Toward a Computational Model for Macroscopic Predictions of Device-induced Thrombosis

Open Access
- Author:
- Taylor, Joshua Ollen
- Graduate Program:
- Bioengineering
- Degree:
- Doctor of Philosophy
- Document Type:
- Dissertation
- Date of Defense:
- October 16, 2015
- Committee Members:
- Keefe B Manning, Dissertation Advisor/Co-Advisor
Richard Scott Meyer, Committee Member
Thomas Neuberger, Committee Member
Dr Brent Craven, Committee Member
Peter J Butler, Committee Member
Donna Hope Korzick, Committee Member - Keywords:
- thrombosis
computational fluid dynamics
medical device
platelet adhesion
magnetic resonance imaging
LDV - Abstract:
- The efficacy of blood-contacting devices has improved significantly over the years; yet, thrombosis remains a concern with current device technology. A primary contributor to device-induced thrombosis is disturbed flow through the device; specifically, regions of high shear stress cause platelet activation and regions of low wall shear stress (WSS) promote thrombus deposition. In an effort to aid medical device developers, a mathematical model of device-induced thrombosis is developed to predict the entire thrombotic process, encompassing platelet activation, platelet adhesion, and thrombus growth. The mathematical model represents a significant reduction of the full thrombotic network and assumes the fluid mechanics of the system is the primary determinant of device-induced thrombosis. The single-scale thrombosis model considers bulk concentrations of platelets (non-activated, activated, and surface adherent) and adenosine diphoshpate (a chemical activator of platelets). A power law model predicts platelet activation based on the local shear stress, and a non-linear weighting function predicts thrombus deposition and growth based on the local WSS. A modified Brinkman term, used to simulate flow through porous media, acts as a momentum sink in the Navier-Stokes equations and is a function of the local aggregation intensity. The thrombosis model is developed with the use of in vitro thrombus size data collected in a backward-facing step geometry with magnetic resonance imaging, and it is implemented using the open-source computational fluid dynamics (CFD) toolbox OpenFOAM. The model is initially tested in a two-dimensional asymmetric sudden expansion, where growth is compared to the in vitro size data. After which, the model is implemented in a rotating disc system, where the governing equation for surface adherent platelets is developed using published in vitro platelet adhesion data. Finally, the thrombosis model governing equations are solved in a three-dimensional asymmetric sudden expansion, which represents the first application of a thrombosis model to a three-dimensional geometry on time scales necessary to observe macroscopic thrombus growth. Overall, a model capable of providing macroscopic predictions of thrombosis is developed and presented with the goal of simplifying and expediting the medical device design process.