Speaker: Daniel B Pike. Advisor: Yan-Ting Shiu, Ph.D.

Abstract:

Arteriovenous grafts (AVG) as vascular access for chronic hemodialysis suffer from significant failure rates due to stenosis that is often caused by neointimal hyperplasia (NH) at the graft-venous anastomosis. No treatment is available to prevent NH-caused stenosis. We propose that an important NH-causing factor is early and sustained ‘disturbed’ flow (i.e. high and low wall shear stress (WSS), high spatial wall shear stress gradient (WSSg), and high oscillatory shear index (OSI)), which often occurs at the venous anastomosis immediately following graft implantation surgery but is not yet characterized in detail. Additionally, previous work has not explored the specific spatial and temporal relationship between the location of early and sustained ‘disturbed’ flow hemodynamics on the location and size of NH in the AVG setting. Our hypothesis is that early and sustained disturbed flow causes the development of NH at specific locations in AVGs. To test this hypothesis, we propose four specific aims, which all use a large animal (porcine) model of AVG stenosis, where hemodynamics are comparable to that in humans. Specific Aim 1 is to determine the location and amount of NH development over time by using contrast-free 3 dimensional (3D) black-blood (BB) magnetic resonance imaging (MRI). Specific Aims 2, 3 and 4 are to correlate the location and amount of NH with early hemodynamics, as determined by computational fluid dynamics (CFD) simulations that do not consider vessel wall deformation and motion (Aim 2), CFD simulations that consider vessel wall deformation and motion (Aim 3), and directly measured velocity fields in the entire AVG lumen by flow-sensitive 4 dimensional (4D) MRI (Aim 4). Each of the three methods provides unique and/or additional hemodynamic information. The CFD method in Aim 2 has been developed in our lab, and CFD with the rigid and immobile wall assumptions has been shown to simulate the flow fields of non-expanding AVG regions accurately in previous research. CFD simulations with the deformable and mobile wall (i.e., fluid-structure interaction (FSI)) in Aim 3 could potentially refine our existing CFD model, which could be particularly important in the vein where wall expansion is greater than in the artery. Flow-sensitive 4D MRI in Aim 4 measures blood velocity in the entire AVG lumen, from which WSS can be calculated directly and does not require assumptions needed for CFD simulation. Hence, 4D MRI could potentially help validate the CFD results. The contrast-free MRI protocol developed in the proposed study can be used in dialysis patients, to whom contrast agents could be toxic. The results of this proposed study will provide insight into the causal effect of disturbed flow on NH development in AVGs, future directions in the development of refined CFD simulations of blood flow in AVGs, and rationales for improving AVG placement surgical techniques to minimize disturbed flow, which could lead to increased patency rates and lowered intervention rates. Additionally, this research could be applied to studying venous stenosis in the AV fistula or vein bypass grafts, which also have aberrant hemodynamics.