Title

A Genetic-Algorithm-Based Design Approach To Minimize Abnormal Hemodynamics Parameters At The End-To-Side Distal Anastomoses Of Synthetic Bypass Grafts

Abstract

Medical records on post-operative performances of synthetic bypass grafting show poor results due to the restenosis occurring at the end-to-side distal (ETSD) anastomoses of the grafts. The re-stenosis is mainly caused by intimal hyperplasia (IH) [1]. A major factor behind the anastomotic IH development is the non-physiological flow patterns, such as separation and re-attachment, resulting from the bifurcating nature of the anastomotic geometry [2]. The spatial wall shear stress gradient (SWSSG) has been experimentally proven to cause proliferation and migration of the endothelial cells [3], resulting in an endothelial dysfunction that potentially leads to the initiation of IH. Several researchers attempted to optimize the shape of end-to-side distal anastomoses in order to mitigate the effects of the abnormal anastomotic flow patterns [4]. However, they only followed direct optimization approaches via comparing a limited number of different designs; those direct approaches cannot claim an optimal shape. As an alternative, the incorporation of a global search optimization approach can accurately yield an optimal solution. In the current work, a meshless computational fluid dynamics (CFD) solver is coupled to a genetic algorithm (GA) to achieve the anastomosis shape optimization purpose. The GA is an optimization tool that performs a search for the optimal shape of the anastomosis within a search range for the anastomosis design variables such as the anastomotic angle and the graft caliber. The GA constantly evolves a set of anastomotic geometric models throughout successive optimization generations. At each optimization generation, the GA modifies the anastomotic geometry based on reducing the SWSSG on the floor of the host artery and then exports the modified geometry to the meshless CFD solver. The meshless CFD solver automatically pre-processes the modified geometry and then solves the flowfield to evaluate the SWSSG. The SWSSG is then passed again to the GA for a new optimization generation. This cycle will repeat itself until the optimal anastomosis shape is located. The meshless notion of the CFD solver arises from the fact that the distribution of the solution points on the boundaries is independent from the distribution of the solution points in the domain interior (figure 1). This order independence between the boundary and internal points allows a simple integration of the GA with the meshless CFD solver in an autonomous optimization loop that does not require any human interaction to identify the optimal anastomosis shape. As this optimization application is in its early stages, the anastomotic flow is treated as steady, incompressible, and laminar. The effects of the arterial wall distensibility and the blood non-Newtonian rehology are neglected. Hence, the flow will be governed by the Navier-Stokes (N-S) equations. The numerical treatment of the N-S equations consists of a third-order explicit time-marching scheme and a radial-basis function (RBF) approach for the spatial discretization [5]. The RBF are used to preset interpolation vectors to determine derivatives needed for the numerical solution of the N-S. A pressure correction scheme is used to satisfy the continuity equation. The preliminary results reported in this abstract comprise of the meshless CFD solutions in a direct anastomosis model as shown in figure 2. The blood flow is simulated at a Reynolds number of 47 based on the graft caliber. The blood density and viscosity are taken to be equal to 1060 kg/m 3 and 0.004 N.s/m respectively. The bypass graft caliber is specified as 9 mm and the host artery diameter is equal to 6 mm, which is within the diameter range of a femoral artery. The anastomotic angle for the current model is chosen to be 45 degrees. The results obtained by the meshless CFD solver are benchmarked with the results obtained by a well-established finite volume method (FVM) CFD solver for validation purposes. The same geometry and flow conditions are used for both methods. The velocity contours of the flow as computed by both methods are shown in figures 3 and 4. Moreover, plots of the x-component of the wall shear stress (WSS) calculated by both methods at the floor of the host artery are shown in figure 5. The velocity and WSS results indicate a high agreement between the meshless method and the FVM. Copyright © 2007 by ASME.

Publication Date

1-1-2007

Publication Title

Proceedings of the ASME Summer Bioengineering Conference 2007, SBC 2007

Number of Pages

761-762

Document Type

Article; Proceedings Paper

Personal Identifier

scopus

DOI Link

https://doi.org/10.1115/sbc2007-176718

Socpus ID

40449083802 (Scopus)

Source API URL

https://api.elsevier.com/content/abstract/scopus_id/40449083802

This document is currently not available here.

Share

COinS