Efficiently Simulating an Endograft Deployment: A Methodology for Detailed CFD Analyses

Abstract

Numerical models of endografts for the simulation of endovascular aneurysm repair are increasingly important in the improvement of device designs and patient outcomes. Nevertheless, current finite element analysis (FEA) models of complete endograft devices come at a high computational cost, requiring days of runtime, therefore restricting their applicability. In the current study, an efficient FEA model of the Anaconda™ endograft (Terumo Aortic, UK) was developed, able to yield results in just over 4 h, an order of magnitude less than similar models found in the literature. The model was used to replicate a physical device that was deployed in a 3D printed aorta and comparison of the two shapes illustrated a less than 5 mm placement error of the model in the regions of interest, consistent with other more computationally intensive models in the literature. Furthermore, the final goal of the study was to utilize the deployed fabric model in a hemodynamic analysis that would incorporate realistic fabric folds, a feature that is almost always omitted in similar simulations. By successfully exporting the deployed graft geometry into a flow analysis, it was illustrated that the inclusion of fabric wrinkles enabled clinically significant flow patterns such as flow stagnation and recirculation to be detected, paving the way for this modelling methodology to be used in future for stent design optimisation.

Validation and Verification of High-Fidelity Simulations of Thoracic Stent-Graft Implantation

“…Then, the catheter was removed to let the device expand: the stent-graft was deployed at once and the free flow ring was not released as last, but simultaneously with all the stent struts. In particular, simulations with Nitinol modeled as linear elastic material ( E = E A = 57,500 MPa 13 , 16 , 21 , 29 ) or with PET as linear elastic material ( E = 1.84 MPa 18 , 20 , 33 and E = 1.84 GPa 24 ) were tested as well as the effect of the absence of the stent pre-stress. 18 , 19 , 33 Also, to investigate the stent-graft releasing phase, a condition of modified “virtual catheter” method was studied where the stent-graft was deployed gradually.…”

“…12 For the stent component, some studies simplify the superelastic Nitinol behavior with a linear elastic material model with Young's modulus equal to the austenitic one. 13,16,21,29 In addition, a few works introduce the pre-stress field in the unloaded stent component. 7,11,17,19,34 This study aims at developing high-fidelity FE simulations to virtually reproduce the TEVAR procedure.…”

“…In the context of endograft numerical simulations, the recent literature involves several finite element (FE) structural studies regarding stent-graft modeling by adopting different strategies. Regarding the stent mesh, either 1D beam elements 11 , 9 , 13 , 21 , 30 , 29 or 3D hexahedral elements 5 , 14 , 18 , 19 , 32 , 33 are adopted; the graft is discretized using 2D triangular or quadrangular shell elements 9 , 13 , 21 , 30 or membrane elements. 5 , 11 , 18 , 19 , 32 , 33 Regarding the material properties, the graft fabric is modelled as a linear elastic material in most cases or as a linearized orthotropic linear elastic material.…”

A Comparative Study on the Hemodynamic Performance Within Cross and Non-cross Stent-Grafts for Abdominal Aortic Aneurysms With an Angulated Neck

“…Third, the computational models were geometrically asymmetrical, phase shift of pressure and different flow splits at two iliac outlets may occur ( Grinberg and Karniadakis, 2008 ). Therefore, the flow split between two iliac branches in the models was regulated by the Windkessel model ( Kyriakou et al, 2020 ), where their RCR parameters were matched for 1 million times by trial-and-error method. The matching blood pressure was the iliac artery outlet blood pressure (100/80 mmHg) given in the previous literature ( Mills et al, 1970 ).…”

“…Matsagkas et al showed that the hemodynamic indexes of different laminating stents might significantly differ from the actual model ( Raptis et al, 2017b ). Finally, previous studies have proposed that the folds caused by the bending of stent grafts will affect the hemodynamics near the wall ( Prasad et al, 2012 ; Kyriakou et al, 2020 ). In vivo limb stents involve many curved parts, but the current models did not consider the impact of folds, which will be the focus of our future research.…”