Patient-specific simulation of stent-graft deployment in type B aortic dissection: model development and validation

Abstract

Thoracic endovascular aortic repair (TEVAR) has been accepted as the mainstream treatment for type B aortic dissection, but post-TEVAR biomechanical-related complications are still a major drawback. Unfortunately, the stent-graft (SG) configuration after implantation and biomechanical interactions between the SG and local aorta are usually unknown prior to a TEVAR procedure. The ability to obtain such information via personalised computational simulation would greatly assist clinicians in pre-surgical planning. In this study, a virtual SG deployment simulation framework was developed for the treatment for a complicated aortic dissection case. It incorporates patient-specific anatomical information based on pre-TEVAR CT angiographic images, details of the SG design and the mechanical properties of the stent wire, graft and dissected aorta. Hyperelastic material parameters for the aortic wall were determined based on uniaxial tensile testing performed on aortic tissue samples taken from type B aortic dissection patients. Pre-stress conditions of the aortic wall and the action of blood pressure were also accounted for. The simulated post-TEVAR configuration was compared with follow-up CT scans, demonstrating good agreement with mean deviations of 5.8% in local open area and 4.6 mm in stent strut position. Deployment of the SG increased the maximum principal stress by 24.30 kPa in the narrowed true lumen but reduced the stress by 31.38 kPa in the entry tear region where there was an aneurysmal expansion. Comparisons of simulation results with different levels of model complexity suggested that pre-stress of the aortic wall and blood pressure inside the SG should be included in order to accurately predict the deformation of the deployed SG.

Keywords: Finite element analysis; TEVAR; Type B aortic dissection; Virtual stent-graft deployment.

Fig. 1

Illustration of the segmentation and reconstruction of aortic dissection geometry from CTA scan. a The transverse view of the descending aorta before TEVAR showed true lumen (TL), false lumen (FL) and the measurement of narrowed true lumen. b The segmentation mask A represents the blood flow domain and mask B encloses the flow domain and the intimal falp. c The aortic wall was created by extruding mask B outwardly by 1.5 mm. d The segmentation of intimal flap was created by performing Boolean subtraction of mask A from mask B. e The aortic wall and intimal flap were combined and trimmed to form the pre-TEVAR aortic dissection geometry

Fig. 2

Summary of the steps in the simulation of stent-graft (SG) deployment and model variations. a The 28–28-150 mm Medtronic Valiant SG was used in TEVAR procedure and was covered by the virtual sheath. b The SG was compressed by the virtual sheath to its crimped state. c A curved tube opened up the local narrowing in the compressed true lumen. d The SG was delivered and deployed at the targeted position. e overall workflow and model variations

Fig. 3

a Definitions of parameters for quantitative assessment of stent-graft configuration. b Numbering of stent strut ends and the definition of landing sections

Fig. 4

Stress–stretch relationships for aortic tissue samples taken from type B aortic dissection patients

Fig. 5

Comparisons of local open area (LOA) at the strut ends measured from model A (a), model B (b) and model C (c) against the corresponding values measured from the post-TEVAR follow-up CT scan

Fig. 6

Deviations of the predicted stent-graft configuration from the three models. a Local open area (LOA) deviation (e_LOA). b Stent strut centre points position deviation (e_c(mm))

Fig. 7

Maximum principal stress maps on the aortic wall (top) and intimal flap (bottom) in pre- and post-TEVAR models. The virtually deployed stent graft is shown in the middle with the intimal flap being highlighted. All results were obtained with model C