Patient-specific multi-scale design optimization of transcatheter aortic valve stents

Abstract

Background and objective: Transcatheter aortic valve implantation (TAVI) has become the standard treatment for a wide range of patients with aortic stenosis. Although some of the TAVI post-operative complications are addressed in newer designs, other complications and lack of long-term and durability data on the performance of these prostheses are limiting this procedure from becoming the standard for heart valve replacements. The design optimization of these devices with the finite element and optimization techniques can help increase their performance quality and reduce the risk of malfunctioning. Most performance metrics of these prostheses are morphology-dependent, and the design and the selection of the device before implantation should be planned for each individual patient.

Methods: In this study, a patient-specific aortic root geometry was utilized for the crimping and implantation simulation of 50 stent samples. The results of simulations were then evaluated and used for developing regression models. The strut width and thickness, the number of cells and patterns, the size of stent cells, and the diameter profile of the stent were optimized with two sets of optimization processes. The objective functions included the maximum crimping strain, radial strength, anchorage area, and the eccentricity of the stent.

Results: The optimization process was successful in finding optimal models with up to 40% decrease in the maximum crimping strain, 261% increase in the radial strength, 67% reduction in the eccentricity, and about an eightfold increase in the anchorage area compared to the reference device.

Conclusions: The stents with larger distal diameters perform better in the selected objective functions. They provide better anchorage in the aortic root resulting in a smaller gap between the device and the surrounding tissue and smaller contact pressure. This framework can be used in designing patient-specific stents and improving the performance of these devices and the outcome of the implantation process.

Keywords: Finite element method; Gaussian process regression models; Multi-objective optimization; Stent; Transcatheter aortic valve.