Finite Element Analysis of Cardiovascular Stent Frames: Identifying Appropriate Mesh Discretization

Brandon Lurie^1*Stuart Kari²Marc Horner³Christopher A Basciano⁴Santhosh Kumar Murugadass⁵Nuno Rebelo⁶

• ¹W. L. Gore & Associates, Inc., Flagstaff, AZ, United States
• ²Medtronic, PLC, Santa Rosa, CA, United States
• ³Ansys, Evanston, IL, United States
• ⁴Becton, Dickinson and Company, Durham, NC, United States
• ⁵Ansys, Bangalore, India
• ⁶Nuno Rebelo Associates, LLC, Fremont, CA, United States

Engineers use finite element analysis (FEA) to predict the deformations, strains, stresses, and resistive forces of metallic stent frames under in vivo, in vitro, and manufacturing-induced loading conditions. The discretization of the geometric model influences the simulation predictions, with the error generally reducing with mesh refinement. This improved accuracy comes with the trade-off of requiring more resources. Since FEA influences decisions that carry patient and business risk, engineers must balance the computational cost against numerical accuracy. This paper explores a methodology for selecting the mesh discretization for a computational model of an implantable stent frame based on discretization error, computational cost, and the risk associated with using the model to inform a specific decision.The methodology covers estimating the exact solution for the numerical model, calculating the discretization error and computational cost for various mesh discretization options, and considering the error and cost when selecting one of the options. The method was applied to a laser-cut nitinol stent model for four different finite element solvers to demonstrate its real-world applicability and that it is agnostic to solver type and developer.We were able to estimate the exact solution to the numerical model with a 95% confidence interval using submodeling, a geometry representative of the full stent frame, and four systematically refined meshes. The selection of the mesh discretization is subjective, with the importance of each model’s computational cost dependent on the number of simulations, resource availability, and risk. Three real-world implantable medical device examples of using FEA to inform a project decision are presented, each with a mesh discretization option suggested and rationalized based on the discretization error and computational costs.FEA’s important role in developing implantable stent frames and providing evidence of their safety to decision makers and regulatory bodies underscores the need for a method to select a suitable mesh discretization. The methodology explored in this paper calculates the error in the model’s prediction due to discretization and the computational cost. A project team can use this information, along with the risk associated with using the model, to select and rationalize a mesh discretization.