Robust optimisation of Coronary Stents

The miniature nature of the coronary stent systems has substantially limited the extent of quantifying their post-procedural outcomes. Fortunately, advances in computational methods, represented by Finite Element Analysis (FEA), have provided powerful tools for estimating the stent performance, predicting longevity and enabling design optimisation of the stent in a patient-specific fashion. What more can computational methods offer? Despite the high success rates, late complications of coronary stents have been reported clinically, resulting in premature mechanical failure. Note that the optimisation method in the current literature has mostly been deterministic, offering sub-optimal solutions incapable of handling uncertain and noise factors in the system. Meanwhile, optimisation relies on building surrogate models accurately and efficiently, but this is hindered by the complexity of the highly nonlinear stent FEA. As a result, most works in the literature have chosen to simplify the FE models with substandard assumptions to reduce computing complexity and cost. These common problems seen in the literature have brought the creditability of the computational analysis results into question. For the above reasons, this thesis aims to develop the nondeterministic numerical models of the coronary stent system and presents a robust optimisation scheme for minimising the impact of uncertainties on stenting outcomes. In this study, we first explore the strategy for surrogate modelling and the optimisation method with a case study. At the same time, we investigate the influence of vessel factors on stent deployment, statistically revealing their significant effect on stenting responses. Then, the robust optimisation of coronary stents is performed in two studies. The first successfully reduces the residual stress in the stent and the vessel post-deployment and improves the stent expansion metrics while minimising performance variances across different biological and surgical uncertainties. The second study presents a robust fatigue optimisation, resulting in a vast improvement in the stent fatigue safety factor, showing high resistance to pulsatile cardiac pressure and low sensitivities to biological and material uncertainties. This thesis’s modelling results and optimisation methodology offer a cost-effective method for optimising the stent design. This can be beneficial for the stent device development, ensuring the longevity of the device and bettering the clinical outcomes