PATIENT-SPECIFIC SIMULATION AND MODEL CREDIBILITY FOR TRANSCATHETER AORTIC VALVE IMPLANTATION

Transcatheter aortic valve implantation, known as TAVI, is a minimally invasive procedure for the treatment of severe aortic stenosis. Its expanding use in patients with complex anatomies and longer life expectancy underscores the need for reliable tools capable of predicting device performance and procedural outcomes. This thesis develops a comprehensive patient-specific computational framework for the analysis of TAVI, integrating anatomical variability, device mechanics, and blood flow dynamics within a unified modeling approach. Statistical shape modeling was employed to characterize interpatient anatomical variability of the aortic root and to generate virtual cohort representative of real clinical populations. Shape-derived features were further combined with machine learning techniques to predict functional indicators and support prosthesis sizing.High-fidelity finite element analysis was implemented to reproduce TAVI deployment using clinically accurate device models, while post-implantation hemodynamics were evaluated through Smoothed particle hydrodynamics. Model credibility was addressed through a rigorous Verification, Validation, and Uncertainty Quantification strategy in accordance with the ASME V&V40 standard. The framework was subsequently applied to investigate redo-TAVI interventions, enabling parametric assessment of implantation depth, valve size, and device–device interactions, with particular attention to coronary flow impairment. Finally, a novel transcatheter leaflet resection approach prior to redo-TAVI was explored in silico, demonstrating its potential to enhance valve expansion and improve hemodynamic performance.Overall, this work establishes a validated in-silico platform that supports device design, procedural optimization, and the development of predictive tools for personalized TAVI planning.