VERIFICATION, VALIDATION AND UNCERTAINTY QUANTIFICATION IN PATIENT-SPECIFIC TAVR MODELS

In transcatheter aortic valve replacement (TAVR), computational models have shown potential to predict transcatheter heart valve (THV) deployment and improve patient outcomes. However, their predictive accuracy is influenced by uncertainties from imaging modalities, model assumptions, and patient variability. Systematic uncertainty analyses are essential for characterizing these uncertainties and quantifying their impact on predictions. Model credibility assessment relies on verification, validation, and uncertainty quantification (VVUQ), as outlined by the FDA and the ASME V&V40 standard, "Verification and Validation in Computational Modeling of Medical Devices" [1]. This study applies the ASME V&V40 framework to establish the credibility of computational models simulating balloon-expandable SAPIEN 3 (S3) TAVR, employing techniques like surrogate modeling, Monte Carlo simulations (MCS), and cumulative distribution functions (CDFs) to compare simulations with clinical data.