Multi-fidelity estimators for coronary circulation models under clinically-informed data uncertainty

• DOI: 10.48550/arxiv.1911.11266
• Publication Date: 2019-01-01
• Authors: Jongmin Seo; Casey Fleeter; Andrew M. Kahn; Alison L. Marsden; Daniele E. Schiav...
• Keywords: Uncertainty Quantification

Numerical models are increasingly used for non-invasive diagnosis and treatment planning in coronary artery disease, where service-based technologies have proven successful identifying hemodynamically significant hence potentially dangerous vascular anomalies. Despite recent progress towards clinical adoption, many results the field still based on a deterministic characterization of blood flow, with no quantitative assessment variability simulation outputs due to uncertainty from multiple sources. In this study, we focus parameters that essential construct accurate patient-specific representations circulation, such as aortic pressure waveform, intramyocardial quantify how their affects clinically relevant model outputs. We deformable left subject prescribed inlet open-loop outlet boundary conditions, treating fluid-structure interaction through an Arbitrary-Lagrangian-Eulerian frame reference. Random input is estimated directly repeated measurements intra-coronary catheterization complemented by literature data. also achieve computational cost reductions propagation thanks multifidelity Monte Carlo estimators interest, leveraging ability generate, at practically cost, one- zero-dimensional low-fidelity appropriate conditions. The demonstrate use multi-fidelity control variate leads variance accuracy improvements respect traditional Monte-Carlo.