Closing the Gap between Reality and CFD Simulations of FFR with Techniques to Quantify and Reduce Uncertainty of Predictions

Coronary artery disease has been reported the most common cause of death worldwide and is an ever-growing problem for global health. Recently, diagnosis of coronary artery disease from CT angiography (CTA) and computational fluid dynamics (CFD), CT-FFR, has emerged as a promising non-invasive alternative to the conventional clinical procedure. The aim of this thesis is to quantify and reduce the uncertainty of CT-FFR. In particular, we focus on uncertainties from the interaction between CFD and the coronary physiology. First, we use a lumped-element model to investigate the uncertainty and sensitivity of CT-FFR to physiological parameters. Second, we perform an in-depth investigation of the governing physiological model for flow distribution in CT-FFR simulations, Murray's law. Third, we propose a new model to reduce the inaccuracies from CTA-invisible coronary arteries. Our results show that uncertainty in physiological parameters has a significant effect on FFR estimates. Moreover, we find that CTA-invisible coronary arteries greatly increase the uncertainty of FFR, but that this effect can be reduced with improved mathematical modeling. Last, we find that CT-FFR is highly sensitive to post-stenotic flow outlets, which highlights the clinical importance of post-stenotic CTA image quality.