Background: Calculation of fractional flow reserve (FFR) using computed tomography (CT)-based 3D anatomical models and computational fluid dynamics (CFD) has become a common method to non-invasively assess the functional severity of atherosclerotic narrowing in coronary arteries. We examined the impact of various inflow boundary conditions on computation of FFR to shed light on the requirements for inflow boundary conditions to ensure model representation. Methods: Three-dimensional anatomical models of coronary arteries for four patients with mild to severe stenosis were reconstructed from CT images. FFR and its commonly-used alternatives were derived using the models and CFD. A combination of four types of inflow boundary conditions (BC) was employed: pulsatile, steady, patient-specific and population average. Results: The maximum difference of FFR between pulsatile and steady inflow conditions was 0.02 (2.4%), approximately at a level similar to a reported uncertainty level of clinical FFR measurement (3–4%). The flow with steady BC appeared to represent well the diastolic phase of pulsatile flow, where FFR is measured. Though the difference between patient-specific and population average BCs affected the flow more, the maximum discrepancy of FFR was 0.07 (8.3%), despite the patient-specific inflow of one patient being nearly twice as the population average. Conclusions: In the patients investigated, the type of inflow boundary condition, especially flow pulsatility, does not have a significant impact on computed FFRs in narrowed coronary arteries.
Three-dimensional modeling and numerical analysis of fractional flow reserve in human coronary arteries
