The Development of a Robust Low Computational Cost Diagnostic Tool to Evaluate Stenosis Functional Significance in Human Coronary Arteries

There is discordance between the anatomic severities of the coronary narrowing and their corresponding functional significance. Fractional Flow Reserve (FFR) is among the physiological parameters invasively measured to assess the hemodynamic significance of a stenosis during maximal hyperemia. FFR values ≤ 0.8 indicate that the downstream heart tissue perfused by this vessel is at risk for ischemia. While measuring FFR is an invasive procedure that is expensive, time consuming, and not without complications, recently, noninvasive estimation of FFR was shown to be possible from comprehensive predictive techniques allowing the computation of in-vivo FFR. However, these non-invasive methods are associated with high computational cost and require high performance computing technology, thus, reducing their wide adoption in clinics. This paper is steered to achieve two main goals: (1) to develop a fast numerical method to aid clinicians assessing ischemia level and determine if coronary revascularization (PCI) is required in human diseased coronary arteries with minimum time and computer resources; (2) to develop a robust method which allows predicting the patient FFR independently of the actual in-vivo physiologic conditions (mainly pressure) of the specific patient. The numerical framework was designed by adopting the finite volume method to generate the discrete model of the Reynolds average form of conservation equations used to predict blood hemodynamics. Two strategies were investigated to reduce computational cost while retaining solution accuracy. The first strategy is based on isolating the diseased artery from its branch tree and simulating it separately without implicitly integrating other arterial segments. A lumped dynamic model with special numerical treatment is coupled to the 3D domain outlet boundary to account for the downstream effects from the vascular bed. The second strategy is based on replacing a full transient simulation by a steady state one performed under mean conditions of pressure and blood volume flow rate. The strategy was applied on a healthy (hypothetical) and stenosed arterial segments with different stenosis severities simulated under rest and hyperemic conditions. An excellent agreement was achieved for FFR values computed from full transient simulations with the ones obtained from steady state simulation (error < 0.2 % was obtained for all test cases). The computational cost for the mean condition scenario was 0.1 that of a full transient simulation. The robustness of the method was tested by varying inflow conditions and reporting their effect on FFR. Interestingly, the predicted ischemia level was not altered when the inlet pressure was increased by 10 % from the base case. An analytical model was derived to explain the FFR independency of patient in-vivo coronary pressure. These promising findings from the numerical tests performed on idealized healthy and stenosed arterial models could significantly impact the applicability of the developed methodology and translating it into future practical clinical applications.Copyright © 2015 by ASME