Coronary artery disease (CAD) is one of the leading causes of mortality worldwide, accounting for about 30% of all deaths [1]. Early diagnosis and correct estimation of the hemodynamic significance of coronary stenoses are crucial to initiate pharmacotherapy and to qualify the patient for possible percutaneous or surgical revascularization.
Currently, invasive pressure-wire-based methods such as fractional flow reserve (FFR) and resting indices, e.g., instantaneous wave-free ratio (iFR), are the standard to assess the significance of coronary stenoses. However, invasive determination of iFR/FFR has disadvantages, including (i) the use of a potentially nephrotoxic contrast agent and ionizing radiation, (ii) vascular access complications in 1.5% of patients, (iii) damage of the atherosclerotic plaque or healthy arterial segment during guidewire passage in 0.5% of patients [2], and (iv) side effects of adenosine administration, required during FFR measurements, in 35% of patients (chest pain, dyspnea, and transient arrhythmias) [2].
The increasing use of computed tomography angiography (angio-CT) to diagnose CAD and plan PCI procedures has triggered the development of novel non-invasive methods for iFRCT/FFRCT evaluation [3–6]. The values of iFRCT/FFRCT can be numerically determined based on the flow simulation and pressure measurements. In the last decade, 4 solutions for non-invasive estimation of the hemodynamic significance of coronary stenoses using FFRCT have become commercially available [7]. However, due to their feasibility only in a central core laboratory necessitating telemedicine and/or high costs, they are not commonly used in clinical practice. To circumvent the disadvantages of FFR, methods to determine iFRCT have also been developed but not yet commercialized [3-6]. Those methods have limitations (Table 1). Therefore, we aimed to assess the effect of simplifications implemented in previous iFRCT software on the estimation of coronary artery stenosis and propose a new method for coronary artery stenosis assessment using iFRCT.
The simplifications in Table 1 yield the following shortages in the methods: lack of information on the vessel complex geometry, lack of blood-wall interaction, non-realistic flow model, and no possible patient comorbidities considered. Those shortages affect reliable estimation of iFRCT.
In contrast to previous solutions (Table 1), our simulations include patient-specific conditions by considering blood pressure, stroke volumes, blood velocities, and heart rate (Fig. 1).
|
Reference |
Present study |
||||
|
Pearson correlation coefficient |
0.65 (iFRCT vs. FFR) |
– |
0.85 (iFRCT vs. iFR) |
0.68 (iFRCT vs. iFR) |
– |
|
Geometrical model |
3D model |
1D model |
3D model |
3D model |
3D model |
|
Flow assumption |
Newtonian fluid |
Newtonian fluid |
Newtonian fluid, laminar |
Newtonian fluid |
Non-Newtonian fluid, turbulent |
|
Coronary artery wall assumption |
Rigid wall |
Wall viscosity coefficient |
Rigid wall |
Rigid wall |
Elastic wall |
|
Simulation boundary conditions |
Resting correlation |
Average data from literature |
Correlation of flow and coronary vessel length |
Resting correlation |
Velocity, Windkessel model |
This approach makes it possible to incorporate the impact of the patient’s comorbidities on the hemodynamic parameters of the flow, enabling the most accurate representation of the physiological coronary blood flow. Based on CT, 3D models were created, taking into account the branches of the coronary arteries. This issue is important because the number, length, and irregularity of the branches affect the value of the pressure behind the stenosis. Despite the promising results obtained with previous numerical methods for determining coronary artery stenosis, it is necessary to be cautious about the potential underestimation of the calculated values [3–6]. We analyzed the effect of different variants of assumptions on the estimation of the iFRCT index, including (i) rigid wall, Newtonian blood flow, (ii) rigid wall, non-Newtonian blood flow, (iii) elastic wall, Newtonian blood flow, and (iv) elastic wall, non-Newtonian blood flow. The estimated iFRCT indices for the considered cases are: (i) 0.97, (ii) 0.99, (iii) 0.89, and (iv) 0.93, respectively (Fig. 2). Clearly, the differences between the index values are considerable, ranging up to 0.10 units, and potentially changing the decision regarding the need for revascularization. Subsequently, our results show that the assumptions related to coronary artery wall, boundary conditions, and blood flow are fundamental because they can lead to inaccurate clinical decisions.
Our new methodology of non-invasive estimation of the iFRCT index is based on numerical simulations of blood flow in the coronary arteries, whose geometry was generated from CT images. Our simulations showed that the above-listed conditions (iv) yielded the best match of the iFRCT index (0.93) to the invasive examination value of iFR 0.92.
We conclude that the assumptions made in the flow numerical modelling significantly affect the results. To reliably estimate the iFRCT index, numerous factors must be taken into account, i.e., a change in velocity over time as the boundary condition at the aortic inlet, boundary conditions at the truncated ends of the coronary arteries and the aortic outlet determined based on the Windkessel model, and fluid-structure interaction between the blood and the artery walls. The Windkessel model makes it possible to take into consideration very important factors determining the boundary conditions, i.e., the elasticity of blood vessels outside the considered artery system geometry, their possible pathology and complex reactions, and the effect of coronary artery wall material on blood flow. The results and the definition of an appropriate methodology can be the basis for the execution of an algorithm, allowing the determination of the iFRCT index or preoperative planning based on CT images [7–10]. Further studies will include a larger number of patients to further develop a new method to accurately identify the hemodynamic significance of coronary artery stenosis and facilitate revascularization decisions.
Author contributions: Anna Nieroda: conceptualization, methodology, investigation, writing — original draft preparation. Krzysztof Jankowski: conceptualization, investigation, writing — original draft preparation. Janusz Domański: conceptualization, investigation. Michał Janiszewski: conceptualization, investigation, data curation. Marcin
Komorowski: conceptualization, investigation, data curation. Krzysztof Lamparski: investigation, data curation. Janusz Kochman: investigation, data curation. Aleksandra Gąsecka: conceptualization, investigation, writing — original draft preparation, writing — review & editing. Marek Pawlikowski: conceptualization, methodology, investigation, writing — original draft preparation, writing — review & editing.
Acknowledgments: The work has been funded by Warsaw University of Technology grant No. 09/2023/RND IB/WMT.
Conflict of interest: None declared.
Funding: The work has been funded by Warsaw University of Technology grant No. 09/2023/RND IB/WMT.
