Investigating the Impact of Coronary Vascular Structure Diversity on Local and Global Hemodynamics

Introduction::
Coronary artery disease (CAD) is the leading cause of global mortality. Patient-specific anatomical features such as vessel diameters, curvatures, and bifurcations influence local hemodynamics and, as a result, promote atherogenesis. While atherosclerotic plaque tends to develop near bifurcations and curvatures, there are no predictive metrics to assess a patient's risk of atherosclerosis based on their coronary anatomy.


Using computational fluid dynamics (CFD), the relationship between anatomical variability and hemodynamics can be noninvasively quantified. Existing results suggest that patient-specific variability in bifurcation angles alone does not completely explain adverse hemodynamics. Instead, additional patient-specific features should be investigated to predict pathological hemodynamics. However, most studies are limited by small sample sizes, idealized geometries, or non-physiological boundary conditions. Additionally, previous studies have often neglected the impact that anatomical variations have on global coronary hemodynamics.


We hypothesize that a systematic sensitivity analysis of coronary anatomy variability will provide greater insight into the correlations between adverse hemodynamic conditions and patient-specific anatomy. In this work, we systematically analyzed how various bifurcation positions (a previously unexplored anatomical feature), bifurcation angles, sex-based branch diameters, tortuosity, and lumen eccentricity alter hemodynamic metrics such as wall shear stress (WSS), low WSS area, and normalized helicity intensity (NHI) volume in the left coronary artery. Furthermore, we investigated how bifurcation positions and angles impact hemodynamics locally near the bifurcation, as well as globally, in the left coronary branch.



Materials and Methods::
We generated 93 left coronary artery (LCA) anatomical models through parametric sweeps of bifurcation positions with angles, branch diameters in males and females, and varying levels of tortuosity and eccentricity. Angle, position, and diameter changes were applied at the left anterior descending (LAD) and left circumflex (LCx) junction (Figure 1A). Geometric parameters were sampled from healthy population angiographic data, and tortuosity was classified using a definition from Eleid et al [1,2]. Other parameters, such as artery diameter or tortuosity, were kept relatively constant. Each synthetic LCA geometry was generated from a baseline model of an adult LCA [1,3]. Each model had two side branches in the LAD and LCx.


Transient 3D blood flow simulations were conducted using SimVascular (a cardiovascular modeling open-source package). Boundary conditions included no-slip walls, a coronary inlet flow waveform with an average heart rate of 70 bpm, and outlet lumped-parameter networks (LPN) to represent coronary physiology (Figure 1A). The LPN was tuned to match a 120/70 mmHg pressure waveform. Blood was modeled as a Newtonian fluid as the average simulation shear rate exceeded 100/s. Rigid walls were assumed.


The WSS, TAWSS, WSS area below 0.5 Pa (ALWSS), and NHI were calculated in a spherical region at each bifurcation. WSS below 0.5 Pa was considered atherogenic. NHI thresholds of ±0.6, ±0.4, and ±0.3 approximated atheroprotective regions at proximal, medial, and distal bifurcations based on decreasing downstream helicity.



Results, Conclusions, and Discussions::
The LAD/LCx bifurcation position showed a strong correlation with ALWSS exposure (R = -0.97, p = 8.99x10-15 ) in comparison to the angle (R = -0.62, p = 0.0011) (Figure 1B). Bifurcation positions had a 3.76 greater impact on ALWSS compared to the angles. ALWSS predominantly formed at the lateral walls of each bifurcation. Similarly, peak TAWSS at the LAD/LCx exhibited negative correlations with position (R = -0.52, p = 0.0085) and angle (R = -0.52, p = 0.0085). Global hemodynamics changes were associated with the LAD/LCx bifurcation position. Proximal LAD/LCx positions strongly correlated with ALWSS at each distal bifurcation (R = -0.93, p = 1.86x10-8 ) compared to a weak correlation with angle (R = -0.14, p = 0.54).


Bifurcation angles were weakly correlated with NHI (R = 0.16, p = 0.46), while position had a stronger correlation (R = 0.79, p = 4.39x10-6 ). Anatomy with higher NHI correlated with lower exposure to pathological WSS near the LAD/LCx bifurcation (R = -0.79, p = 2.4x10-6). Peak ALWSS at the LAD/LCx bifurcation occurred in anatomies with large diameter differences between the LAD and LCx. Smaller diameter branches had statistically higher ALWSS irrespective of which branch was larger (p < 0.05). NHI increased with branch diameters (R2 = 0.86, p = 2.7x10-12)) There was no statistical difference between male and female ALWSS. Branch-wise tortuosity and eccentricity had a negligible effect on ALWSS.


In conclusion, LAD/LCx bifurcation positions and LAD/LCx diameter differences greatly impacted coronary hemodynamics. Locally, ALWSS positively correlated with upstream bifurcation positions and large LAD/LCx diameter variations. NHI correlated with lower ALWSS. Distal bifurcation ALWSS depended on upstream LAD/LCx position and diameters, not the angle. Future simulation work will incorporate non-culprit lesions in the left coronary arteries. These findings will be generalized to patient-specific models. Finally, these simulations will inform a data-driven model to predict regions of high atherosclerotic risk for patient-specific anatomy.



References (Optional): :
[1] P. Medrano-Gracia et al., “A computational atlas of normal coronary artery anatomy,” EuroIntervention, vol. 12, no. 7, pp. 845–854, Sep. 2016, doi: https://doi.org/10.4244/eijv12i7a139.


[2] M. F. Eleid et al., “Coronary Artery Tortuosity in Spontaneous Coronary Artery Dissection,” Circulation: Cardiovascular Interventions, vol. 7, no. 5, pp. 656–662, Oct. 2014, doi: 10.1161/CIRCINTERVENTIONS.114.001676.


[3] "Vascular Model Repository," Vascular Model Repository, accessed Apr. 24, 2023. [Online]. Available: https://www.vascularmodel.com.