Computational Fluid Dynamics Simulations of Hemodynamics in Plaque Erosion

Abstract

Purpose: We investigated whether local hemodynamics were associated with sites of plaque erosion and hypothesized that patients with plaque erosion have locally elevated WSS magnitude in regions where erosion has occurred.

Methods: We generated 3D, patient-specific models of coronary arteries from biplane angiographic images in 3 human patients with plaque erosion diagnosed by optical coherence tomography (OCT). Using computational fluid dynamics, we simulated pulsatile blood flow and calculated both wall shear stress (WSS) and oscillatory shear index (OSI). We also investigated anatomic features of plaque erosion sites by examining branching and local curvature in x-ray angiograms of barium-perfused autopsy hearts.

Results: Neither high nor low magnitudes of mean WSS were associated with sites of plaque erosion. OSI and local curvature were also not associated with erosion. Anatomically, 8 of 13 hearts had a nearby bifurcation upstream of the site of plaque erosion.

Conclusions: This study provides preliminary evidence that neither hemodynamics nor anatomy are predictors of plaque erosion, based upon a very unique dataset. Our sample sizes are small, but this dataset suggests that high magnitudes of wall shear stress, one potential mechanism for inducing plaque erosion, are not necessary for erosion to occur.

Keywords: Computational fluid dynamics; atherosclerosis; endothelium; plaque erosion; wall shear stress.

Figures

Figure 1. Plaque erosion in right coronary artery

Images from an angiogram in a patient who presented to Emory University Hospital and underwent cardiac catheterization. A filling defect from a large thrombus was identified in the right coronary artery, (circled region). Subsequent OCT imaging identified the source of the thrombus as an eroded plaque.

Figure 2. Optical coherence tomography identification of plaque erosion

Intravascular OCT reveals the cross-sectional anatomy of a coronary artery (top) and a longitudinal view of the vessel (bottom). Thrombus appears at the bottom of the cross-section as positive signal with jagged edges. We approximated the vessel boundary (yellow line) although underneath the thrombus, exact segmentation is impossible. In the longitudinal view, extent of the thrombus is marked with a horizontal white line.

Figure 3. Segmentation of culprit vessels in Paieon software

We semi-automatically identified the silhouette of vessels in biplane angiograms for each of three patients using Paieon software. Filling defects and blockage sites are circled.

Figure 4. Coronary velocity waveform

We prescribed a blunt velocity profile through an inlet flow extension into each CFD model. This waveform was derived from a typical patient, as patient-specific velocity data were not available.

Figure 5. Identification of local curvature

For each lesion identified in x-ray angiograms (circled), we tallied whether the local region was serpentine (left), gradually curving (middle), or locally straight (right) based on the number of inflection points near the lesion.

Figure 6. Wall Shear Stress at Sites of Plaque Erosion is Neither High Nor Low

Box plots of temporal mean WSS over the cardiac cycle with superimposed WSS magnitude at site of thrombus formation revealed that extreme magnitudes of WSS were not associated with plaque erosion in these patients. Whiskers represent minimum and maximum WSS of all elements, the box represents the 25th/75th percentile of elements, and the middle line represents median WSS.

Figure 7. Mean Wall Shear Stress

We computed the temporal mean WSS for all three patients. All patients’ erosions were at sites containing low magnitudes of WSS, but this is not a unique feature to the erosion sites.

Figure 8. Oscillatory shear index

We calculated oscillatory shear index, a measure of the duration of the cardiac cycle when flow deviates from its mean direction by more than π/6 radians. No remarkable patterns in OSI related to plaque erosion location were identified.

Figure 9. Transient helical flow structures

Occasional helical flow structures (arrows) appeared at the site of plaque erosion when visualized by streamline analysis. The zoomed region encompasses the site of erosion where the filling defect was observed in the angiogram. Other regions of the same coronary artery (not shown) also exhibited transient helical flow structures.