Detection of aortic wall inclusions using regional pulse wave propagation and velocity in silico

Abstract

Monitoring of the regional stiffening of the arterial wall may prove important in the diagnosis of various vascular pathologies. The pulse wave velocity (PWV) along the aortic wall has been shown to be dependent on the wall stiffness and has played a fundamental role in a range of diagnostic methods. Conventional clinical methods involve a global examination of the pulse traveling between two remote sites, e.g. femoral and carotid arteries, to provide an average PWV estimate. However, the majority of vascular diseases entail regional vascular changes and therefore may not be detected by a global PWV estimate. In this paper, a fluid-structure interaction study of straight-geometry aortas was carried out to examine the effects of regional stiffness changes on PWV. Five homogeneous aortas with increasing wall stiffness as well as two aortas with soft and hard inclusions were considered. In each case, spatio-temporal maps of the wall motion were used to analyze the regional pulse wave propagation. On the homogeneous aortas, increasing PWVs were found to increase with the wall moduli (R2 = 0.9988), indicating the reliability of the model to accurately represent the wave propagation. On the inhomogeneous aortas, formation of reflected and standing waves was observed at the site of the hard and soft inclusions, respectively. Neither the hard nor the soft inclusion had a significant effect on the velocity of the traveling pulse beyond the inclusion site, which supported the hypothesis that a global measurement of the average PWV could fail to detect regional abnormalities.

Pulse-Wave Propagation in Straight-Geometry Vessels for Stiffness Estimation: Theory, Simulations, Phantoms and In Vitro Findings

“…Therefore, the simulation results here also reflect the initial transient of the dynamic wave behavior. Nonetheless, the high correlation established in simulations for all homogenous and nonhomogeneous wall properties studied as previously reported [11], and the high correlation established between the simulation and experimental findings as well as the Moens-Korteweg theoretical relationship strongly confirm that the simulation framework accurately represents the wave propagation. As part of our ongoing studies related to the normal or pathological aortas, the 2D mesh is being considered in axisymmetric geometries, which allows for an affordable computational cost.…”

“…The initial time increment was 4.85 Â 10 À7 s with the average stable time increment and total CPU time, respectively, varying between 7.41 Â 10 À6 and 14.75 Â 10 À6 s, and 1.37 Â 10 6 and 1.46 Â 10 6 s, across different simulations. Additional details on using CEL for modeling the arterial pulse wave propagation can be found elsewhere [11]. In the radial direction (wall thickness), only one point was selected along the entire length of the tube to segment the wall and to measure the average wall displacement at the wall middle path (similar to the experimental and in vivo PWI studies).…”

“…Scotti et al have extracted temporal wall displacement variation using dynamic FSI; however, no spatial information was obtained and therefore no PWV measurements were extracted [25]. Preliminary studies on performing FSI simulations of aortic pulse wave propagation have been reported by our group [11,26]. The objective of this paper is to verify the use of such simulations of pulse wave propagation along the straight-geometry linearly elastic aortic walls against similar PWI studies on phantoms and in vitro canine aortas.…”

“…Due to a number of factors, such as the lack of knowledge on the exact arterial geometry and the assumption of a single longitudinal flow direction between the carotid and the femoral sites, such methods may lead to an overestimation/underestimation of the true distance [10]. Most importantly, the wall stiffness has been shown to vary along the arterial tree, and therefore assuming a global PWV may not detect early onset or focal disease [11]. These and other fundamental limitations of the carotid-femoral methods have been reviewed in details elsewhere [12].…”

“…The same as in the phantom and in vitro studies, the fluid was modeled as water with density q f ¼ 993 kg/m 3 , sound speed c 0 ¼ 1523 m/s, and viscosity g ¼ 0.007 N s/m 2 [28,29]. Additional details can be found in Appendix A.4, as well as in a previous publication on using CEL for modeling the arterial pulse wave along inhomogeneous walls [11].…”