Spatial correlations between MRI-derived wall shear stress and vessel wall thickness in the carotid bifurcation

• Published in: European radiology experimental Vol. 2; no. 1; pp. 27 - 10
• Main Authors: van Ooij, Pim, Cibis, Merih, Rowland, Ethan M., Vernooij, Meike W., van der Lugt, Aad, Weinberg, Peter D., Wentzel, Jolanda J., Nederveen, Aart J.
• Format: Journal Article
• Language: English
• Published: Cham Springer International Publishing 10.10.2018; SpringerOpen
• Subjects: Atherosclerosis; Carotid artery; Diagnostic Radiology; Imaging; Internal Medicine; Interventional Radiology; Mechanical stress; Medicine; Medicine & Public Health; Neuroradiology; Original; Original Article; Radiology; Ultrasound; Wall thickness; Carotid artery; Wall thickness; Mechanical stress; Atherosclerosis
• ISSN: 2509-9280; 2509-9280
• DOI: 10.1186/s41747-018-0058-1

Background To explore the possibility of creating three-dimensional (3D) estimation models for patient-specific wall thickness (WT) maps using patient-specific and cohort-averaged WT, wall shear stress (WSS), and vessel diameter maps in asymptomatic atherosclerotic carotid bifurcations. Methods Twenty subjects (aged 75 ± 6 years [mean ± standard deviation], eight women) underwent a 1.5-T MRI examination. Non-gated 3D phase-contrast gradient-echo images and proton density-weighted echo-planar images were retrospectively assessed for WSS, diameter estimation, and WT measurements. Spearman’s ρ and scatter plots were used to determine correlations between individual WT, WSS, and diameter maps. A bootstrapping technique was used to determine correlations between 3D cohort-averaged WT, WSS, and diameter maps. Linear regression between the cohort-averaged WT, WSS, and diameter maps was used to predict individual 3D WT. Results Spearman’s ρ averaged over the subjects was − 0.24 ± 0.18 ( p  < 0.001) and 0.07 ± 0.28 ( p  = 0.413) for WT versus WSS and for WT versus diameter relations, respectively. Cohort-averaged ρ, averaged over 1000 bootstraps, was − 0.56 (95% confidence interval [− 0.74,− 0.38]) for WT versus WSS and 0.23 (95% confidence interval [− 0.06, 0.52]) for WT versus diameter. Scatter plots did not reveal relationships between individual WT and WSS or between WT and diameter data. Linear relationships between these parameters became apparent after averaging over the cohort. Spearman’s ρ between the original and predicted WT maps was 0.21 ± 0.22 ( p  < 0.001). Conclusions With a combination of bootstrapping and cohort-averaging methods, 3D WT maps can be predicted from the individual 3D WSS and diameter maps. The methodology may help to elucidate pathological processes involving WSS in carotid atherosclerosis.