Simultaneous Measurements of Vascular Strain and Wall Shear Stress in the Carotid Artery based on Vector Flow Imaging and Vessel Wall Tracking in Duplex Mode

Abstract

Vascular strain (VS) is an elastic index reflecting the arterial stiffness. Wall shear stress (WSS) is a hemodynamic parameter which is considered a friction force exerted by blood flow on the arterial wall. Although the VS and WSS are different types of clinical parameters, both are related to atherosclerosis and intima thickening. Currently, it is difficult to measure both parameters at the same cardiac cycle since they are implemented based on very different algorithms. In clinic, simultaneous measurements of VS and WSS could maintain high consistency of vascular health assessment. The paper presented a framework, where interleaved transmissions of focus and plane waves were employed to generate B-mode images and blood flow velocities, which were used to calculate VS and WSS, respectively. Therefore, simultaneous calculations for VS and WSS can be implemented. In order to directly visualize the VS and WSS results on the arterial wall, a convolutional neural network (CNN) was used to segment the vessel wall of common carotid artery (CCA). Both the B-mode images and Doppler flow signals were used in the segmentation of arterial walls. Totally, 95 samples with 4580 ultrasound images were used in the CNN-based segmentation of vessel walls. On a labeled test dataset with 485 CCA images, the segmentation result in which 31 images were affected by artifacts inside the blood vessel, achieved a dice similarity coefficient of 97.21 %,and was effectively improved by using Doppler signals. Simultaneous measurements of the dual parameters might open a new clinical value in the pathogenesis of vascular disease and early prevention of atherosclerosis.