Motion correction in contrast-enhanced ultrasound scans of carotid atherosclerotic plaques

Contrast Enhanced Ultrasound has shown great potential in detecting vulnerable atherosclerotic plaques, but an effective motion correction crucial for identifying relevant plaque features is still lacking. In this article, a semi-automated algorithm capable of non-rigidly registering contrast enhanced ultrasound scans of the carotid artery is proposed, employing information about the lumen and its deformation at different phases of the cardiac cycle, in order to segment the lumen and improve the registration process. The proposed registration is compared to state of the art algorithms. Significant improvements can be achieved, especially in the regions within and surrounding the plaques. A 77% reduction has been reported in grayscale variability for a stationary scatter, compared to reference algorithms.