Dynamic velocity vector and relative pressure estimation in the left ventricle with dynamic contrast-enhanced ultrasound of low frame rates

Abstract

Clinical studies indicate that dysfunction of the left ventricle occurs at an early stage of heart-failure. As ventricular dysfunction and cardiac disease affect the intraventricular flow patterns, analysis of blood flow and relative pressure may aid diagnosis. Several visualization tools exist that are generally based on speckle-tracking techniques using ultrafast ultrasound acquisitions. Yet, there is a need for techniques that can deal with low frame rates, in particular when envisioning future translation to 3D imaging. Here, blood flow velocity and relative pressure are estimated with low-frame-rate dynamic contrast-enhanced ultrasound (DCE-US). Different from standard speckle-tracking techniques, here we propose a method based on estimation of time-delays between indicator dilution curves measured from a set of neighboring pixels in the DCE-US video. An improved estimation of the velocity vectors is then obtained by Navier-Stokes regularization. Using the regularized velocity vectors, relative pressures can be calculated. Blood flow patterns were characterized in terms of vorticity, quantified by the curl of the velocity vectors, and changes in relative pressure (dp/dt) around the aortic valve, since these are linked to the ejection efficiency. Eighteen patients, who underwent cardiac resynchronization therapy (CRT) (9 for CRT-responder and 9 for CRT-non-responder), were tested with the algorithm. In the responder group, vorticity and peak dp/dt were significantly increased after CRT, while no significant difference was found in the non-responder group.