A comparison of multidimensional flow estimation techniques using computational fluid dynamics: Speckle tracking versus vector Doppler

Vascular territories susceptible to atherosclerosis often show regions of disturbed flow. Hence, non-invasive imaging of complex blood flow patterns can ameliorate the understanding of atherogenesis, as well as improve early diagnosis of cardiovascular disease. In this work, a simulation environment is presented, where established and experimental ultrasound imaging methods can be thoroughly investigated towards a ground truth for realistic flow patterns. By coupling computational fluid dynamics (CFD) and ultrasonic simulations (Field II), RF-signals resulting from realistic flow in a 3D carotid bifurcation model were obtained, and used for the evaluation of established and experimental velocity estimators: the 1D autocorrelation method (CFI), speckle tracking (ST), and vector Doppler (VD). Simulations showed that CFI failed to portray complex swirling flow in the stenosed bifurcation. Both ST and VD were able to resolve also the lateral velocity component for the complex flow under ideal imaging conditions. Further work will focus on an indepth investigation of the estimators, including common effects that influence estimator performance such as noise, clutter, and vessel depth and angle.