Compensation for Velocity Underestimation in 2D Super-Resolution Ultrasound

Super-resolution ultrasound imaging (SRI) has shown the potential for visualization of the vasculature with sub-wavelength resolution. Microbubble (MB) tracking in SRI can improve image quality and enable velocity estimation. However, it is usually neglected that the size of the vessel might be relatively small compared to the full width half maximum resolution in the elevation direction (FWHMy) and the velocity estimated by tracking algorithms is usually lower than the true velocity. Considering the small size of desired vessels in SRI, it is hypothesized that the velocity is underestimated in SRI when the MB contributions are averaged in the elevation direction. In this work, a theoretical model is introduced, which shows peak velocity of a 3D parabolic velocity profile underestimated by 33% in 2D imaging when the size of vessel is smaller than the FWHM
y
. The hypothesis was verified using Field II simulations and hydrogel 3D printed micro flow phantom measurement. A GE L8-18i-D linear array transducer with FWHMy of approximately 770 µm at the target imaging depth and a Verasonics Vantage
$256^{\text{TM}}$
scanner were used. Simulation and measurement results were compared to the theory. Simulations with various parabolic velocity profiles showed that the peak velocity was underestimated by 26% ± 5% (mean ± standard deviation), and in the measurement the underestimation was 34% ± 8%. The results were in correspondence to the theoretical value of 33%. Therefore, in this case the theoretical model can be used for compensation of velocity estimates.