Sparsity-driven super-localization in clinical contrast-enhanced ultrasound

Ruud J. G. van Sloun,Oren Solomon,Yonina C. Eldar,H. Wijkstra,M. Mischi
DOI: https://doi.org/10.1109/ULTSYM.2017.8092875
2017-09-01
IUS
Abstract:Super-resolution (SR) ultrasound enables detailed assessment of the fine vascular network by pinpointing individual microbubbles (MBs), using ultrasound contrast agents (UCAs). The information in SR images is determined by the density of localized MBs and their localization accuracy. To obtain high densities, one can evaluate extremely sparse subsets of MBs across thousands of frames by using a very low MB dose and imaging for a very long time, which is impractical for clinical routine. While ultrafast imaging somewhat alleviates this problem, long acquisition times are still required to enhance the full vascular bed. As a result, localization accuracy remains hampered by patient motion. Recently, Sparsity-based Ultrasonic Super resolution Hemodynamic Imaging (SUSHI) achieved comparable spatial resolution with a sub-second temporal resolution. However, in the current implementation of SUSHI this temporal resolution was achieved using very high frame-rate, e.g. plane-wave imaging, which is not currently widely available in clinical scanners. The aim of this work is twofold. First, to attain a high MB localization accuracy on dense contrast-enhanced ultrasound (CEUS) data using a clinical dose of UCA and a widespread clinical scanner. Second, to retain a high resolution by motion compensation.
What problem does this paper attempt to address?