Ultrasound imaging of blood flow based on high frame rate acquisition and adaptive signal processing

• Main Author: Ekroll, Ingvild Kinn
• Format: Doctoral Thesis
• Language: English
• Published: Norges teknisk-naturvitenskapelige universitet, Institutt for sirkulasjon og bildediagnostikk 2013
• Online Access: http://urn.kb.se/resolve?urn=urn:nbn:no:ntnu:diva-20512; http://nbn-resolving.de/urn:isbn:978-82-471-4203-5 (printed ver.); http://nbn-resolving.de/urn:isbn:978-82-471-4205-9 (electronic ver.)

Ultrasound imaging of blood flow is in widespread use for assessment of atherosclerotic disease. Imaging of the carotid arteries is of special interest, as blood clots from atherosclerotic plaques may follow the blood stream to the brain with fatal consequences. Color flow imaging and PW Doppler are important tools during patient examination, providing a map of the mean velocities in an image region and the full velocity spectrum in a small region of interest respectively. However they both suffer from limitations which may hamper patient diagnostics. Recent technological advances have enabled an increased acquisition rate of ultrasound images, providing possibilities for further improvement in robustness and accuracy of color flow and PW Doppler imaging. Based on these advances, we aimed to utilize the high acquisition rate to enable robust vector Doppler imaging, where both velocity magnitude and direction is estimated. Additionally, we wanted to incorporate information from several parallel receive beams in spectral Doppler, which is currently limited to velocity estimation in a limited region of a single beam. Two limitations in conventional PW Doppler are especially considered, namely the trade-off between temporal and spectral resolution, and the increased spectral broadening in situations of high velocity or large beam-to-flow angles. By utilizing information from several parallel receive beams, we show that by applying adaptive spectral estimation techniques, it is possible to obtain high quality PW Doppler spectra from ensembles similar to those found in conventional color flow imaging. A new method to limit spectral broadening is also presented, and we show spectra with improved resolution and signal-to-noise ratio for a large span in beam-to-flow angles. Plane wave vector Doppler imaging was investigated using both realistic simulations of flow in a (diseased) carotid artery bifurcation, and in vivo studies. It was found that the plane wave approach could provide robust vector velocity estimates at frame rates significantly higher than what is found in conventional blood flow imaging. The technique was implemented in a research ultrasound system, and a feasibility study was performed in patients with carotid artery disease. Promising results were found, showing an increased velocity span and the successful capture of complex flow patterns. All together, the proposed techniques may provide more efficient clinical tools for vascular imaging, as well as quantitative information for research into new markers for cardiovascular disease.