Evaluation of new ultrasound techniques for clinical imaging in selected liver and vascular applications

Abstract

This PhD project is based on a longstanding collaboration between physicists and engineers from the Center of Fast Ultrasound Imaging (CFU) at the Technical University of Denmark and medical doctors from the department of Radiology at Rigshospitalet. The intent of this cooperation is to validate new ultrasonic methods for future clinical use. Study I compares two B-mode ultrasound methods: the new experimental technique Synthetic Aperture Sequential Beam-forming combined with Tissue Harmonic Imaging (SASB-THI), and a conventional technique combined with THI. While SASB reduces the amount of data transformation, thus enabling wireless transmission, THI can improve resolution and image contrast, and creates fewer artifacts. Thirty-one patients with verified liver tumors were scanned and recordings with and without visible pathology were performed. Subsequently, eight radiologists evaluated blinded to information about the technique, which B-mode images they preferred, as well as detection of pathology. Evaluation showed that the techniques were preferred equally and tumor could be detected equally well. Study II deals with the ability of vector flow imaging (VFI) to monitor patients with arteriovenous fistulas for hemodialysis for upcoming stenosis. VFI is an angle-independent method for determining blood flow direction and velocity. Volume can be determined by integrating the velocity profile multiplied by the cross-sectional area. Nineteen patients were monitored monthly over a period of six months, and VFI estimates were compared with the reference ultrasound dilution technique (UDT). VFI volume flow values were not significantly different from UDT and had a better precision. Concordance between VFI and UDT was high when large volume flow changes (greater than 25%) oc curred between dialysis sessions. However, the methods could not be regarded as interchangeable. Study III deals with VFI's ability to determine peak velocity in the portal vein. The commonly used ultrasound method for this is spectral Doppler, which is known to overestimate peak velocity when the angle between the blood vessel and the beam is more than 70 degrees; this overestimation becomes even larger when the angle becomes larger. VFI can determine the peak velocity angle independently. Thirty-two healthy volunteers were scanned with spectral Doppler and VFI with two portal vein scan positions (intercostal and subcostal). The study showed that VFI estimates the same peak velocity as spectral Doppler. Furthermore, VFI has better precision and can estimate the same peak velocity with a scan position, where spectral Doppler cannot. Finally, inter- and intra-observer agreement is higher for VFI. All three studies indicate that the techniques can be used in the clinic and probably will be part of everyday practice in the near future.