Development of an optimised subharmonic dynamic contrast-enhanced ultrasound imaging technique for liver cancer

Abstract

Dynamic Contrast-Enhanced Ultrasound (DCEUS) using ultrasound contrast agents (UCA) is a non-invasive method of imaging the microvasculature, typically employing harmonic imaging. Subharmonic imaging (SHI) detects signals which are generated by the UCA but not by tissue, and holds potential for improving quantification accuracy. However, optimal transmission parameters have not yet been established. An optimised subharmonic imaging dynamic contrast-enhanced ultrasound technique for liver cancer is presented in this thesis. The technique was first developed in a controlled bench-level environment and then translated to a clinical ultrasound scanner. Methodologies: Two variations of the UCA SonoVue® (Bracco, Italy) were investigated in this work: the ‘Native’ UCA, which was made up as per the manufacturer’s instructions, and an ‘Altered’ UCA, which was altered to a narrower size distribution of 1 – 4 µm using the method of decantation. The size distribution of both UCA were measured using a Coulter Counter™ (Beckman Coulter™, Life Sciences, USA). Bench-level bubble acoustic characterisation was carried out using a narrowband through-transmission system, which allowed control of the transmit beam parameters, transmit centre frequency and pulse length. Characterisation was performed across 1.8 – 5 MHz and 1 – 8 cycles. Frequency spectra were determined with the Welch periodogram and corrected for transmit and receive sensitivities of the single-element transducers (Olympus, USA). The subharmonic amplitude was measured at bandwidths defined at half the transmit frequency full-width half-maximum and full-width tenth-maximum. The optimum transmit parameters were translated to the clinical ultrasound system (Aixplorer®, Supersonic Imaging, France). A simple system was used to compare the ability of the system to detect changes in the concentration of the ‘Native’ and ‘Altered’ UCA. An anatomically-realistic perfusion phantom and a hepatic artery attenuation phantom were developed and manufactured. The two phantoms were used to test the ability of the clinical system to implement the optimised subharmonic imaging technique and detect differences in liver tissue. Results: The ‘Altered’ UCA had reduced polydispersity (1 – 4 µm; 99 % / 82 % bubble count / volume, respectively), compared to ‘Native’ (57 % bubble volume 4 – 10 µm). Both UCA generated the highest subharmonic signal when insonated at a pulse length of 3 cycles. The maximum measured ‘Native’ subharmonic signal peaked at a transmit frequency of 1.9 MHz, corresponding to subharmonic at 0.95MHz. This frequency increased in the ‘Altered’ UCA, to the range 2.3–2.5 MHz, bringing the subharmonic centre frequency above 1 MHz. Results supported that there is a lower subharmonic generation threshold when insonated near the resonance frequency, measured at 2.1 / 2.4 MHz for ‘Native’ / ‘Altered’ UCA. The ‘Altered’ UCA exhibited a more linear relationship with concentration in both subharmonic and subharmonic-to-fundamental ratio measurements for the raw and scan-converted images than the ‘Native’ UCA, and had lower error values associated with the measurements. The subharmonic signal measured from the ‘Altered’ UCA was found to be a better indicator of the concentration than that measured for the ‘Native’ UCA: for a given signal level it was more likely that a predicted value based on the linear increase in signal level with increased amount of UCA in an imaged region would be correct, meaning that the ‘Altered’ UCA was more appropriate for use in quantification measurements. Time intensity curves mimicking the perfusion in different types of liver tissues were successfully generated using the perfusion phantom. An improvement due to the newly developed SHI technique was demonstrated, in terms of quantification of signals at depth. The new SHI technique had less shadowing artefacts present as well as better suppression of ‘tissue’ signals generated by the PVAc (poly-vinyl-alcohol cryogel) tissue mimicking material (TMM). The ‘Altered’ UCA solution was found to result in more consistent subharmonic generation which may hold potential for improved quantification accuracy. Through the use of the hepatic liver attenuating flow phantom, it was found that the clinical ultrasound system was not sensitive to subharmonic signals which were significantly attenuated by imaging to clinically relevant depths for liver imaging. Conclusions: The developed SHI DCEUS technique, and research into the influence of the UCA size distribution on both the optimum subharmonic imaging transmit frequency and the potential for improving quantification accuracy with the technique, have clearly indicated the steps to be taken to clinically implement an optimised SHI DCEUS quantification technique for liver disease. By implementing these recommendations, in terms of optimum transmit parameters and producing SHI specific UCA, the use of SHI DCEUS holds potential for indicating differences in perfusion dynamics which may be used in staging disease/tracking response to treatments in liver cancer patients.