| Summary: | Focused ultrasound surgery (FUS) or high intensity focused ultrasound (HIFU), is a promising technique for less- or non-invasively destroying unhealthy tissue deep inside the body, without damage to the skin or surrounding tissues. The procedure has been performed under both diagnostic ultrasound and MRI guidance. Treating cancers and metastases in the liver that are unresectable is a potential application for FUS. However the respiratory motion hindered FUS treatment of liver to become a completely non-invasive technique. The method is currently limited to breath-hold treatments under general anaesthesia that is uncomfortable for patients. The purpose of this study is to investigate key issues of US and MRI guided real-time target ablation when the target is in free breathing motion state which is similar to human liver motion. For the ultrasound guided focused ultrasound (USgFUS), diagnostic ultrasound B-mode image was used to track a moving target. The possibility of using strain sonoelastography to assess FUS lesion formation was explored. Multi-layered tissue mimicking phantoms were designed and fabricated to mimic the graphical features of tumours in human livers in diagnostic ultrasound images. The phantom was then fixed onto three motion setups: 1) controllable 1D reciprocal motion stage, 2) controllable 2D reciprocal motion stage, and 3) ventilator driven balloon to mimic breath motion. Active snake tracking was developed to follow the moving phantom to evaluate the tracking accuracy and speed. This method can achieve a speed of 5~6 frames/second with an error less than 1.0 mm. Strain sonoelastography is selected to assess lesion formation for FUS. Through comparisons of the elastograms between pre- and post-FUS around the focal zone, useful information about the FUS-induced lesions could be extracted from the elastographic artefacts. The performance of elastography to assess FUS lesion in egg-white Polyacrylamide (PAA) phantoms and fresh sheep livers was tested. The FUS lesions in the experiment samples (PAA phantoms and fresh sheep livers) were recognizable under strain sonoelastography after image processing. For MRI guided focused ultrasound (MRgFUS), a moving target with similar graphical features of tumours in human liver was tracked via analysing MRI scans. Then letting the ultrasound beam lock onto a moving target was realized via beam-steering by a phased-array HIFU transducer. An MR compatible robotic arm-INNOMOTION was introduced. A fast localization method was developed to make the robotic arm guided HIFU transducer more efficiently. What is more, it becomes a controllable reciprocal moving setup for investigating the raised issues of MRgFUS for motion tracking in this study. Two normal volunteers were scanned via MR scanner. The data was used to 1) design tissue mimicking phantoms with similar graphical features to the volunteer livers, 2) design respiratory motion simulator based on the estimated liver motion parameters, 3) and develop motion tracking algorithm based on the image features of the volunteer livers. The tissue mimicking phantoms appeared to be similar to the structures of volunteer livers in the MR echo planar imaging (EPI) scans. An experiment setup, in which the tissue mimicking phantoms was controlled to move reciprocally, was designed. The off-line MATLAB algorithm based on cross correlation proved to have an acceptable error less than 1.0 mm. A synchronization system between the target motion and beam-steering was built. Several key problems for motion tracking were studied including how to realize beam-steering with a phased-array transducer, how to map target location in the MR frame to the focus position in the transducer frame, and how to use a step-by-step local sonication series to approximate continuous beam-steering. The system’s performance was tested with a series of sonications, in which temperature rises were compared between when the target was moving with and without tracking. A primary conclusion can be made that tracking could decrease the impact of target movement in focused ultrasound ablation. Tracking could be considered as a compensatory method to liver motion caused by respiration during MRgFUS treatment. In conclusion, the thesis proposed a promising research direction to solve the issue of target motion in FUS treatment of human livers and other abdominal organs. The study achieved the target motion tracking both with diagnostic ultrasound and MRI guidance. The focus steering of HIFU transducer was realized accordingly in the MRgFUS, which can allow the focused ultrasound beam to follow a moving target. The strain sonoelastography had proved to become a potential method to assess FUS lesion formation. This study also brings more issues to be solved, e.g. the noise in diagnostic ultrasound during USgFUS tracking, real-time sonoelastography monitoring lesion formation, and new MRI thermometry that is less susceptible to target motion. The real-time image guided FUS would be more promising by overcoming these technical difficulties.
|
|---|
