Optimal phased array pattern synthesis for non-invasive cancer ablation of liver tumors using high-intensity focused ultrasound

In recent years, phased arrays have been employed with reasonable success to apply high intensity focused ultrasound (HIFU) beams to tumors surrounded by homogeneous media to achieve a desired thermal treatment. In the case of tumors surrounded by inhomogeneities, such as liver tumors shadowed by the rib cage, the focusing scheme is highly complicated and has not been addressed in the literature. The problem stems from the need to find an efficient focusing scheme capable of generating well behaved interference patterns in the interior region of the rib cage concurrently with minimizing power deposition over the rib surfaces. In this work, high frequency techniques are employed to predict the ultrasonic fields within the computational domain. Two synthesis techniques are developed and implemented for single and multiple focus pattern synthesis. After experimental validation of the proposed computational and synthesis approaches, realistic modeling of the problem geometrical features and parameters is considered, analyzed, and tested for single and multiple focusing: schemes. This realistic modeling employs exact scattered field representations to predict the ultrasonic fields within the computational domain. Utilizing the proposed computational and synthesis approaches, design curves for selecting the optimal treatment frequency have been constructed, allowing practioners to design treatment plans according to the tumor location and the tissue attenuation. This thesis demonstrated the feasibility of achieving deeply localized HIFU treatment in the presence of the rib cage while minimizing direct power incidence over the rib surfaces, and achieving optimal phased array pattern features. A major conclusion of the thesis is that high frequency ray approaches can be efficiently employed for predicting the ultrasonic fields within the domain of interest.