REAL-TIME MODELING AND ANALYSIS OF A SCANNING FOCUSED ULTRASOUND HYPERTHERMIA DELIVERY SYSTEM.

Accurate knowledge of power delivered to the tumor and the resulting three-dimensional temperature field is necessary to control hyperthermia and to develop optimal treatment plans. Work in this thesis addresses theoretical and experimental approaches to predict temperature distributions in tumors and in tumor phantoms from limited thermocouple data in real-time and to estimate unknown model parameters during the hyperthermia treatment. Efficient numerical solutions of the bioheat equation are developed that are solvable in real-time on a microcomputer. The accuracy of the numerical solutions is evaluated by comparing numerical solutions to known analytical solutions for simple geometries. Parameter estimation is utilized to converge the predicted temperature distribution with the measured temperature. Phantoms for ultrasonic hyperthermia are examined. Attenuation coefficient, absorption coefficient, conduction velocity, density, and melting point of the phantoms are measured. Errors associated with measurements are presented. A number of different phantom materials are investigated. A scanning focused ultrasound hyperthermia delivery system is implemented to evaluate model performance. Spherical phantoms in a water bath are used as targets. Design considerations for focused ultrasound hyperthermia are addressed. Scanning approaches are investigated and implemented. Modeling and control algorithm results in phantoms are presented