Design of control systems for focused and unfocused intracavitary ultrasound arrays for the thermal treatment of prostate disease

Abstract

Heat transfer in tissue during induced hyperthettnia treatments for prostate disease can be modeled as a multi-input multi-output (MIMO) system. For both focused and unfocused ultrasound arrays, the control of this MIMO system can follow either of two paths. First, a method of system identification can be used to derive a system model from which a gain matrix for a controller can be calculated with methods like that of the linear quadratic regulator (LQR). Second, a trial and error controller can be designed and tuned for each new treatment. Although the first option has clear benefits, this work shows that, with MRI thermometry, the ti.me delay in the temperature feedback loop is sufficient to make any method of system identification difficult within the ti.me constraints of a clinical setting. A proportional plus integrator plus derivative (PID) controller was designed to control the system in computer simulations. This single-input, single-output (SISO) controller was weighted to better control the system by providing parameters scaled to compensate for the geometry of the treatment and thermal coupling of tissue. While the controller performed very well under the conditions for which it was tuned, the level of performance decreased when the parameters of the simulation, such as the rate of perfusion in the tissue were altered. For the unfocused array, a program was written in C++ to run the array and implement the controller for a hyperthermia treatment using MRI thermometry for the feedback signal.