Nonlinear model predictive thermal dose control of thermal therapies: experimental validation with phantoms

A novel approach to the control of thermal therapies is evaluated experimentally using a one-dimensional agar phantom. According to the proposed approach, the thermal dose delivered to the target is controlled directly without attempting temperature control. The control problem is constrained by the transducer saturation constraints, and the constraints on the maximum allowable temperature in the normal tissue surrounding the target. The proposed thermal dose controller has a cascade structure with a linear constrained model predictive temperature controller in the inner loop, which receives the reference commands from a nonlinear thermal dose controller in the main loop. The controller is designed to allow near time-minimal control of the treatment. A single, fixed, focused ultrasound transducer is used as a heating modality. The ultrasound power deposition and the perfusion in the phantom are identified experimentally, and used in the developed controller. The ability of the controller to deliver the desired thermal dose to the specified tumor region without violating the temperature constraints in the surrounding normal tissue was evaluated using a series of experiments. The results demonstrate that the proposed approach is effective at delivering the desired thermal dose in a near minimal treatment time without violating the constraints on the maximum allowable temperature in normal tissue. The experiments show that the controller can effectively operate even with substantial plant-model mismatch.