Offset-Free Model Predictive Temperature Control for Ultrasound-Based Hyperthermia Cancer Treatments

Heating cancer cells over an extended period of time, referred to as hyperthermia, has been proven to enhance the effects of chemotherapy and radiotherapy without inducing additional toxicity or undesirable side effects, and is therefore considered a highly valuable adjuvant therapy in cancer treatment. In this work, a model predictive control (MPC) setup is developed for improving performance and robustness in regulating the temperature for magnetic-resonance-guided high-intensity focused ultrasound (MR-HIFU) hyperthermia treatments. The proposed control design incorporates a disturbance estimator as encountered in offset-free MPC that is able to remove the steady-state temperature error caused by plant-model mismatch. For the considered healthcare application, such modeling errors are inevitable in practice due to the high variability of tissue properties in patients, some of which even exhibit time- and temperature-dependent behavior due to the body’s thermoregulatory response, combined with the fact that extensive model identification is undesirable in the clinic. The controller’s performance is demonstrated by means of
in vivo
experiments on a porcine thigh muscle using a clinical MR-HIFU treatment setup.