Model predictive control for MR-HIFU-mediated, uniform hyperthermia.

Area of Science:

• Medical Physics
• Biomedical Engineering
• Thermal Medicine

Background:

• Precise temperature control is crucial for effective local hyperthermia treatments.
• Current methods may lack the voxel-level precision needed for optimal outcomes.

Purpose of the Study:

• To introduce and evaluate a model predictive control (MPC) algorithm for voxel-level temperature management in MR-HIFU.
• To compare the performance of the MPC algorithm against the current state-of-the-art hyperthermia controller.

Main Methods:

• Evaluating model detail's impact on prediction quality and controller runtime.
• Characterizing a tissue-mimicking phantom and assessing modeling errors in silico and in phantom.
• Conducting side-by-side experiments comparing MPC with the current controller.

Main Results:

• A four-voxel heat exchange model provided high predictive performance and real-time suitability.
• MPC performance degraded with erroneous model parameters but was high when models were derived from low-power sonication thermometry.
• MPC demonstrated smaller tracking errors and tighter temperature distributions compared to the state-of-the-art controller.

Conclusions:

• The proposed MPC algorithm, using thermal models from low-power sonications, performs well in phantoms.
• MPC offers superior heating patterns, leading to more stable and uniform hyperthermia.
• This represents a significant advancement for MR-HIFU thermal therapy control.