Optimization of Clinical Radiofrequency Hyperthermia by Use of MR-Thermography in a Hybrid System

The method to acquire MR-temperature datasets by the proton-resonance-shift method is outlined and verified in phantoms and patients. An online adaptation process has been developed to achieve agreement between planning calculations and MR-temperature measurements. This is used as a basis to optimize the pattern by a control loop. This procedure is successful after the second iteration step in phantoms. In patients an increase of SAR (specific absorption rate) in the tumor relative to the surroundings has been demonstrated for the MR-temperature increase as optimization variable. The optimization results are even improved under clinical conditions, if perfusion and thermal conduction are considered during the optimization procedure. The mathematical background is presented.