Development of a Treatment Planning Framework for Laser Interstitial Thermal Therapy (LITT)

PURPOSE

Develop a treatment planning framework for neurosurgeons treating high-grade gliomas with LITT to minimize the learning curve and improve tumor thermal dose coverage.

METHODS

Deidentified patient images were segmented using the image segmentation software Materialize MIMICS©. Segmented images were imported into the commercial finite element analysis (FEA) software COMSOL Multiphysics© to perform bioheat transfer simulations. The laser probe was modeled as a cylindrical object with radius 0.7 mm and length 100 mm, with a constant beam diameter. A modeled laser probe was placed in the tumor in accordance with patient specific patient magnetic resonance temperature imaging (MRTi) data. The laser energy was modeled as a deposited beam heat source in the FEA software. Penne's bioheat equation was used to model heat transfer in brain tissue. The cerebrospinal fluid (CSF) was modeled as a solid with convectively enhanced conductivity to capture heat sink effects. In this study, thermal damage-dependent blood perfusion was assessed. Pulsed laser heating was modeled based on patient treatment logs. The stationary heat source and pullback heat source techniques were modeled to compare the calculated tissue damage. The developed bioheat transfer model was compared to MRTi data obtained from a laser log during LITT procedures. The application builder module in COMSOL Multiphysics© was utilized to create a Graphical User Interface (GUI) for the treatment planning framework.

RESULTS

Simulations predicted increased thermal damage (10-15%) in the tumor for the pullback heat source approach compared with the stationary heat source. The model-predicted temperature profiles followed trends similar to those of the MRTi data. Simulations predicted partial tissue ablation in tumors proximal to the CSF ventricle.

CONCLUSION

A mobile platform-based GUI for bioheat transfer simulation was developed to aid neurosurgeons in conveniently varying the simulation parameters according to a patient-specific treatment plan. The convective effects of the CSF should be modeled with heat sink effects for accurate LITT treatment planning.

Validation of a Temperature-Feedback Controlled Automated Magnetic Hyperthermia Therapy Device

We present in vivo validation of an automated magnetic hyperthermia therapy (MHT) device that uses real-time temperature input measured at the target to control tissue heating. MHT is a thermal therapy that uses heat generated by magnetic materials exposed to an alternating magnetic field. For temperature monitoring, we integrated a commercial fiber optic temperature probe containing four gallium arsenide (GaAs) temperature sensors. The controller device used temperature from the sensors as input to manage power to the magnetic field applicator. We developed a robust, multi-objective, proportional-integral-derivative (PID) algorithm to control the target thermal dose by modulating power delivered to the magnetic field applicator. The magnetic field applicator was a 20 cm diameter Maxwell-type induction coil powered by a 120 kW induction heating power supply operating at 160 kHz. Finite element (FE) simulations were performed to determine values of the PID gain factors prior to verification and validation trials. Ex vivo verification and validation were conducted in gel phantoms and sectioned bovine liver, respectively. In vivo validation of the controller was achieved in a canine research subject following infusion of magnetic nanoparticles (MNPs) into the brain. In all cases, performance matched controller design criteria, while also achieving a thermal dose measured as cumulative equivalent minutes at 43 °C (CEM43) 60 ± 5 min within 30 min.