Adaptive real-time bioheat transfer models for computer-driven MR-guided laser induced thermal therapy

Area of Science:

• Biomedical Engineering
• Computational Science
• Medical Physics

Background:

• Laser-induced thermal therapies (LITT) require precise treatment times determined by computational models.
• Partial differential equation (PDE)-constrained optimization is crucial for predicting LITT outcomes.
• Real-time patient-specific calibration is needed to adapt models to individual tissue properties.

Purpose of the Study:

• To investigate the convergence behavior of a bioheat transfer constrained calibration problem.
• To assess the feasibility of real-time, patient-specific calibration for LITT.
• To evaluate the accuracy of computational predictions against in vivo thermal imaging data.

Main Methods:

• Utilized multiplanar thermal images from nondestructive in vivo canine prostate heating.
• Developed calibration techniques to adaptively recover biothermal heterogeneities.
• Performed a comprehensive calibration study with varied model parameters and nonlinearities.

Main Results:

• Calibration problems involving thousands of parameters converged within three minutes.
• Reduced the norm of the difference between predicted and measured temperatures to a patient-specific level.
• Demonstrated feasibility for real-time, patient-specific calibration in bioheat transfer models.

Conclusions:

• Real-time, patient-specific calibration for LITT is feasible using PDE-constrained optimization.
• The developed methods can accurately recover tissue biothermal properties.
• This approach enhances the precision and applicability of computational predictions in thermal therapies.