A methodology for thermal dose model parameter development using perioperative MRI.

Area of Science:

• Medical Imaging
• Oncology
• Biophysics

Background:

• Post-treatment imaging is crucial for evaluating thermal ablation lesions.
• Magnetic resonance temperature imaging (MRTI) offers real-time feedback but requires a thermal dose model for damage estimation.
• Existing models use empirical parameters not directly reflecting post-treatment appearance.

Purpose of the Study:

• To investigate a novel technique for deriving thermal dose model parameters from perioperative MR data.
• To tailor thermal dose models to the radiologic appearance of thermal lesions on contrast-enhanced imaging.
• To improve the accuracy of thermal lesion assessment after image-guided thermal ablation.

Main Methods:

• Analysis of perioperative MR data from five patients undergoing magnetic resonance-guided laser-induced thermal therapy (MRgLITT).
• Segmentation of the enhancing ring on post-treatment T1-weighted imaging and registration into MRTI geometry.
• Modeling post-treatment appearance using a coupled Arrhenius-logistic model and non-linear optimization to determine kinetic parameters and dose thresholds.

Main Results:

• Derived parameter values and thresholds were consistent with previous studies.
• The average difference between predicted and segmented lesion boundaries was approximately one pixel (1 mm).
• Optimized model predictions for lesion areas were within 1 mm of clinically used dose models.

Conclusions:

• This technique enables the use of clinically acquired data to refine thermal dose models.
• The developed model accurately predicts thermal lesion boundaries based on radiologic appearance.
• This approach offers a clinically relevant endpoint for improving thermal ablation treatment evaluation.