Probabilistic finite element method for large tumor radiofrequency ablation simulation and planning

Area of Science:

• Biomedical Engineering
• Computational Biology
• Medical Physics

Background:

• Radiofrequency ablation (RFA) is crucial in liver surgery, but accurately predicting lesion formation is challenging due to tissue property variations.
• Large tumors (>10 cm) exacerbate the difficulty in estimating RFA lesion dimensions.

Purpose of the Study:

• To develop a probabilistic bio-heating finite element (FE) model for predicting RFA lesion shapes and sizes in liver tissue.
• To incorporate uncertainties in thermal-electrical properties to generate confidence levels for lesion predictions.

Main Methods:

• A probabilistic bio-heating finite element (FE) model was developed to simulate RFA lesion formation.
• The model accounts for uncertainties in five key thermal-electrical liver properties: thermal conductivity, density, specific heat, blood perfusion, and electrical conductivity.
• The mean-value first-order second-moment (MVFOSM) method was integrated to generate confidence levels for lesion shapes and sizes.

Main Results:

• The probabilistic FE model successfully generated confidence levels for RFA lesion shapes and sizes.
• A workflow for RFA planning was introduced, enabling 3D visualization of predicted lesions within the liver.
• The method allows for interactive simulation and confidence level selection for precise RFA needle placement.

Conclusions:

• The proposed probabilistic FE model enhances the accuracy of RFA lesion prediction in liver surgery.
• This approach provides clinicians with a valuable tool for preoperative planning and optimizing RFA needle placement.
• The integration of uncertainty quantification improves the reliability of RFA treatment planning, especially for large tumors.