Evaluation of a numerical simulation for multibipolar radiofrequency ablation

Area of Science:

• Medical Physics
• Biomedical Engineering
• Computational Biology

Background:

• Hepatic radiofrequency ablation (RFA) is a key treatment for liver tumors.
• Accurate prediction of ablation zones is crucial for treatment efficacy and safety.
• Existing models often struggle to incorporate complex physiological factors like vascular cooling.

Purpose of the Study:

• To assess the accuracy of a custom numerical simulation for hepatic multibipolar radiofrequency ablation (mbRFA).
• To evaluate the model's ability to predict vascular cooling effects using an ex vivo porcine liver model.
• To validate the simulation against experimental data under standardized conditions.

Main Methods:

• Developed a GPU-accelerated finite-difference time-domain (FDTD) simulation.
• Coupled quasi-static Maxwell and Pennes' bioheat equations to model mbRFA.
• Performed ex vivo mbRFA on a porcine liver model with standardized blood flow simulation.

Main Results:

• The numerical simulation achieved 96% conformity between simulated and experimental ablation areas.
• High similarity scores (Dice: 0.92, Jaccard: 0.85) were observed between simulated and actual ablations.
• The model accurately accounted for vascular cooling effects, with minor overestimation (4%) and underestimation (10%) of the ablation zone.

Conclusions:

• The custom numerical model demonstrates strong spatial agreement with experimental hepatic mbRFA.
• The simulation is suitable for quantitative prediction of ablation geometry, including complex vascular influences.
• This validated model can aid in optimizing RFA treatment planning and outcomes.