Validation of Software for Patient-Specific Real-Time Simulation of Hepatic Radiofrequency Ablation

Area of Science:

• Medical Physics
• Interventional Radiology
• Oncology

Background:

• CT-guided radiofrequency ablation (RFA) is a key treatment for liver cancer.
• Accurate visualization of destroyed tissue is challenging for large or irregular tumors, limiting RFA success due to local recurrence.
• Real-time intraprocedural guidance is needed to improve RFA efficacy.

Purpose of the Study:

• To validate a real-time software-simulated ablation volume for intraprocedural guidance during CT-guided RFA.
• To assess the accuracy of the software model in predicting ablation volumes in phantoms and in vivo.
• To compare the software's predictive accuracy against manufacturer-provided ablation charts.

Main Methods:

• Calculated simulated ablation volumes using RFA physics software in 17 agar-albumin phantoms and 6 porcine ablations.
• Compared software-modeled volumes to actual ablation volumes (physical lesions in agar, contrast CT in porcine models).
• Quantified error as the surface distance between true and modeled ablation volumes.

Main Results:

• The simulation demonstrated an average maximum error of 2.8 mm in phantoms, accurately modeling the heat-sink effect of simulated vessels.
• In porcine models, the software's average maximum error was 5.2 mm, significantly lower than the manufacturer's model (7.8 mm, p=0.009).
• The real-time model accurately predicted ablation volumes in phantoms (<3 mm error) and showed superior accuracy to manufacturer maps in vivo.

Conclusions:

• A real-time RFA simulation model accurately predicts ablation volumes, incorporating factors like probe position, energy, and vessel proximity.
• This software offers improved intraprocedural guidance for CT-guided RFA, particularly for complex tumors.
• The validated software has the potential to enhance RFA effectiveness and reduce local tumor recurrence in liver cancer treatment.