GPU accelerated simulation of the temperature field in RFA developed by NE Scientific. In the illustration is visible an electrode for RFA with deployable tines for capturing a larger volume of tissues. Simulation and visualization of the tissue temperatures allows physicians to understand visually which tissues will be necrotized and which not.

NE Scientific is active in the development of a computerized system for guidance of percutaneous ablation of liver cancer. Liver cancer is the 3rd leading cause of cancer mortality. Radio Frequency Ablation (RFA) is a technique based on the insertion in the tissues of a needle-shaped electrode, through which RF energy is applied to the tissues and a volume of tissues of 1cm to 5cm in diameter is killed. The therapeutic effect is achieved if all the malignant tissues, ans a safety margin around them, are completely necrotized.

While RFA is preferable to surgery, as less invasive, and preferable to chemotherapy, as it does not present side-effects, it is challenging for physicians in RFA to consistently necrotize all the tissues that need to be treated. For smaller tumors (less than 2cm in size) RFA is highly successful, but for medium sized tumors (2cm to 5cm in size) in about 27% of the cases the physicians are not able to completely kill all the malignant tissues [1], leading to recurrence of the cancer.

This illustration shows a simulated 14mm tumor (red) and the volume of the ablation (yellow) computed for a 3cm Boston Scientific LeVeen RFA electrode. Three dimensional visualization of the ablation volume allows an immediate evaluation of whether all the malignant tissues are treated or not, allowing physicians to determine the optimal placement of the electrode with respect to the tumor.

One idea to improve outcomes of RFA is that of simulating the physics taking places in RFA (electrical and thermal) and to use this information to indicate superimposed to CT images of the patient which tissues have been killed and which not to any point in the procedure, making it straightforward for physicians to identify which tissues still need treatment.

While attractive, this idea has never been practical as simulation of RFA required ten of minutes to hours of computing time. NE Scientific, leader in computer simulations, was the first entity in the world to demonstrate better than real-time performance in simulating RFA physics. In 2014 NES has developed a library able to compute the volume and geometry of RFA ablations in less than 3 minutes [2] (now, with further optimizations, 45 seconds). Based on these results we have been awarded an SBIR Phase I grant from the National Cancer Institute allowing further refinement and validation of this library. Having successfully completed the SBIR Phase I, we applied last week for Phase II funding. This funding will allow us to build a full clinical system and to validate in a clinical trial. It is expected that by clearly indicating to physicians, on screen and in a three-dimensional way, which tissues might still need treatment, it will be much easier for them to treat all the tissues that need treatment with consistency – and that recurrence rates might be reduced to 5-10% from the current rate of 27%.

REFERENCES:

[1] Kim, Y. S., Lee, W. J., Rhim, H., Lim, H. K., Choi, D., & Lee, J. Y. (2010). The minimal ablative margin of
radiofrequency ablation of hepatocellular carcinoma (> 2 and< 5 cm) needed to prevent local tumor progression:
3D quantitative assessment using CT image fusion. American Journal of Roentgenology,195(3), 758-765.

[2] Borsic, A., Hoffer, E., & Attardo, E. A. (2014, April). GPU-Accelerated real time simulation of Radio
Frequency Ablation thermal dose. In 2014 40th Annual Northeast Bioengineering Conference (NEBEC) (pp. 1-
2). IEEE.