NE Scientific (NES) has received SBIR funding from the National Cancer Institute and is collaborating with physicians from Dartmouth College to improve the outcomes of Radio Frequency Ablation (RFA). Last week NES has filed a non-provisional patent application with the US Patent Office to protect core technologies that will aid physicians in performing RFA to treat cancer. In this post some of these technologies are illustrated.
RFA is a thermal form of ablation, where RF energy is applied to tissues through needle-shaped electrodes, heating and necrotizing the tissues. This permits minimally invasive approaches, where the electrode is inserted through the skin into the volume of the tumor. The goal is to completely treat the tumor and a margin of tissues (e.g. 5mm) around the tumor. Complete treatment of the target tissues (tumor plus margins) reduces chances of local recurrence to a minimum. One challenge that physicians face is that there are no good ways currently to ensure that all target tissues have been necrotized (adequacy of the procedure). For liver tumors of a size between 3cm and 5cm the probability of a local recurrence is 24% . This less than perfect outcome is caused primarily by incomplete narcotization of target tissues – where physicians are not able to evaluate by imaging means whether all the target tissues have be treated. NES is addressing this problem by developing electrical and thermal models of RFA physics which are able to predict the expected volume of treated tissues; by using these predictions it is possible to highlight on CT images which tissues have been treated and which tissues still need treatment during an intervention, so that it is immediately clear to physicians whether some target tissues still need treatment. This computer aided evaluation of adequacy will in principle guarantee that no target tissues are left untreated, and result in potentially much improved outcomes.
A primary application of RFA is treatment of liver cancer. In this procedure an electrode is inserted in the tissues through the skin, and Computed Tomography (CT) guidance is used to acquire a few images that for positioning the electrode in relation to the tumor volume. The image below (Fig. 1) shows a tumor appearing as a white shadow in a CT image. Normally a liver tumor is not visible in CT images, but intravenous contrast enhances the tumor, and makes it visible. Contrast is also needed to differentiate between which tissues have been treated and which not.
As contrast is toxic, physicians are limited in the amount they can administer to a patient. In a typical procedure a few (e.g 2 or 3) contrast images are acquired to confirm the position of the RFA electrode with respect to the tumor volume. With the exception of these images, acquired before the RFA electrode is activated, physicians are “blind” with respect to the tumor and with respect to which part of the tumor and tissue margins have been treated. The image below, Fig. 2, shows a computer model of an RFA electrode (light blue). The computer model of the electrode has been automatically generated from the underlying CT image, which captures an actual electrode deployed in the tissues. The electrode is formed by a cannula which deploys a number of thin metallic filaments, called tines, in an umbrella-like fashion. It is to be noticed that the tumor is not visible, as no contrast was used to acquire this image.
It is challenging therefore for physicians to evaluate whether the position of an RFA electrode will result in an ablation which will destroy completely the tumor, or whether multiple electrode positions can be used to perform multiple ablations which will eventually result in the complete treatment of the target tissues. Nor tumor nor the treated tissues are visible in non-contrast images.
Computer modeling can be used to address the above problem. The image below (Fig. 3) shows 3D computer model (darker yellow) of a tumor and 5mm margins around the tumor (lighter yellow) which have been generated from the contrast-CT image of Fig.1. Transferring the generated tumor model to other images, like the image in Fig. 2, allows a visualization of the tumor (model) and electrode (model reflecting the true position of the electrode in the tissues) like in Fig. 3 below. From images like in Fig. 3 it is now possible to appreciate the spatial relationship between the tumor and the electrode, and whether the position of the electrode might be optimal.
Further, the software developed by NES is able to simulate in approximately 30 seconds, by using advanced algorithms implemented on GPUs, the complex electrical and thermal phenomena taking place in RFA, and to accurately estimate the ablation volume accounting for patient specific factors like the heat-sink effect of vasculature and the effect of tissue perfusion. A computed ablation volume for the electrode in Fig. 2 is displayed in Fig. 4 superimposed to a CT image. The image shows in orange and yellow respectively the tumor and margins. It is immediately clear that a single ablation with the selected electrode cannot treat completely the volume of target tissues.
Further supporting visualizations are possible, for example, the volume of the treated tissues can be subtracted from the volume of target tissues, generating views like in Fig. 5. The underlying CT image in Fig. 5 is a non-contrast image, as normally available to the physician at this stage of the procedure, and the tumor is not visible. The computer generated graphics though highlights which tissues still need to be treated, and makes the evaluation of adequacy (have all target tissues been treated ?) a visual and immediate task. In the example of Fig. 5 a single ablation with the selected RFA electrode is not sufficient to treat completely the volume of the tumor (orange) and the volume of the margins (yellow). Multiple overlapping ablations will be needed for the procedure to be adequate.
By providing these tools to physicians NES aims to significantly reduce tumor recurrence in percutaneous RFA. Recurrences originate from untreated tissues inadvertently left behind, as no tools are currently available to physicians for a reliable intraoperative evaluation of which tissues have been treated and which not.
 AJR Am J Roentgenol. 2010 Sep;195(3):758-65. doi: 10.2214/AJR.09.2954. 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. Kim YS, Lee WJ, Rhim H, Lim HK, Choi D, Lee JY.