Dr.-Ing. Christian Rieder

Patient-specific Planning of RFA

(Dienstag, 08.30 Uhr, Schinkelsaal)

Christian Rieder

Percutaneous radiofrequency ablation (RFA) is a minimally invasive, image-guided therapy for the treatment of primary and secondary liver tumors. Electric energy is locally induced via electrodes in order to destroy all tumor cells by a local resistive heating of the tissue. The success of thermal ablations strongly depends on a well-planned treatment strategy before the intervention. Typically, this planning procedure is done manually by the radiologist. However, the consideration of multiple criteria such as the distance to critical structures and the tumor coverage is a challenging task.

Therefore, a software prototype to plan and optimize patient individually the access path of the inserted RF applicator is presented. To define critical regions for the therapy, segmentation of structures such as ribs, lesions and vessels is integrated. In order to avoid recurrences, the determination of the expected thermal necrosis of the RF applicator has to be considered. It is estimated either utilizing an accurate bio-physical mathematical model or a fast interactive approximation. In both cases, the heat-sink effect of the surrounding blood vessels is taken into account. An optimal access path can be automatically determined under consideration of additional clinical requirements which are defined by the physician. Among other criteria, the distance of possible paths to the extracted risk structures as well as the path length are evaluated based on image processing. The influence of all these criteria can be interactively adapted by the radiologist via some weighting factors. Subsequently, the system evaluates all possible trajectories, which fulfill the specified criteria. Therefore, a mathematical optimization method is utilized in order to determine the best solution under the chosen criteria that affect safety and practicability of the procedure.

The proposed methods are integrated into a clinical software assistant. The radiologist may interact with the system by iteratively adapting the target values. A combined visualization of the patient anatomy, the currently selected access path and the expected coagulation necrosis incorporating the heat-sink effects of the surrounding blood vessels is updated in real-time.

Dr. Christian Rieder
Fraunhofer MEVIS Bremen