Requirements and prototype for supporting the planning of patient specific thermal ablation interventions

Abstract

Background Thermal ablation is the process of destroying pathological tissue by either high temperatures of approximately 105 °C as achieved in radiofrequency ablation or low temperatures of approximately - 40 °C as used in cryotherapy. Ablations are widely used in clinical practice and provide a safe and generally well tolerated minimal invasive treatment if surgery is not an option. Thermal ablations are usually performed under image guidance, either by ultrasound, CT or MR. Even though ablations are widely used, very little textbook knowledge is available.<br />Because of the treatment complexity there is a need for a well defined process which can be followed by an experienced radiologist as well as an inexperienced one. There is also a need for a planning platform which is capable of supporting the physician in planning the intervention on the basis of the patient's anatomy. For additional benefit this platform should also provide the means for estimating the final coagulation zone by simulations based on the patient's anatomy. The most widely used method to simulate the extend of a coagulation zone is by the usage of finite element analysis (FEA). FEA uses a defined geometry with the physical properties of the tissue and the ablation modality to create a model which can then be solved to make estimations about the extend of the final coagulation zone. Method and Results To deal with the problem of ablation knowledge being only available in distributed form, a workflow was abstracted and translated into diagrams. These workflow diagrams visualize the required steps and decisions when performing thermal ablations. The workflow is split into a planning, applicator placement, ablation and result evaluation phase.<br />The information gained from this knowledge is then used to define the requirements for a platform which is capable of helping the physician when performing the ablation. In the next step I examined the possibility to increase an ablation's coagulation zone volume by using whole-body hyperthermia to heat the patient. This study was done using FEA by comparing the outcome of different models with an initial defined body temperature of 36.7 °C, 37.0 °C, 37.5 °C and 40 °C respectively.<br />The study shows that there is little difference in coagulation zone dimensions with body temperatures in the normal range (36.7 °C to 37.5 °C) but show a significant increase (5.4 mm in the short axis, 3.4 mm in the long axis) in initial body temperatures of 40.0 °C. Since one of the most prevalent shortcomings of FEA simulations of ablations is low accuracy when estimating the effect of microvascular perfusion during ablations, another FEA project was set-up to improve this situation. A algorithm using a first-order kinetic Arrhenius model based on current experimental research was developed to increase simulation accuracy. The results gained from this study show a temperature profile, which is closer to in-vivo ablation results when compared with previous studies.<br />Based on the ablation observations and interviews with physicians requirements for a platform for supporting physicians during ablations where collected. Requirements were split into non-functional, data acquisition, image processing, visualization and treatment simulation requirements. These requirements were the foundation for a prototype implemented using the Slicer Framework which provides the necessary functionality for providing additional information for ablation planning. User feedback showed, that an accurate coagulation zone simulation is often not feasible during an intervention because of the narrow time-window of such a procedure. A simplified simulation approach were the resulting coagulation zone is approximated using values provided by the vendor was developed. This presents a time saving alternative which received good feedback from clinicians. The proposed method is also well suited for post-treatment coagulation zone evaluation if used with image registration.<br />Conclusion FEA is established as a robust technique for simulation and treatment outcome estimation for ablations. Because of the time required to simulate an ablation using FEA, it is not the technique of choice when used in a clinical setting. Since ablations are a time sensitive process, treatment planning must not impose additional effort on the physician. The simplified approach as defined during the course of the requirements engineering has received positive feedback and may be a suitable alternative to time consuming FEA simulations. Because of the free availability of the base software and the low hardware requirements, this solution is also ideally suited for medical training.