Michael Schwenke

• Focused ultrasound (FUS) is a noninvasive method for tissue ablation that has the potential for complete and local tumor destruction with minimal side effects. Already being used for the treatment of static organs, compensating target motion is not yet clinically available due to the complexity of the treatment. We here propose a numerical model of the therapy effects during respiratory motion to study FUS for moving liver targets. A focus lies on incorporating the motion and the computational efficiency of the simulations. A temperature model is proposed predicting the temperature distributions efficiently on the graphics processing unit by mapping the problem from the moving physical world to a static motion reference state. We also investigate the accuracy of ultrasound modeling in the highly heterogeneous propagation domain including ribs. A novel angular spectrum approach for heterogeneous media is proposed as the widely used hybrid angular spectrum method is found to be ineffective. For real-time applications, we propose an approximate ultrasound propagation model. An integrated FUS model is developed combining these model with an abdominal motion model, tissue damage, and a parameter model. The patient anatomy is automatically derived from CT images. Two clinical use-cases of the integrated model are given: A simulation-driven planning tool allows a clinician to interactively explore treatment options. And a study is performed using the model to optimize the placement of the FUS device. The model is furthermore used to study a novel motion-compensation FUS treatment system by replacing hardware and patient by model predictions. We estimate the efficiency of the treatment system in combination with a clinically available FUS device and MR imaging device (6.67 Hz image rate, 20 Hz FUS control rate) to be above 80%. This estimated efficiency of the new treatment system is expected to be already suited for clinical applications.