Christian Hansen

• This thesis contributes to the field of computer-assisted liver surgery with a focus on surgical planning and intraoperative support. Resection planning for liver surgery has become an essential tool in the clinical routine. Based on 3D reconstructions from radiologic images of the patient, different resection plans can be simulated and assessed before an intervention. During an intervention, surgical navigation systems align the planning information with the patient's liver and thus support the realization of a preoperative plan. New techniques for preoperative risk assessment in liver surgery are presented in this work. Thereby, the determination of safety margins around tumors is addressed which is a challenging task for surgeons. The spatial relation of tumors to the intrahepatic vascular anatomy and the amount of remnant liver volume are important factors when deciding whether a tumor-free safety margin can be achieved. Besides a method to define safety margins with uniform width, a method to define non-uniform safety margins is proposed while considering robustness and sensitivity of vascular risk. In addition, techniques for intraoperative support of liver interventions are introduced. First, the intraoperative adaptation of surgical planning data is focused. A software assistant to quickly adapt risk analyses and resection proposals in case of intraoperatively detected findings is presented. In this context, appropriate interfaces for surgeon-computer interaction and an approach for automatic generation of virtual resection surfaces are proposed. Second, new visualization techniques designed for intraoperative use are addressed. Intraoperative visualizations have to follow specific requirements such as the workflow and cognitive load of the surgeon, which together call for a context-driven reduction of complexity and a focus on critical areas. Therefore, surgical risk maps and an approach for illustrative augmented reality are presented. The developed visualization techniques provide a new and objective basis for the assessment of risks during liver surgery. Third, new techniques for auditory support for navigated liver surgery are introduced. Auditory support has the potential to reduce the dependency on visual presentations and freeing the surgeon to focus attention on the situs rather than on a monitor. The transfer of these methods from academic research to applicability in clinical routine was an important goal of this dissertation project. A great value was set on including clinicians in the development process by using their feedback to define design requirements, creating new concepts, and finally evaluating the developed methods in a clinical environment. Therefore, the development was accompanied by a number of clinically oriented tests.