Advanced planning and intra-operative validation for robot-assisted keyhole neurosurgery In ROBOCAST

ROBOCAST is a multi-national project comprising several institutes which aim at outlining and implementing a prototype system for advanced, robot-assisted keyhole neurosurgery. This paper reports mainly on software and sensor aspects of the system in the pre- and intra-operative stage. We describe a comprehensive workflow of planning steps provided to the surgeon in a wizard-like manner. As a novelty, we present a method for automatic trajectory planning based on a statistical and patient-specific risk atlas. Intra-operative monitoring of system uncertainty is proposed. Furthermore, the usage of 2D and 3D freehand ultrasound for intra-operative validation is motivated and theoretically outlined. After the first of three years project runtime, we present current work in progress and preliminary results on several patient studies concerning automatic path planning, trajectory localization errors as well as ultrasound imaging on one, six and three patients, respectively. The results underline the usefulness and significance of proposed methods, both within the scope of the ROBOCAST project as well as for conventional keyhole neurosurgery.