The term radiosurgery refers to very focussed delivery of ionizing radiation in order to treat a target region while sparing surrounding tissue. With the robotic CyberKnife system (Accuray Inc.) a large number of non-coplanar, non-isocentric beams can be generated. While the systems allows for excellent conformality and flexible treatments, the planning problem is challenging. We study various methods to further improve treatment plan optimization.
We investigate treatment planning methods, in which the human planner actively interacts with a virtual three-dimensional scene to locally affect the planning outcome, by manipulating the isodose surfaces.
Robotic radiosurgery is characterized by a large number of beams delivered by a robot mounted linear accellerator. While the fexibility in beam delivery allows for very tumor conformal dose distributions it is challenging during treatment planning, where a large optimization problem needs to be solve. We develop efficient beam generation approaches considering the patient geometry and the planning objectives.
While active beam motion can be used to compensate for the most severe artifacts of organ motion, changes in the target shape or in the motion pattern need to be addressed by adaptive planning methods. 4D planning refers to pre-fraction adaptation to the prevalent motion pattern, and 5D planning to intra-fraction adaptation.