An iterative beam placement approach for image guided robotic radiosurgery

Robotic radiosurgery enables the use of a large number of small beams from many directions to precisely deliver a high dose of radiation to a clinical target volume Given the robot's capability to place beams non-isocentric and non-coplanar the number of potential beams becomes very large, and typically a randomized heuristic is employed to select a set of candidate beams considered during inverse planning. We present a new iterative approach based on a more complex heuristic. The new heuristic uses information from the current plan to identify beams that could improve the plan quality. Our results show that the new method leads to improved results compared with randomized planning.