Investigating the fundamentals of IMRT decomposition using ten simple collimator models

The fundamentals of IMRT collimation have been studied using ten conceptual collimators. Spanning a range of complexities from the LINAC jaws alone to a full multi-leaf collimator (MLC), these collimators were designed with two abilities in mind: (1) to be able to define arbitrary field shapes, and (2) to be able to irradiate multiple, disconnected regions in a single segment. The collimators were tested by finding decompositions of random and clinical intensity-modulated beams (IMBs), and collimator performance was measured using both the number of segments required to complete the IMB and the monitor-unit efficiency of the treatment. The decompositions were run on 10 x 10 IMBs with integer bixel values randomly between 1 and 10, and clinical IMBs of varying sizes from lung, head and neck, and pelvic patients taken from a Pinnacle treatment-planning system. Results confirmed that although treatment performance improves with increased collimator complexity, it is not solely dependent on the number of segment shapes deliverable by the collimator but instead on how well these shapes lend themselves to IMRT delivery.