(Wednesday, 8/1/2018) 4:30 PM - 6:00 PM
Room: Room 207
Purpose: To develop a novel system for patient-specific combined optimization of couch, collimator, and gantry start/stop angles for VMAT treatment planning in order to reduce OAR doses and produce highly compact dose distributions from optimally sampling the 4Ï€ space.
Methods: By merging objective functions that optimize the couch rotation and collimator angles, a 3D cost space (the CODA cube) can be constructed with axes of gantry, couch, and collimator angles. At each voxel in this cube, the cost of implementing a combination of axes positions in a trajectory is quantified. The cube is sampled and explored using a modified constrained Bellman-Ford algorithm to suggest low-cost fixed candidate arcs on each plane of the space, from which 10-arcs are chosen using k-means clustering. These arcs are then imported into the Eclipse treatment planning system (v11) and inverse-optimized. Six plans with three to four targets were generated from a test patient with eight artificial cranial targets. The CODA plans were compared to standard 4-arc VMAT cranial SRS plans for the same sets of targets; maximum dose to each OAR and total MUs were compared. Both planning methods were inverse-optimized with identical objectives.
Results: Compared to standard 4-arc VMAT, the CODA plans resulted in a mean reduction in maximum OAR dose of 20.6% (p < 0.01), with brainstem dose decreased by 2.63 Gy (p = 0.031). Mean reduction in total MU was 8.6% (p = 0.16) and a mean decrease in normal brain V12 was 3.9% (p = 0.16), when compared to standard 4-arc VMAT methods.
Conclusion: The optimization of couch, collimator, and gantry angles simultaneously using a 3D optimization space achieves a statistically significant sparing to OAR maximum doses. The synergistic combination in optimization of multiple axes may yield superior results to independent optimization of the individual axes.
Funding Support, Disclosures, and Conflict of Interest: Components of this research have been licensed by Brainlab AG; This research has been financially supported by the Atlantic Canadian Opportunities Agency.
Keywords
Stereotactic Radiosurgery, Optimization, Treatment Planning
Taxonomy
TH- External beam- photons: VMAT dose optimization algorithms
Contact Email
