(Sunday, 7/29/2018) 4:00 PM - 4:55 PM
Room: Karl Dean Ballroom A2
Purpose: Volumetric modulated arc therapy (VMAT) is a highly utilized form of external radiation therapy. During planning optimization however, there is no explicit control of aperture shape, which may lead to the formation of irregular aperture configurations. Complex apertures can result in avoidable dosimetric inaccuracies around aperture edges. Here, we present enhancements to the aperture control objective algorithm. We compare our aperture control penalization to a standard complexity metric that penalizes excessive amounts of perimeter per unit of area, prognosing possible dosimetric inaccuracies.
Methods: A previously developed column-generation based heuristic VMAT algorithm was enhanced. Motivated by a possibility to improve upon treatment quality, we initially assigned control points distant and unrelated from each other. In order to improve upon this solution; iterations of aperture refinement were performed. Each iteration sequentially removes a set of control points with low contribution to treatment quality, and replaces them with more suitable aperture shapes.
Results: Comparisons of aperture complexity measurements suggest consistency between our edge metric and standard aperture shape irregularity measures that use perimeter per unit of area. The correlations being initially 0.61, and increasing to 0.72 when our penalization is enforced. Our experiments on a large problem show that we can reduce the standard edge metric penalty by 20% with negligible decrease in treatment quality as measured by DVHs. We anticipate that comparable reductions in dosimetric inaccuracies are achieved due to the nature of the standard metric. The runtime of our python implementation is 4 hours per iteration refinement due to the large size and density of control points in the problem, but there is ample room for parallelizations.
Conclusion: A previously developed algorithm that allows aperture control for VMAT has been enhanced. We tested it on a very large problem and achieved results consistent with reduction of dosimetric inaccuracies.
Funding Support, Disclosures, and Conflict of Interest: Martha Matuszak has a research grant from Varian Medical Systems, unrelated to the current work. Kelly Young is listed in the same grant. This project is funded in part by NIH P01-CA059872
Keywords
Radiation Therapy, Optimization, Inverse Planning
Taxonomy
TH- External beam- photons: VMAT dose optimization algorithms
Contact Email
