Program Information
A New Intensity Modulation Radiation Therapy (IMRT) Optimizer Solution with Robust Fluence Maps for MLC Segmentation
D Nguyen1*, D O'Connor2 , V Yu1 , D Ruan1 , M Cao1 , D Low1 , K Sheng1 , (1) Deparment of Radiation Oncology, University of California Los Angeles, Los Angeles, CA, (2) Deparment of Mathematics, University of California Los Angeles, Los Angeles, CA
Presentations
TH-EF-BRD-5 (Thursday, July 16, 2015) 1:00 PM - 2:50 PM Room: Ballroom D
Purpose: Converting theoretical fluence maps to deliverable MLC segments while maintaining plan quality has posed a significant challenge for highly complex IMRT plans involving many fields. Rather than simplifying individual fluence maps, we have formulated the problem as a global dose domain fluence map optimization with anisotropic total variation regularization (FMOATVR) problem and then solve the large scale optimization problem.
Methods: The 4π planning platform was used to select and optimize 20 non-coplanar beams for 2 GBM, 2 head & neck, and 2 lung patients with varying prescription doses, including H&N case #2 that had 3 prescription dose levels of 54 Gy, 59.4 Gy, and 69.96 Gy. The plans were re-run using FMOATVR and the same 20 4π beams. The L2 norm of dose deviation from the ideal prescription with zero dose to OARs was used as the objective function, which was subjected to anisotropic total variation regularization to encourage piecewise smoothness in the fluence maps. The fluences were stratified into equal step sizes to create 300 MLC segments. The dose distributions were normalized so the prescription dose covered cover 95% of the PTV. H&N case #2 was normalized to cover the 59.4 Gy PTV, which was included in the statistics.
Results: Compared to the directly stratified 4π plans (4π DS), the 4π FMOATVR method reduced the average mean and max dose to the OARs by 3.54% and 4.15%, respectively, of the prescription dose. Dose coverage was also improved, evidenced by an average D98 and D99 increase of 1.63% and 2.74% of the prescription dose, respectively. The average PTV homogeneity was increased from 0.833 to 0.905.
Conclusion: The global dose domain fluence map optimization formula and its solver provides a straightforward yet potent method to attain high quality while remaining robust to MLC segmentation for complex IMRT.
Funding Support, Disclosures, and Conflict of Interest: This work is supported in part by Varian Medical Systems, Inc.
Contact Email: