Author: Dharmin D. Desai, Ellis Lee Johnson, Philip J. Kallenberg, Kapil M. Mishra π¨βπ¬
Affiliation: Spectrum Medical Physics, University of Kentucky, Varian Advanced Oncology Solutions π
Purpose: The βAsk For Itβ (AFI) manual optimization strategy has been shown to generate high quality SRS plans within Eclipse. The purpose of this work is to directly compare manually generated Eclipse-AFI plans to auto-generated plans by the Brainlab Elements TPS.
Methods: Using the CT data set provided by the Radiosurgery Society, 9 PTVs were chosen for this study ranging in volume from 0.19cm3 to 8.05cm3. Five individual SRS cases were constructed using various combinations of the 9 PTVs (1-target, 2-target, 4-target, 5-target, and 9-target) and planned with both Eclipse and Brainlab Elements. The Eclipse plans were generated using 6xFFF beam and VMAT optimization using the AFI strategy. The Brainlab plans were generated using a Pencil Beam Algorithm, a 6xFFF beam, and 3D conformal arcs utilizing the Arc Trajectory Optimization Tool. The Brainlab DICOM doses were imported into Eclipse for evaluation. The Eclipse-AFI and Brainlab plans were compared via intermediate dose spill (R50%), prescription conformality (CIRTOG and CIPaddick), and PTV prescription coverage. In addition, Wilcoxon signed rank tests were performed to assess the differences in the R50%, CIRTOG, and CIPaddick values.
Results: The Eclipse-AFI generated plans showed improved R50%, CIRTOG, and CIPaddick when compared to the Brainlab plans, while maintaining similar PTV coverage (Eclipse-AFI: mean R50%=6.24, mean CIRTOG=1.26, mean CIPaddick=0.78; Brainlab: mean R50%=7.40, mean CIRTOG=1.58, mean CIPaddick=0.64). Based on Wilcoxon signed rank tests, Eclipse-AFI plans showed improvement at the 0.05 confidence level for all three metrics.
Conclusion: BrainLab Elementβs ATO generates high quality plans very quickly. Eclipse-AFI plans require a high level of technical expertise to produce better intermediate dose spill and plan conformality than plans generated with Brainlab Elements using the same SRS cases. However, this project studied a small number of SRS cases, and a larger study (in progress) is needed to provide a more definitive conclusion.