Program Information
A Plan Comparison Tool to Ensure Robustness and Deliverability in Online-Adaptive Radiotherapy
P Hill*, Z Labby, R A Bayliss, M Geurts, J Bayouth, University of Wisconsin, Madison, WI
Presentations
SU-F-BRB-7 (Sunday, July 12, 2015) 4:00 PM - 6:00 PM Room: Ballroom B
Purpose: To develop a plan comparison tool that will ensure robustness and deliverability through analysis of baseline and online-adaptive radiotherapy plans using similarity metrics.
Methods: The ViewRay MRIdian treatment planning system allows export of a plan file that contains plan and delivery information. A software tool was developed to read and compare two plans, providing information and metrics to assess their similarity. In addition to performing direct comparisons (e.g. demographics, ROI volumes, number of segments, total beam-on time), the tool computes and presents histograms of derived metrics (e.g. step-and-shoot segment field sizes, segment average leaf gaps). Such metrics were investigated for their ability to predict that an online-adapted plan reasonably similar to a baseline plan where deliverability has already been established.
Results: In the realm of online-adaptive planning, comparing ROI volumes offers a sanity check to verify observations found during contouring. Beyond ROI analysis, it has been found that simply editing contours and re-optimizing to adapt treatment can produce a delivery that is substantially different than the baseline plan (e.g. number of segments increased by 31%), with no changes in optimization parameters and only minor changes in anatomy. Currently the tool can quickly identify large omissions or deviations from baseline expectations. As our online-adaptive patient population increases, we will continue to develop and refine quantitative acceptance criteria for adapted plans and relate them historical delivery QA measurements.
Conclusion: The plan comparison tool is in clinical use and reports a wide range of comparison metrics, illustrating key differences between two plans. This independent check is accomplished in seconds and can be performed in parallel to other tasks in the online-adaptive workflow. Current use prevents large planning or delivery errors from occurring, and ongoing refinements will lead to increased assurance of plan quality.
Contact Email: