CREATING A METHOD OF TREATMENT PLAN SELECTION FROM THE LIBRARY OF PLANS ON THE MR-LINAC

Abstract

ABSTRACT BACKGROUND AND PURPOSE: The Unity 1.5 T MR-Linac is potentially the technique of choice for online Adaptive radiotherapy (ART). The MR-Linac workflows for performing plan need to be made shorter to ensure safe and consistent delivery of radiation dose. The purpose of this work was to examine the feasibility of identifying which of the library of plans (LoP) is most similar to today's anatomy to investigate if the plan from the most similar anatomy can be simply recalculated for today’s treatment. METHODS: A retrospective planning study was carried out with two patients with clinical evidence of prostate cancer. The similarity scores from ADMIRE software were used to highlight treatment plan similarities. The data was extracted from the on-line treatment planning system which stores all the daily adapted plans and daily images. The ADMIRE software was used to deform the clinical contours of the target and organs at risk from the planning CT. The adapted plans for the most, median and least similar plans were then copied across and recalculated, LoP method. The original plan was copied and recalculated on all fractions as a baseline. The analysis of the data were performed using Python. RESULTS: The percentage dose difference of the adapted plans and baseline plans of the least, median, and most similar fractions were compared to the original plan. The results of the LoP method compared to the adapted plans and baseline plans show mixed results. It was expected that as the similarity increased that the plan quality would also increase. Overall, this has not been seen. CONCLUSION: This study generates an accelerated method for treatment plan adaptation without the need for creating a new plan for every treatment fraction. The results presented here are mixed, further investigation and more patient data is needed to find the underlying cause.