Room: AAPM ePoster Library
Purpose: In proton therapy, density changes can severely jeopardize the effective delivery of the planned dose distribution. The dose distortion effects can be compensated by treatment adaptation. In this study, the clinical benefit of an automatic on-line adaptive strategy called dose restoration (DR) is investigated. In general, DR aims at compensating proton range distortions by accurately reproducing the planned dose in repeated-CTs, without considering re-contouring. Our objective is to assess to what extent DR can replace the need for a comprehensive off-line adaptive strategy and to estimate the impact of neglecting anatomical deformations.
Methods: Robust fully-automated DR was evaluated on 14 lung cancer patients characterized by a planning-4DCT and two repeated-4DCTs(1rCT,2rCT). Initial plans were 4D-robustly optimized (including breathing motion, setup and range errors). DR used isodose contours from the reference dose with minimum and maximum objectives to mimic the initial dose in repeated-CTs. Robustness evaluations were performed for the initial, not-adapted and restored (adapted) plans in two contours sets: the copy-pasted original structures (RS_ref) and the re-contoured real structures (RS_real).
Results: Nominal and worst-case values resulting from DVH-bands showed an overall improvement of DVH-metrics and optimal robustness levels in restored plans, with respect to not-adapted plans for both contours sets. RS_real presented higher variability and number of outliers due to the imperfect overlapping between both structure sets. According to CTV coverage criteria (D95%>95%Dprescription) in the nominal scenario, 35% (5/14) of the cases needed adaptation. After DR, median(D95%) was increased by 1.5 Gy and only 1 patient out of 14 (less than 10%) still needed off-line adaptation because of important anatomical changes.
Conclusion: Although the results are less spectacular in RS_real, it was shown that DR meets its objective of maintaining CTV coverage within clinical limits and reducing the off-line adaptation rate.
Not Applicable / None Entered.