Online adaptive dose restoration in intensity modulated proton therapy of lung cancer to account for inter-fractional density changes

• Published in: Physics and imaging in radiation oncology Vol. 15; pp. 30 - 37
• Main Authors: Borderías Villarroel, Elena, Geets, Xavier, Sterpin, Edmond
• Format: Journal Article
• Language: English
• Published: Netherlands Elsevier B.V 01.07.2020; Elsevier
• Subjects: Adaptive proton therapy; Adaptive radiation therapy; Automated adaptation; Original; Proton therapy; Range uncertainties; Range uncertainties; Proton therapy; Adaptive proton therapy; Adaptive radiation therapy; Automated adaptation
• ISSN: 2405-6316; 2405-6316
• DOI: 10.1016/j.phro.2020.06.004

•An online adaptive workflow based on automatic dose restoration was implemented.•Restoration aims at compensating density changes by reproducing the planned dose.•Our tool mitigates the impact of inter-fractional density changes in proton therapy.•The need of comprehensive offline adaptation is reduced by introducing restoration.•Dose restoration improves target coverage and robustness respect to no adaptation. In proton therapy, inter-fractional density changes can severely compromise the effective delivery of the planned dose. Such dose distortion effects can be accounted for by treatment plan adaptation, that requires considerable automation for widespread implementation in clinics. In this study, the clinical benefit of an automatic online adaptive strategy called dose restoration (DR) was investigated. Our objective was to assess to what extent DR could replace the need for a comprehensive offline adaptive strategy. The fully automatic and robust DR workflow was evaluated in a cohort of 14 lung IMPT patients that had a planning-CT and two repeated 4D-CTs (rCT1,rCT2). Initial plans were generated using 4D-robust optimization (including breathing-motion, setup and range errors). DR relied on isodose contours generated from the initial dose and associated patient specific weighted objectives to mimic this initial dose in repeated-CTs. These isodose contours, with their corresponding objectives, were used during re-optimization to compensate proton range distortions disregarding re-contouring. Robustness evaluations were performed for the initial, not-adapted and restored (adapted) plans. The resulting DVH-bands showed overall improvement in DVH metrics and robustness levels for restored plans, with respect to not-adapted plans. According to CTV coverage criteria (D95%>95%Dprescription) in not-adapted plans, 35% (5/14) of the cases needed offline adaptation. After DR, Median(D95%) was increased by 1.1 [IQR,0.4] Gy and only one patient out of 14 (7%) still needed offline adaptation because of important anatomical changes. DR has the potential to improve CTV coverage and reduce offline adaptation rate.