ESTRO 2025 - Abstract Book

S3100

Physics - Inter-fraction motion management and offline adaptive radiotherapy

ESTRO 2025

1245

Digital Poster Accuracy of adaptive protocol in radiotherapy for locally advanced non-small cell lung cancer Liv B. Hysing 1 , Vilde Ragnvaldsen 1 , Tuva B Johannessen 2 , Kristine Fjellanger 1 , John-Vidar Hjørnevik 1 , Inger Marie Sandvik 1 , Turid H Sulen 1 , Tone Nybø 1 , Johanna A Hundvin 1 , Sara Pilskog 1 1 Oncology and medical physics, Haukeland University Hospital, Bergen, Norway. 2 Institute of physics and technology, University of Bergen, Bergen, Norway Purpose/Objective: Radiotherapy for locally advanced non-small cell lung cancer (LA-NSCLC) can be challenging because of breathing motion and anatomical changes occurring during the treatment course. Adapting the treatment plan may improve the outcome for these patients. In 2019, we implemented a traffic-light protocol for adaptive radiotherapy. The aim of this study was to evaluate its accuracy in detecting patients in need for adaptation. Material/Methods: The 38 first LA-NSCLC patients treated with the protocol were prospectively included in this study from 2019-2022. They received IMRT with prescribed dose of 60-70 Gy in 2 Gy fractions. The CBCT at each fraction was matched on bony anatomy to the planning CT (pCT). The RTTs assessed the impact of the anatomical changed by assigning a traffic-light colour: yellow for observed changes that did not require further actions, orange for requirement of a physicist to assess the need for adaption before next fraction, and red for treatments that cannot be given before evaluation by an oncologist. To mitigate bias during training, the traffic-light for the first 19 patients in the cohort were replaced with offline evaluations by an experienced RTT. A cross-sectional study design with two time points, in the first (w1) and third (w3) treatment week, was used. To evaluate delivered target dose at these time points, repeat CTs (rCT) from w1 and w3 were registered to the pCT using bony alignment. Target volumes were re delineated on the 73 rCTs and plans were re-calculated. Poor target coverage was defined by CTV V95% < 98% of the prescribed dose. This indicated that adaptions were necessary and orange or red should have been assigned to the corresponding CBCT. Results: Six out of 73 (8,2%) recalculated plans had CTV V95% < 98, mostly from w1 (Figure). The traffic-light identified these fractions as one yellow, four orange and one red. Three of the plans with poor target coverage were treated with a soft tissue match and two were replanned. In total 13 evaluated fractions were assigned orange or red, of which five had poor target coverage and eight had sufficient coverage. Sixty-two plans were correctly assigned green or yellow. The sensitivity and specificity of the traffic light protocol were 83.3% and 88.6% respectively, giving a balanced accuracy of 85.7%.