ESTRO 2021 Abstract Book

S963

ESTRO 2021

manual DSCs for P-CT and CBCT were 0.83 and 0.80, respectively. The time consumed in manual delineation was significantly higher than the time consumed by the algorithm- operator.

Conclusion The automated segmentation algorithm provides a clinical reliable, user friendly and time-saving tool for delineation of pulmonary lesions in SBRT.

PO-1161 Feasibility of stereotactic body radiotherapy of (ultra)central lung tumors using an 1.5 T MR- linac L. Merckel 1 , S. Hackett 1 , A. van Lier 1 , M. van den Dobbelsteen 1 , M. Rasing 1 , L. Snoeren 1 , C. van Es 1 , M. Fast 1 , P. van Rossum 1 , J. Verhoeff 1 1 UMC Utrecht, Radiation Oncology, Utrecht, The Netherlands Purpose or Objective SBRT is an important modality for the radical treatment of malignant tumors in the lungs. Treatment with highly ablative doses in central or ultracentral tumors has been demonstrated to result in high rates of toxicity. MRI-guidance during treatment with daily plan adaptation may aid accurate delineation of OARs and target volumes in proximity of the mediastinum with daily plan adaptation allowing for safer treatment. In this study, we report the first clinical experiences on the safety and feasibility of SBRT of (ultra)central lung tumors on an 1.5 T MR-linac. Materials and Methods Both patients with primary NSCLC and patients with an oligoprogressive metastatic lung nodule or mediastinal lymph node were eligible for treatment on the Unity MR-linac (Elekta AB, Stockholm, SE) if their PTV was within 2 cm of the mediastinum. Pre-treatment imaging with 4D-CT and MRI was performed in treatment position and a pre-treatment offline IMRT plan was created in Monaco v5.40.01. Patients were treated to a stereotactic dose of 60 Gy in 8 or 12 fractions. An in-house developed T2-weighted 3D sequence acquired during free breathing was used for online delineation and treatment planning. For each fraction, contours of ITV and OARs were propagated using deformable image registration. A radiation oncologist modified the