ESTRO meets Asia 2024 - Abstract Book

S336

Physics – Motion management and adaptive radiotherapy

ESTRO meets Asia 2024

Keywords: Tumour motion,DIBH,Mediastinal tumour,ITV margin

409

Digital Poster

Evaluation of DIR strategies for OAR dose accumulation in MR-guided pancreas SABR

Mairead Daly 1 , Eliana Vasquez Osorio 1 , Ananya Choudhury 1,2 , Alan McWilliam 1 , Ganesh Radhakrishna 2 , Cynthia L Eccles 3 1 Division of Cancer Sciences, The University of Manchester, Manchester, United Kingdom. 2 Department of Clinical Oncology, The Christie Hospital NHS Foundation Trust, Manchester, United Kingdom. 3 Radiotherapy Department, The Christie Hospital NHS Foundation Trust, Manchester, United Kingdom

Purpose/Objective:

Daily adaptation on the MR Linac offers dosimetric advantages for pancreatic stereotactic ablative body radiotherapy (SABR), such as improved dose to the planning target volume (PTV) whilst maintaining organ at risk (OAR) constraints (1). With daily adaptation, the delivered dose throughout a course of radiotherapy is likely to vary from planned. To assess relationships between delivered dose and patient outcomes, it is necessary to estimate the cumulative delivered dose for a treatment course (2). Mapping daily fraction (#) doses using deformable image registration (DIR) is one strategy to estimate accumulated delivered dose over a course of radiotherapy. DIR accuracy is heavily impacted by the underlying rigid initialisation, also referred to as frame of reference (FOR) registration (3). Furthermore, the accuracy of DIR strategies has not been investigated for estimation of accumulated doses within the abdomen. The aims of this study were to: 1) quantify the difference between accumulated dose and planned dose for OAR in MR-guided pancreatic SABR patients, and 2) evaluate the geometrical performance of three DIR strategies for dose accumulation: global and organ-wise (OW) based on a global FOR registration, and OW based on an OW FOR registration. 10 patients with locally advanced non-metastatic pancreatic cancer treated with 4000 cGy in 5# using online adaptive SABR on the MR Linac were included. All patients were enrolled with informed consent in an ethics approved international registry study (NCT04075305 (4)). Magnetic resonance imaging (MRI) sequences acquired included motion compensated balanced steady-state free procession gradient echo (bFFE) 3-dimensional (3D) VANE and fat-suppressed 3D VANE. Planning MRI (pMRI), reference plan (Ref Plan), daily pre-treatment MRI (FxMRI) as well as structure sets and doses were imported to Raystation v11B. OARs were recontoured on each FxMRI. Global rigid grey level frame of reference (FOR) registrations were created between pMRI and each FxMRI based on a best-fit match of abdominal anatomy. Each FxMRI was deformably registered to pMRI using this global FOR via three strategies: 1) Global DIRs (GLOBAL), where the user changed similarity metric parameters (mutual information, correlation coefficient) to achieve the best registration possible based on visual assessment, 2) OW DIR (GL_OW), where each OAR was used as control structure in one independent registration per organ, per scan using the Raystation scripting interface and 3) OW rigid grey level FOR registrations were then created using the scripting interface and used to initialise OW DIRs similar to step 2 above (OW_OW). Geometrical performance of Material/Methods: