ESTRO 2020 Abstract book

S956 ESTRO 2020

Conclusion This method can be used in a prospective way to estimate the accuracy of DIR based dose reconstruction and provides a method to objectively assess the adequacy of treatment in the context of anatomical changes or variations in patient positioning. The DAU based on DIR can be accurately quantified and incorporated in DVHs. PO-1648 Impact of image deformation workflow in high-field MR-Linac treatment of head and neck cancer R.L. Christiansen 1,2 , J. Johansen 3 , R. Zukauskaite 3 , C.R. Hansen 1,2 , F. Mahmood 1,2 , A.S. Bertelsen 1 , T. Schytte 2,3 , O. Hansen 2,3 , C. Brink 1,2 , U. Bernchou 1,2 1 Odense University Hospital, Laboratory of Radiation Physics, Odense, Denmark ; 2 University of Southern Denmark, Department of Clinical Research, Odense, Denmark ; 3 Odense University Hospital, Department of Oncology, Odense, Denmark Purpose or Objective Deformable image registration (DIR) and contour propagation is used in MR-Linac (MRL) treatment to minimize time spent on online treatment adaptation. The precision of propagated contours may vary, depending on the chosen workflow (WF) that can affect the quantity of required manual corrections. This study investigated the effect of three different WFs of DIR and contour propagations produced by the clinical treatment planning system for high-field MRL in head and neck cancer patients. Material and Methods Seventeen patients referred for curative RT for oropharyngeal cancer were MR scanned in the treatment position for planning (pMR) and at the 10 th (MR 10 ), 20 th (MR 20 ) and 30 th (MR 30 ) fraction (± 2) on a diagnostic 1.5 T scanner. An oncologist delineated the primary tumour (GTV-T), the largest lymph node (GTV-N) and larynx on the planning CT (pCT) and on each set of T2 weighted images. An RTT delineated the following organs at risk (OAR): brain stem, spinal cord, bilateral parotids, submandibular glands and thyroid. All delineations were repeated on the pMR by the same person at least one month later to determine the intra-observer variation (IOV). The images were deformed using Monaco 5.40 (Elekta AB, Stockholm, Sweden) following the three WF strategies illustrated in Figure 1. In WF1, only the planning images and contours were used as reference for DIR and propagation to MR 10,20,30 . In WF2, the image set acquired at the previous fraction with its propagated, uncorrected contours were mapped to the current MR. WF3 was identical to WF2, except that here the manually delineated contours from the previous imaging session were mapped to the current session using DIR. Dice similarity coefficient (DSC), mean surface distance (MSD) and Hausdorff distance (HD) were calculated for each structure in each model and tested for statistical significant differences by the Wilcoxon signed rank test (α = 0.05).

Results WF1, WF2 and WF3 yielded similar DSC for GTVs and OARs, generally (Table 1), but inferior to the IOV for all investigated structures. WF3 provided better MSD and DSC overall than WF1 and WF2. MR-MR propagation showed better results than MR-CT for OAR evaluated by DSC and MSD regardless of WF with p < 0.05 for all structures except for thyroid gland propagations of WF2. For mean of median HDs the DIR results was better than the IOV, which is in contrast to the DSC and MSD measures. Compared to WF1, WF2 introduced more uncertainty due to multiple, successive DIR. WF3 reduced the uncertainty by using corrected structures, compared to WF1 and 2.

Conclusion The MR-MR DIR produced the most precise structure propagation. Revision of contours is necessary, but WF3 provides the best starting point for fast online dose planning. The IOV is still a factor which should be considered when ensuring target coverage. PO-1649 Dose comparison between hybrid plan and deformable images' plan in nasopharyngeal cancer patients W. Iabsakul 1 , J. Petsuksiri 1 , T. Jaikuna 1 , C. Kakanaporn 1 , W. Phongprapun 1 , K. Thephamongkhol 1 , J. Setakoranukul 1 , J. Kittikornchaichan 1 , P. Dankulchai 1 1 Employee, Radiation Oncology, Bangkok, Thailand Purpose or Objective Common problems that we encounter during the radiation treatment process are changing body contours, targets,