ESTRO 2020 Abstract Book

S218 ESTRO 2020

Purpose or Objective MRI-guided radiotherapy (RT) enables superior soft-tissue outlining of both target and organs at risk. However, geometric distortions in MR images warrant caution in MRI- guided treatment planning. The MRI-system related B 0 non-uniformity and gradient non-linearity (GNL), and patient (pt) induced B 0 disturbance are the main contributors to the total geometric distortion. Further, distortions related to the B 0 non-uniformity (system and pt related) scale linearly with the strength of the main magnetic field, and thus may be important to account for in high field MR-Linac treatments. The aim of this study is to estimate the tumour-site specific geometric distortion in pts treated for cancer in the adrenal gland, liver and prostate, respectively, on a 1.5 T MR-Linac. Material and Methods For this study 25 pts eligible for MR-Linac treatment were included prospectively (target was adrenal gland in 7 pts, liver metastasis in 3 pts and prostate in 15 pts). A 3D GRE dual-echo MRI sequence with echo times (TE) 4.6 and 9.2 ms (in-phase), was used to acquire phase and magnitude images on a 1.5 T MR-Linac (Unity, Elekta AB). B 0 maps were calculated with in-house developed software in MATLAB using the dual-echo method. The B 0 maps were converted to spatial distortion maps using the pixel bandwidth (693 Hz/pixel) and the pixel size (1.2 mm in readout direction) of a clinical 3D T2W SE sequence for the abdomen/pelvic region. Phase unwrapping was performed using a MATLAB function (sunwrap), and the axial through- target slice containing least unresolved phase wraps was used for distortion estimation. GNL related distortion (after 3D system correction) was calculated in 2D (axial plane) using the gradient reversal method and considered constant between pts. Results Representative target specific distortion maps related to B 0 (Figure 1a-c) and systematic GNL (Figure 1d) indicate large inter-patient variation. The mean distortions in the GTV are in the ranges 0.17 -1.92 mm (adrenal gland), 0.69 -1.55 mm (liver), and 0.06 - 1.86 mm (prostate) (Table 1), with no clear target dependent differences. Same levels of B 0 related distortion is seen in a 30 mm region around the GTV, indicating a spatial robustness of the estimates. The distortions related to the systematic GNL depend on the location of the tumour. For target sites within 100 mm from the iso-center, the distortion is below 1.0 mm, and for target sites within 150 mm the distortion can increase up to 1.4 mm. For target sites far off the iso-center (> 150 mm) the GNL may induce distortions up to 3.5 mm (Figure 1d). The largest GNL related distortion in presented dataset is 2.5 mm (pt 2, liver metastasis).

scanners for EBRT, whereas in all other countries <25% reported dedicated scanners for EBRT and were more reliant on collaboration with radiology for MRI access. Anatomical sites receiving MRI for EBRT varies internationally and the most common are shown in table 1. Commissioning and QA of both image registration and MRI scanners varied greatly, as did MRI sequences performed, staffing models and training given to different staff groups.

The most common barrier for increasing MRI scans for radiotherapy across all centres (figure 1) was MRI access and a lack of financial reimbursement, except DK/SE where lack of clinical interest/local knowledge was the main barrier.

It was reported that within 5 years a median of 29% (range 9-50%, absolute number 57) of centres per country are planning for a new MRI scanner dedicated for EBRT A limited number of sites in NL/DK/SE/FN/AU/BE currently employ MRI-only EBRT planning; over the next 5 years MRI- only EBRT planning is expected to be taken up in >50% of centres in NL/SE/DK/FN but < 35% in UK/NZ/BE/FR/AU/IT. MR-linac technology is being clinically employed in DK/UK/NL and within 5 years expected uptake varies between 63% in DK, 59% in NL and <35% in UK/FN/SE/NZ/BE/AU/IT/FR. Conclusion The current international use of MRI for EBRT has been surveyed in 10 countries. Variations in practice/equipment/QA/staffing models have been identified. These are likely due to differences in funding as well as a lack of consensus or guidelines in the literature. For most countries the lack of MRI access and funding is the limiting factor on the number of patients who benefit from MRI as part of their EBRT treatment planning. Despite these challenges, significant interest remains in increasing MRI-assisted EBRT planning over the next 5 years. PH-0407 Tumour-site specific geometric distortions in high field MR-Linac treatments S.W. Hasler 1,2 , E. Van Veldhuizen 1,2 , A. Bertelsen 1,2 , U. Bernchou 1,2,3 , T. Schytte 1,2,3 , V.N. Hansen 1,2 , C. Brink 1,2,3 , F. Mahmood 1,2,3 1 Odense University Hospital, Department of Oncology, Odense, Denmark ; 2 Odense University Hospital, Laboratory of Radiation Physics, Odense, Denmark ; 3 University of Southern Denmark, Department of Clinical Research, Odense, Denmark