ESTRO 2025 - Abstract Book

S3762

Physics - Radiomics, functional and biological imaging and outcome prediction

ESTRO 2025

1945

Digital Poster Repeatability and reproducibility of diffusion-weighted MRI of rectal cancer on a MR-Linac Jonas Habrich 1 , Cihan Gani 2,3 , Johanna A. Hundvin 4,5 , Marcel Nachbar 1 , Monica Lo Russo 2 , Simon Böke 2,3 , Maximilian Niyazi 2,3 , Daniela Thorwarth 1,3 1 Section for Biomedical Physics, Department of Radiation Oncology, University Hospital Tübingen, Tübingen, Germany. 2 Department of Radiation Oncology, University Hospital Tübingen, Tübingen, Germany. 3 German Cancer Consortium (DKTK), partner site Tübingen; and German Cancer Research Center (DKFZ), Heidelberg, Germany. 4 Department of Oncology and Medical Physics, Haukeland University Hospital, Bergen, Norway. 5 Institute of Physics and Technology, University of Bergen, Bergen, Norway Purpose/Objective: Mean apparent diffusion coefficient (ADC) value as a quantitative imaging biomarker (QIB) derived from diffusion weighted (DW)-MRI has shown to be predictive of response to radiotherapy in patients with rectal cancer [1]. Additionally, interobserver variability in the delineation of gross tumor volumes (GTVs) on hybrid MR-Linacs can be significantly improved through a transanally applied rectal ultrasound gel filling [2]. To further evaluate the use of ADC values on MR-Linacs as a QIB in rectal cancer, the aim of this study was to evaluate repeatability and reproducibility of ADC values derived from a 1.5 T MR-Linac with and without rectal filling. Material/Methods: A total of 17 patients with rectal cancer were treated and imaged on a 1.5 T MR-Linac between 04/2021 and 05/2024. Pretreatment DW-MRI using a single-shot spin-echo echo-planar-imaging sequence (b-values 0, 150 and 500 s/mm 2 ) was acquired twice (Test/Retest) without rectal filling and one additional time with filling within the same imaging session. ADC maps were calculated with an in-house developed python script using b150 and b500 images. Using 3DSlicer, GTVs were delineated on every b500 image by an experienced radiation oncologist and mean ADC values were extracted (Figure 1). Absolute and relative repeatability (RC, relRC) and reproducibility coefficients (RPC, relRPC) [3] including Bland-Altman analysis were calculated in Matlab2023a comparing both images without filling (RC, relRC) and with/without filling (RPC, relRPC).

Results: Mean ± standard deviation of GTV Test /GTV Retest volumes without filling resulted in 18±11/ 16±9 cm filling showed 19±17 cm 3 . Mean ADC without filling was 1284±143/ 1313±185×10 -6 mm 2 /s for GTV Test /GTV Retest and 1514±238×10 -6 mm 2 /s for GTVs with filling. Repeatability without filling resulted in RC and relRC of 278×10 -6 mm 2 /s and 22% while comparison of ADC with and without filling yielded RPC of 679×10 -6 mm 2 /s and relRPC of 47%, respectively. Bland-Altman analysis showed mean ADC bias without filling of -24.5±144×10 -6 mm 2 /s and - 254±245×10 -6 mm 2 /s with filling (Figure 2). 3 and GTVs with