ESTRO 2025 - Abstract Book

S3815

Physics - Radiomics, functional and biological imaging and outcome prediction

ESTRO 2025

3315

Poster Discussion Multi-centre longitudinal changes in the apparent diffusion coefficient for short course radiotherapy in rectal cancer. Anne L. H. Bisgaard 1 , Chavelli M. Kensen 2 , Marielle E. P. Philippens 3 , Martijn P. W. Intven 3 , G. J. Meijer 3 , Corrie A. M. Marijnen 2 , Uulke A. van der Heide 2 , Erik van der Bijl 4 , Pètra Braam 4 , Petra J. van Houdt 2 , Faisal Mahmood 1,5 1 Laboratory of Radiation Physics, Odense University Hospital, Odense, Denmark. 2 Department of Radiation Oncology, The Netherlands Cancer Institute, Amsterdam, Netherlands. 3 Department of Radiotherapy, University Medical Hospital Utrecht, Utrecht, Netherlands. 4 Department of Radiation Oncology, Radboud University Medical Center, Nijmegen, Netherlands. 5 Department of Clinical Research, University of Southern Denmark, Odense, Denmark Purpose/Objective: The apparent diffusion coefficient (ADC) derived from diffusion-weighted MRI (DWI) is a potential biomarker for predicting response to neoadjuvant radiotherapy in rectal cancer, enabling personalized treatment [1]. This study aims to determine whether longitudinal changes in DWI parameters can be detected during short-course radiotherapy in a multi-centre cohort of patients with rectal cancer treated on a 1.5T MRI-linac. Material/Methods: This retrospective study included patients with rectal cancer across three centres who received radiotherapy on the MRI-Linac. All patients were enrolled in the ongoing MOMENTUM trial (clinicaltrials.gov, NCT04075305 [2]). In total, 200 patients were evaluated with respect to tumour characteristics and treatments. Patients with primary tumours who received short course radiotherapy (5 fractions of 5Gy) with at least 2 fractions on the MRI-Linac, and no other treatment prior to this, were selected for further analysis (n=141). GTVs were rigidly propagated from pre-treatment T2-weighted MRI to DWI for each treatment fraction. Within a 5 mm margin around the GTV, a region of interest (ROI) was defined on the high b-value DWI image using a threshold based delineation tool [3]. ADC voxel-values were extracted within this ROI using b-values in the range [100, 800] s/mm 2 . For each patient, median ADC values within the ROI were reported and the ADC time-trend across fractions was extracted using linear fitting. Acquisition-related ADC variation was addressed by grouping sequences according to scan parameters and applying median scaling of ADC histograms from all tumour voxel-values between the groups [4]. Results: Seven DWI sequences were identified across centers and grouped into 4 sequence groups (Figure 1). Important differences between the sequence groups were different echo times, repetition times and fat suppression techniques. Median ADC values varied were 1.13, 1.19, 1.07 and 1.16 mm 2 /s for sequence group 1, 2, 3 and 4, respectively. Relatively small scaling factors (range: 0.98-1.09) were needed to account for this variation. The median (range) ADC change during radiotherapy was 19.8% (-23.7%-102.7%) (Figure 2).