ESTRO 2025 - Abstract Book

S3717

Physics - Radiomics, functional and biological imaging and outcome prediction

ESTRO 2025

respectively. All studies used single-shot echo-planar-imaging (SS-EPI) sequence to acquire DWI and additional non EPI DWI sequences were used in 2 studies, while with considerable variabilities in b-value setting. Apparent diffusion-coefficient (ADC) was used in all studies. The majority of technical studies focused on feasibility, acquisition optimization, distortion assessment/correction, image quality assessment, and ADC accuracy/repeatability/reproducibility. Five clinical studies, none conducted within clinical trials, reported DWI utilization and preliminary findings in clinical use, focusing on clinical feasibility/applicability and treatment response assessment/prediction, involving brain tumors (patient n=34), prostate cancer (n=43), sarcoma (n=30) and HNSCC (n=30). Data heterogeneity prevented further meta-analysis. Conclusion: The integration of DWI into MRgRT on MR-LINACs presents a promising avenue for improving treatment planning, adaptation, response moitoring/assessment, and prediction. Further acquisition standardization/optimization and image quality improvement are necessary. ADC repeatability/reproducibility also needs to be improved. Despite its potential benefits, current clinical findings are preliminary. Clinical evidences are still limited. Further validation and trials are warranted. This systematic review underscores the need for continued exploration of DWI as a biomarker in MRgRT, with the aim of optimizing patient outcomes and personalizing treatment approaches. Poster Discussion Prediction of radiation-induced esophagitis during chemoradiation therapy for NSCLC: Independent external validation using the RTOG 0617 trial data Laura Cella 1 , Serena Monti 1 , Giuseppe Palma 2 1 Institute of Biostructures and Bioimaging, National Research Council, Naples, Italy. 2 Institute of Nanotechnology, National Research Council, Lecce, Italy Purpose/Objective: Radiation-induced esophagitis (RE) is a dose-limiting complication associated with concurrent chemoradiation therapy for Non-Small-Cell Lung Cancer (NSCLC). Our aim was to externally validate previously published results on dosimetric predictors as well as dose patterns associated with RE of grade ≥ 2 (RE2+) after concurrent chemo radiotherapy for NSCLC [1]. Material/Methods: We retrospectively analyzed RTOG 0617 trial patient data retrieved from The Cancer Imaging Archive. Patients received concurrent chemotherapy with or without cetuximab, and 60- versus 74-Gy radiation doses [2]. The average esophageal dose-volume histograms (DVHs) for patients with and without RE2+ were compared. Voxel based analysis (VBA) was performed to assess the spatial patterns of the dose differences between patients with and without RE2+ following the pipeline described in [3]. Dose were converted into Biologically Effective Dose (BED) using an alpha-beta ratio of 10 Gy for acute reacting tissue. The generalized linear model (GLM) was designed to include BED maps and each non-dosimetric variables selected by a Least Absolute Shrinkage and Selection Operator (LASSO) regularized GLM of RE2+. A non-parametric permutation test of the maximum threshold-free cluster-enhanced statistic accounting for multiple comparisons was performed, the significance p - maps generated and compared using DICE over volume metric [4] with previously published VBA results on NSCLC patients from the NCT00915005 trial. The predictive performance of esophageal dosimetric factors were also assessed in terms of discrimination (ROC AUC). Keywords: Diffusion weighted imaging, MRgRT 475

Results: Among 495 patients in the trial, 326 patients had complete RT treatment and passed VBA image quality check for