ESTRO 2025 - Abstract Book

S3057

Physics - Image acquisition and processing

ESTRO 2025

Conclusion: The use of GBCA is essential for adaptive MRI-guided radiotherapy of glioblastoma. Half-dose contrast-enhanced MRI may be sufficient for monitoring during radiotherapy, but full-dose is preferable for adaptation. A feasible low dose workflow may include full-dose MRI at the first fraction and midway through treatment, with optional half dose MRI at the end. For small lesions, only full-dose images are recommended and should be required more frequently.

Keywords: MRI contrast-dose, glioblastoma, adaptive RT

References: [1] Bernchou U et al. Evolution of the gross tumour volume extent during radiotherapy for glioblastomas. Radiother Oncol. 2021, PMID: 33848564 [2] Mahmood F et al. Safety of gadolinium based contrast agents in magnetic resonance imaging-guided radiotherapy - An investigation of chelate stability using relaxometry. Phys Imaging Radiat Oncol. 2022, PMID: 35243039

[3] Tweedle MF. Gadolinium Retention in Human Brain, Bone, and Skin. Radiology. 2021, PMID: 34128728

[4] de Mol van Otterloo SR et al. The MOMENTUM Study: An International Registry for the Evidence-Based Introduction of MR-Guided Adaptive Therapy. 2020, PMID: 33014774

[5] Otsu N. A Threshold Selection Method from Gray-Level Histograms. IEEE Trans. Syst. Man Cybern. 1979, 9, 62-66.

3560

Digital Poster Image-based assessment of breast density as a risk factor for radiation-induced breast cancer in young women treated with mediastinal radiotherapy Rory A D Bell 1 , Tanwiwat Jaikunaa 1,2 , Edward G A Henderson 1 , Hannah Chamberlin 1 , Sacha Howell 1 , Eliana M Vasquez Osorio 1 , Marianne Aznar 1 1 Division of Cancer Sciences, University of Manchester, Manchester, United Kingdom. 2 Division of Radiation Oncology, Department of Radiology, Faculty of Medicine Siriraj Hospital, Mahidol University, Bangkok, Thailand Purpose/Objective: High breast tissue density (BD), as estimated on mammograms (MG), is associated both with a higher risk and a lower detectability of breast cancer (BC)[1]. It is hypothesized it could also lead to a higher risk of radiation-induced BC, e.g., in young female lymphoma patients treated with mediastinal radiotherapy (RT). However, mammograms are unavailable at the time of RT planning for these patients, limiting the application of risk models developed for mammographic BD. In this feasibility study, we propose a novel approach using Gaussian Mixture Models (GMMs) to calculate BD on CT scans (BD CT ) and compare results to an established MG-based tool (BD MG ). Material/Methods: 20 young (<40 years old) female unilateral BC patients with both CT and MGs acquired at the time of diagnosis were included in this study. We used the LIBRA tool [2, 3] to calculate BD MG from cranio-caudal mammograms and selected the breast with lowest density as the contralateral (i.e. tumour-free) breast. On CT scans, breast tissue was contoured using RayStation’s Breast Deep Learning model to produce masks. Fat and fibroglandular tissues were segmented using a two-peaked individualised GMM, selected for its efficacy in distinguishing the two tissues based on voxel intensity. BD was then calculated as the ratio of fibroglandular tissue to total breast tissue. BD MG and BD CT were then compared and their correlation was assessed using Spearman’s rank coefficient. To mimic the evaluation