S2475
Physics - Radiomics, functional and biological imaging, and outcome prediction
ESTRO 2026
Conclusion: Repeatable and reproducible T1 and T2
Linac devices at 3 institutions. B1 mapping (DREAM [9] ) was assessed for its utility to correct T1 and T2 measurements. T1 and T2 maps were then registered to each other. The prostate and adjacent obturator internus muscle regions-of-interest (ROI) were contoured and propagated, with a 1-voxel erosion to minimize interface averaging effects. T1 and T2 values from the ROI-enclosed voxels were extracted and used to calculate summary statistics as well as repeatability and reproducibility metrics according to QIBA recommendations [10]. Results: Figures 1 and 2 present mean and standard deviation whole gland prostate and muscle T1 and T2 measurements for each volunteer session. For T1, DESPOT1 presented with promising repeatability (prostate: 61 ms; muscle: 72 ms) and reproducibility (prostate 4%; muscle 6%), and T1 values aligned with the published literature (prostate: 1408+-71 ms; muscle: 1268+-34 ms). For T2, TSE and DESPOT2 both showed promising consistency metrics (repeatability: 3 and 7% for prostate; reproducibility: 0.5 and 2% for muscle), but the actual values were low by published standards (45+-2 ms and 63+-14 ms for prostate; 17+- 2 ms and 23+-5 ms for muscle). MIRACLE was promising, but its performance was skewed by biased values local to the rectum in 2 acquisitions.
measurements are achievable on the Elekta Unity MR- Linac, which supports the continued incorporation of quantitative MRI biomarkers into clinical trials and biologically-guided radiation therapy on this unique platform. References: [1] Bentzen et al. Semin Radiat Oncol 2011 21, 101– 110.[2] Kooreman et al. Radiother Oncol 2020 153, 106–113.[3] Raaymakers et al. Phys Med Biol 2017 62, L41–L50.[4] van Rijssel et al. Radiother Oncol 2025 206, S3000–S3002.[5] Deoni et al. Magn Reson Med 2003 49, 515–526.[6] Nguyen et al. Magn Reson Med 2017 78, 518–526.[7] Carr et al. Phys Rev 1954 94, 630– 638.[8] Meiboom et al. Rev Sci Instrum 1958 29, 688.[9] Nehrke et al. Magn Reson Med 2012 68, 1517– 1526.[10] Shukla-Dave et al. J Magn Reson Imaging 2019 49, e101–e121. Keywords: relaxometry, MR-linac, multi-institution Temporal Stability of MR-Derived Radiomic Features During Radiotherapy for Locally Advanced Cervical Cancer Jasmine Morrison 1 , Alan McWilliam 1 , Michael Dubec 2,1 , Eliana Vasquez-Osorio 1 , Anthea Cree 3 , Peter Hoskin 1,4 , Marcel Van Herk 1 1 Division of Cancer Sciences, University of Manchester, Manchester, United Kingdom. 2 Medical Physics, The Christie NHS Foundation Trust, Manchester, United Kingdom. 3 Clinical Oncology, The Clatterbridge Cancer Centre NHS Foundation Trust, Liverpool, United Kingdom. 4 Clinical Oncology, The Christie NHS Foundation Trust, Manchester, United Kingdom Purpose/Objective Cervical cancer remains one of the most diagnosed cancers in women worldwide. Magnetic resonance imaging (MRI) is used throughout the treatment pathway, providing an opportunity to develop imaging biomarkers for patient stratification [1]. However, imaging biomarkers require crossing translational gaps, such as demonstrating repeatability and Digital Poster 3874 sensitivity to image changes [2]. Radiomics have been widely studied, however feature repeatability during radiotherapy has rarely been evaluated. Here, we evaluate repeatability and minimal detectable change (MDC) of radiomic features in locally advanced cervical cancer (LACC) during radiotherapy. Material/Methods 10 LACC patients undergoing external beam radiotherapy attended MRI scans (1.5T) prior to radiotherapy and at mid and end of treatment. At each session, two T2-weighted images were acquired,
Made with FlippingBook - Share PDF online