S2592
Radiobiology - Preclinical biomarkers
ESTRO 2026
[95% CI 0.74 - 0.87] on the development cohort, and 0.80 [95% CI 0.67 - 0.90] on the validation cohort. Application of the final combined signature obtained a significant risk stratification on local tumor control in both cohorts (p < 0.001) (Figure 1).
Digital Poster 1499 MRI-based radiomics enables prediction of tumor control after radiotherapy in non-small cell lung cancer (NSCLC) subcutaneous xenograft models Cylia Ouadah 1,2 , Claire Gordziel 1,3 , Michael Ramirez Parra 1,3 , Nathalie Borgeaud 1,3 , Lydia Koi 2,4 , Michael Baumann 5,6 , Henning Willers 7 , Rebecca Bütof 1,4 , Antje Dietrich 1,3 , Mechthild Krause 3,4 , So ň a Michlíková 1,2 , Alex Zwanenburg 1,8 , Steffen Löck 1,2 1 OncoRay - National Center for Radiation Research in Oncology, Faculty of Medicine and University Hospital Carl Gustav Carus, TUD Dresden University of Technology, Helmholtz-Zentrum Dresden- Rossendorf, Dresden, Germany. 2 Institute of Radiooncology - OncoRay, Helmholtz-Zentrum Dresden-Rossendorf, Dresden, Germany. 3 German Cancer Consortium (DKTK), Partner Site Dresden, and German Cancer Research Center (DKFZ), Heidelberg, Germany. 4 Department of Radiotherapy and Radiation Oncology, Faculty of Medicine and University Hospital Carl Gustav Carus, TUD Dresden University of Technology, Dresden, Germany. 5 German Cancer Research Center (DKFZ), and German Cancer Consortium (DKTK), Core Center Heidelberg, Heidelberg, Germany. 6 Division of Radiooncology/Radiobiology, German Cancer Research Center (DKFZ), Heidelberg, Germany. 7 Department of Radiation Oncology, Massachusetts General Hospital, Harvard Medical School, Boston, Massachusetts, USA. 8 National Center for Tumor Diseases (NCT), NCT/UCC Dresden, a partnership between DKFZ, Faculty of Medicine and University Hospital Carl Gustav Carus, TUD Dresden University of Technology, and Helmholtz-Zentrum Dresden-Rossendorf (HZDR), Dresden, Germany Purpose/Objective: Radiomics is an innovative approach to tumor characterization via quantitative analysis of clinical imaging data. Preclinical subcutaneous mouse xenograft models enable the comparison of such imaging features with the biological properties of tumors. In this proof-of-principle study we developed and validated a magnetic resonance imaging (MRI)-based radiomics signature to predict tumor control after radiotherapy in three non-small cell lung cancer (NSCLC) xenograft models with varying radiosensitivity. Material/Methods: A heterogeneous xenograft tumor cohort was established in NMRI (nu/nu) mice by subcutaneous injection of cell suspensions derived from three NSCLC cell lines. (NCI-H1703, NCI-H441, and NCI- H23). The tumor-bearing mice received fractionated radiation combined with targeted therapy or conventional chemotherapy over 6 weeks. Tumor control was determined based on measurements of tumor volumes over a period of 180 days. Tumors at a diameter of 5-6 mm in 111 mice were examined pre-therapeutically using MRI (Bruker BioSpec PET/MR 3T) with a T1-weighted 3D sequence with contrast agent. The tumors were contoured in µ-RayStation (Raysearch). MIRP software [1] was used to extract 161 statistical, morphological, and texture-based features. We trained a Cox model, within a framework of 30 repetitions of a 3-fold cross-validation, on the development cohort (n = 87). A radiomic signature was defined by considering the median signature size across all repetitions and a stability rank aggregation method. Finally, a combination of the radiomic signature, radiation dose and treatment information was used to train a Cox model, which was validated on previously unseen data (n = 24). The performance was evaluated using the concordance index (C-index) and the log-rank test. Results: The identified radiomic signature included 2 intensity-based features: ‘stat_min’ and ‘ih_skew_fbn_32’, which were found to be differentially associated with the three tumor models. The final Cox model trained on the combined signature achieved a C-index of 0.81
Conclusion: This study demonstrated that radiomics analyses based on pre- therapeutic MRI can be used for the prediction of tumor control after radiotherapy in preclinical NSCLC models. In further analyses the biological interpretability of the developed signature will be investigated by correlating the selected radiomics features with histological stainings.* Cylia Ouadah and Claire Gordziel contributed equally to this work References: References:[1] Zwanenburg et al., JOSS (2024). https://doi.org/10.21105/joss.06413 Keywords: radiomics, xenograft, tumor control probability Digital Poster 1778 TGF-β as a biomarker of radiotherapy response in ductal carcinoma in situ Tanja Marinko 1,2 , Manca Pirc 2 , Š pela Plestenjak 2 , Maja Marin č i č 3 , Vita Dol ž an 3 , Katja Gori č ar 3 1 Department of Radiotherapy, Institute of Oncology Ljubljana, Ljubljana, Slovenia. 2 Faculty of Medicine, University of Ljubljana, Ljubljana, Slovenia. 3 Pharmacogenetics Laboratory, Institute of Biochemistry and Molecular Genetics, Faculty of Medicine, University of Ljubljana, Ljubljana, Slovenia Purpose/Objective: Adjuvant radiotherapy (RT) after breast-conserving surgery is a standard treatment for patients with ductal carcinoma in situ (DCIS), but there are interindividual differences in treatment response and occurrence of adverse events. Important biological pathways that could contribute to these differences include the inflammatory response, as RT alters the expression of various anti-inflammatory and pro-inflammatory factors. One of the key cytokines is transforming growth factor beta (TGF- β ), which helps to maintain tissue homeostasis. The aim of our study was to determine if plasma TGF- β concentration is altered after RT and if genetic variability contributes to differences in TGF- β concentration. We also evaluated the association of TGF- β concentration with acute and late adverse events of RT in DCIS patients. Material/Methods: Our prospective longitudinal study (NCT06648148) included 175 DCIS patients treated with adjuvant RT. Adverse events were evaluated immediately after RT and at one-year follow-up. Skin adverse events were assessed using Common Terminology Criteria for Adverse Events (CTCAE) v.5 and The Late Effects on Normal Tissue -Subjective, Objective Management and Analytic (LENT-SOMA) scale. Cardiac adverse events were assessed using New York Heart
Made with FlippingBook - Share PDF online