Abstract Book

S1266

ESTRO 37

EP-2294 Serum miRNA-based model predicts xerostomia in patients treated for head and neck cancer. B. Tomasik 1,2 , A. Papis-Ubych 3 , R. Bibik 4 , J. Fijuth 2 , W. Fendler 1 1 Medical University of Lodz, Department of Biostatistics

and Translational Medicine, Lodz, Poland 2 Medical University of Lodz, Department of Radiotherapy, Lodz, Poland

3 Regional Center of Oncology and Hematology, Department of Radiotherapy, Lodz, Poland 4 Oncology Center of Radom, Department of Radiation Oncology, Radom, Poland Purpose or Objective One of the most frequently reported side effect of radiotherapy in the head and neck region is xerostomia. Higher than grade 2 acute xerostomia is noted in about 75% patients treated for head and neck cancer. There is an increasing interest in the prevention of radiation- induced xerostomia because reduced salivary output may result in problems with mastication, digestion, swallowing and increase the risk of oral infections. Hence, the aim of this study was to create an efficient serum miRNA-based test predictive for xerostomia in patients treated for oropharyngeal squamous cell carcinoma (OPSCC). Material and Methods Our study is designed as a prospective cohort study which enrolled OPSCC patients treated with IMRT (total dose of 70Gy or equivalent) from June 2016. Until now, we have recruited 44 OPSCC patients and 31 patients (70%) developed acute grade≥3 xerostomia. We randomly selected a group of 10 patients with grade≥3 xerostomia and matched a comparative group of 10 patients without severe xerostomia. We performed microRNA profiling of serum samples collected within 24 hours after the last fraction of RT using qPCR arrays (miRCURY LNA, Human panels I & II, Exiqon, Copenhagen, Denmark). Expression data were normalized toward the average expression of microRNAs detectable in all samples. Acute side effects were prospectively assessed using EORTC QLQ-C30 and EORTC H&N-35 questionnaires. The primary endpoint was moderate-to-severe (grade ≥3) patient-rated xerostomia reported 3 months after the end of treatment. The choice of candidate microRNAs for the classifier was based on differential expression, false discovery rate was controlled using Benjamini-Hochberg method. The classifier for patient-rated xerostomia was created using 5-fold cross-validated logistic regression model built after significance-based feature selection. The study was funded by the National Science Centre within the framework of Preludium 11 project (2016/21/N/NZ5/01938). Results Ten patients (50%) suffered from moderate-to-severe xerostomia 3 months after treatment. Out of 540 miRNAs measured in at least one sample, 49 were detected in all 20 samples. After correction for multiple hypothesis testing, two microRNAs: miR-425-5p and miR-26b-5p, were found to be differentially expressed as depicted on Figure. The logistic regression model based on miR-425- 5p and miR-26b-5p showed nearly perfect separation of the groups with AUC 0.98 (95%CI: 0.93–1.00) and maintained its performance in 5-fold cross validation (AUC=0.96, 95%CI: 0.88–1.00). Hosmer-Lemeshow’s test showed a very good fit of the model (p=1.00).

Conclusion A classification model based on serum expression of miR- 425-5p and miR-26b-5p allowed nearly perfect prediction of xerostomia 3 months after treatment, however further studies with an external validation set are needed to confirm the model’s accuracy in a larger population. EP-2295 HNSCC derived cell lines and subclones differ in genomic copy number changes and radiation response O. Klymenko 1,2 , P. Baumeister 2,3 , H. Zitzelsberger 2,4 , K. Unger 2,4 , J. Heß 2,4 , U. Schötz 5 , C. Belka 1,2 , K. Lauber 1,2 1 Department of Radiation Oncology, University Hospital- LMU Munich, Munich, Germany 2 Clinical Cooperation Group Personalized Radiotherapy in Head and Neck Cancer, Helmholtz Research Center for Environmental Health, Neuherberg, Germany 3 Department of Otorhinolaryngology- Head and Neck Surgery, University Hospital- LMU Munich, Munich, Germany 4 Research Unit of Radiation Cytogenetics, Helmholtz Research Center for Environmental Health, Neuherberg, Germany 5 Department of Radiotherapy and Radiooncology, Philipps-University Marburg, Marburg, Germany Purpose or Objective Head and neck squamous cell carcinomas (HNSCCs) account for approximately 90% of all head and neck neoplasms. Risk factors include tobacco use and alcohol consumption as well as high-risk human papilloma virus infection. Current therapy of HNSCC involves radio(chemo)therapy in definitive or adjuvant settings. However, 5-year survival rates remain limited to 45-50% due to therapeutic failure. In this regard, the emerging issue of intratumoral heterogeneity receives increasing interest as sub-populations of tumor cells with distinct characteristics are considered to drive therapy resistance