ESTRO 38 Abstract book

S253 ESTRO 38

Purpose or Objective Circulating tumor cells (CTCs) are detectable in many cancers, including breast and lung cancer, where they can have prognostic significance. However, because of the lack of a suitable detection method, there are no useful data on CTCs in patients undergoing radiotherapy. The only US Food and Drug Administration-approved methodology, the CellSearch platform, uses epithelial cellular adhesion molecule EpCAM, exploiting the positivity of carcinomas to common epithelial markers. The monitoring of CTCs under chemotherapy showed a correlation of persistent CTCs and shorter survival in breast cancer. This indicates stronger defense mechanisms in the remaining CTCs. More efficient DNA repair capacity could contribute to such stronger resistance. The aim of this study is to identify relevant DNA damage response pathways in CTCs and peripheral blood lymphocytes under radiotherapy and their possible implications for the adjustment of future therapies. Material and Methods Up to now 47 patients with brain metastases of breast and lung cancer (n=21/26) receiving radiotherapy were included in the study. Blood samples were collected before, at the end of radiotherapy and at the first follow- up (87 blood samples so far). The number of CTCs at the first follow-up was compared with clinical treatment response (e.g. MRI/CT and performance status). Enumeration and characterization of CTCs were done using the CellSearch®system. Apoptosis was measured in CTCs ( in vivo irradiation) with the help of the M30 antibody in the CellSearch®system and DNA damage repair analyzed by yH2AX and 53BP1 foci detection was analyzed in primary lymphocytes ( ex vivo irradiation). Results CTCs were detectable in 19% of lung cancer and 38% of breast cancer patients with brain metastases before start of radiotherapy. Quantitative changes in the number of CTCs under local radiotherapy were measurable in all patients. In the lung cancer group, 40% of patients showed an increase of CTCs after irradiation. This percentage was much higher in the breast cancer group with 75%. To specify whether the observed increase in number was due to vital or lethal CTCs an apoptotic marker (M30) was stained in addition in the same samples. In 82% the increase in CTC number was accompanied by apoptosis. After correction for this the number of vital CTCs decreased, indicating treatment response. This is mirrored by the clinical follow up via MRI/CT (evaluation under way) after different radiation schedules (whole brain vs. stereotactic treatment). The same treatment response was detectable in the DNA-damage response assessment parameters. Conclusion The results indicate that monitoring DNA repair in CTCs and primary lymphocytes is already showing promising potential for judging treatment response after radiotherapy in the metastatic state of disease. The increase in apoptotic cells under radiotherapy suggests ineffective DNA repair and thus a local response to therapy. On the other hand, if persistent CTCs are present, this indicates efficient DNA repair and a poor prognosis. OC-0494 Genetic variants associated with radiation- induced xerostomia in head and neck cancer: a GWA study E. Naderi 1,2 , A.P. Crijns 1 , R.J. Steenbakkers 1 , H.P. Bijl 1 , J.F. Van den Hoek 1 , M. Dieters 1 , M.H. Boezen 2 , B.Z. Alizadeh* 2 , J.A. Langendijk* 1 1 The University of Medical Center Groningen, Radiation Oncology, Groningen, The Netherlands ; 2 The University of Medical Center Groningen, Epidemiology, Groningen, The Netherlands

Purpose or Objective We aimed to find genetic variants associated with radiation-induced late Xerostomia in head and neck cancer (HNC) patients population treated with definitive or postoperative radiotherapy (RT). Material and Methods We included 1,061 HNC patients treated with definitive or postoperative RT with or without chemotherapy from a prospective cohort study. All baseline patient- tumor- and treatment characteristics and acute and late toxicity was prospectively scored. Patients were followed up to 5 years after treatment. Patient phenotype data were imputed to correct for missingness of samples. Patients were genotyped on Illumina human hap 550k v.3.0 (n=607) or Illumina global screening array (n=464). Genotypes were imputed on Haplotype Reference Consortium reference panel version R1.1. Principal component analysis (PCA) was performed to identify population substructures among samples. Eventually, 957 patients and 6,334,277 SNPs passed quality controls which were included in the final analysis. Xerostomia was defined as moderate to severe xerostomia at 6 months (XER6M). We fitted first a phenotype model (model I) to identify the significant non- genetic clinical and bassline factors associated with XER6M by using imputed samples. Next logistic regression was used to estimate the association of the additive effect of genetic variants with XER6M while adjusting for other predictors from the model I and the top four eigenvectors obtained from PCA analysis. A p-value<0.05 for clinical co- variable, and a genome-wide p-value<5.0x10 -8 for genetic variants was considered statistically significant. Results Full data were available in 763 patients, of whom 280 (36.7%) had XERM6. These cases were compared to 483 patients without XERM6. Clinical factors including N stage (OR=0.29; 95%CI 0.12 to 0.68; p=0.004), volume surrogate (5.97; 3.11 to 11.46; p=7.16x10 -8 ), definitive radiotherapy (0.22; 0.15 to 0.32; p=1.61x10 -15 ) and baseline xerostomia (2.60; 2.00 to 33.36; p=5.48x10 -13 ) were significantly associated with XER6M. In total, 26 variants across eight genomic regions showed a suggestive association at (5.0x10 -8