ESTRO 37 Abstract book

ESTRO 37

S584

PO-1039 Meta-analysis of toxicity and small-bowel radiotherapy dose-volume: “omnibus consequentia” D. Holyoake 1 , M. Partridge 1 , M. Hawkins 1 1 CRUK MRC Oxford Institute for Radiation Oncology, Department of Oncology- University of Oxford, Oxford, United Kingdom Purpose or Objective The limited radiotherapy tolerance of the small-bowel causes toxicity for many patients receiving abdominal or pelvic radiotherapy. We conducted a systematic review and meta-analysis using published datasets of dose- volume parameters with toxicity outcomes for the small bowel to analyse and model the relationship with toxicity. Material and Methods SCOPUS, EMBASE & MEDLINE were searched using keywords and subject headings. 952 unique results were screened to identify 16 publications reporting small- bowel dose-volumes and toxicity. Where suitable data were available (mean values with parametric measures of error) fixed-effects inverse-variance meta-analysis was used to compare irradiated small-bowel dose-volumes for cohorts of patients with and without grade ≥3 toxicity. For all other data, non-parametric correlation between irradiated small-bowel dose-volume and incidence of grade ≥3 toxicity was examined. Results On fixed-effects meta-analysis of three studies (total 203 patients), each of the dose-volume measures V 5Gy –V 40Gy were significantly greater (p < 0. 00001) for patients who suffered Grade ≥3 toxicity compared to those who did not. The absolute difference is largest for the lowest dose-volume parameter and decreases with increasing dose, however the relative difference in volume actually increases with increasing dose, (table). When data from all studies were examined together, no significant correlations were found. However, when data from studies of rectal cancer radiotherapy were examined separately, the V 10Gy , V 30Gy , V 35Gy , and V 40Gy all correlated with incidence of toxicity, most strongly for the V 10Gy (Spearmans correlation coefficient 0.888, CI 0.489-0.980, p = 0.003), for which the figure shows the fit of a univariate logistic regression model. Table. Results of meta-analysis including absolute and relative differences in volume of small bowel irradiated to threshold dose VxGy, between patients with Grade 0-2 and those with Grade ≥3 Toxicity.

Conclusion Analysis of published clinical cohort dose-volume data with toxicity outcomes provides evidence for a significant dose-volume-toxicity response effect for a range of clinically-relevant doses in the treatment of rectal cancer. The largest absolute differences in irradiated volume are seen at low doses, but the largest relative differences at high doses. Significant correlations of volume with toxicity risk are seen at both low dose (V 10Gy ) and high dose (V 30Gy , V 35Gy , and V 40Gy ), which has important implications for minimising risk of toxicity in these patients. PO-1040 Comprehensive bioinformatics analysis about the radiosensitivity in esophageal carcinoma M. Li 1 , B. Li 1 1 Shandong Cancer Hospital, radiation department, Jinan, China Purpose or Objective Esophageal carcinoma (EC) is a common malignancy and radiotherapy plays an important role in its treatment, while few biomarker could predict the radiotherapy response in current clinical practice. To explore the radiation response associated genomic characterizations, we performed a comprehensive bioinformatics analysis using the TCGA EC data with radiation records. Material and Methods All the clinical and genomic data of TCGA esophageal carcinoma were obtained from GDC Data Portal. To avoid confounding factor, we firstly used correlation test or χ 2 test to assess the interdependency of clinical information. The survival time related to radiation therapy were defined from the start of radiation therapy to the last follow up or death, and log rank test and Cox model were adopted for survival analysis. The level 3 MAF files and genome wide segmented SNP chips were used for demonstrate mutation and copy number variations profiles. For mRNA-seq counts data, DESeq2 were used for differential expression analysis, while we adopted CIBERSORT tool to estimate the infiltrating immune cells. We used minfi to handle the methylation data and limma was employed for normalization and comparation for other expression profiles. PCA and MDS were used for exploration of high dimension genomic profiles, while co- expression analysis and hierarchical cluster were used to demonstrate the genomic internal connectivity. We further performed gene set enrichment analysis and biological network analysis for mechanism exploration. Results Fifty TCGA EC patients received radiotherapy, among whom 35 categorized to radio-sensitive and 14 radio- resistant group according to their radiation response. For clinical variables, the radiation response strongly associated with chemo-resistant status. In survival analysis adjusted by clinical parameters, despite the relative low statistic power due to small proportion

Figure. Plot of toxicity rate and small-bowel V 10Gy [cm 3 ] for rectal cancer radiotherapy, with logistic regression model (dashed line) and 69% confidence interval (dotted lines). Studies are labelled with the number experiencing toxicity and the total number in the cohort.