PS02.210: EVALUATION OF PREDICTION SYSTEM IN TREATMENT EFFECT AND PROGNOSIS OF ESOPHAGEAL CANCER BASED ON RADIOEPIGENOMICS THEORY

Background We tried integrated analysis of epigenomic information (Epigenomics) and image information (Radiomics) as Ragioepigenomics for esophageal cancer. Methods 1, Expression analysis of miR, evaluation as prognostic predictor (Epigenomics), 2, Image information analysis (Radiomics): a, Normalization of contrast CT images, feature quantities by 3D gray level co-occurrence matrix extraction. b, Normalization of SUV of FDG-PET image. Extraction of feature quantity by 3D gray level co-occurrence matrix. c, Image analysis of features. 3, Integrated analysis of Epigenomics and Radiomics information: We analyzed the relationship between miR expression level and variables obtained from image analysis. Results As a result of Cox regression by the selection method using elastic net, three variables were selected from contrast CT, and the results of both were consistent. Furthermore, when the survival curve and the log rank test were performed by the Kaplan-Meier method for the cluster shade high value group and the low value group, a significant difference was observed in the 2 groups. Conclusion A multimodal approach so called Radioepigenomics, which is a system for analyzing molecular information from image analysis will be applied to future Precision Medicine. Disclosure All authors have declared no conflicts of interest.