Gene expression profiling of Japanese gastric cancer patinets using Asian Cancer Research Group classification.

Abstract

72 Background: Gastric cancer is well known as having heterogeneous features. Recently, the Asian Cancer Research Group (AGRG) had proposed a new classification scheme based on the gene expression profile of the tumor. However, the genomi/expression profiling of gastric cancer in Japanese patients is still unknown. We started a comprehensive molecular profiling study that analyzes genome and transcriptome of tumor obtained from cancer patients admitted to Shizuoka Cancer Center from 2014. We already had evaluated more than 1,500 samples from various type of cancer. Among them, 104 gastric cancer patients were analyzed. Methods: Fresh surgically resected tumor/normal samples and peripheral blood were obtained and whole-exome sequencing (Ion Proton, Life Technologies) and gene expression profiling (DNA microarray, Agilent Technologies) were performed. Patients were grouped based on the gene expression profile according to AGRG classification, and clinicopathological features were compared among the group. Results: Patients were classified into MSI in 14, MSS/EMT in 15, MSS/TP53+ in 38 and MSS/TP53- in 37, respectively. There was no significant difference of sex among the group. Age was significantly younger in MSS/EMT and MSS/TP53-. In MSI, tumor tended to be located at antrum, and differentiated type tumor was predominant. In MSS/EMT, advanced T stage (T4) and undifferentiated type of tumor was predominant. In MSS/TP53+, relatively less advanced stage and localized macroscopic type tumor was predominant. In MSS/TP53-, relatively advanced stage and invasive macroscopic type tumor was predominant. Although the follow-up period is insufficient, relapse-free survival was the worst in MSS/EMT and no patient recurred in MSI. Conclusions: Classification of gene expression profiling based on ACRG was possible in Japanese gastric cancer. Distribution and tumor characteristics were almost identical to ACRG cohort. Gene -expression profiling may be comprehensively used for tumor classification and further clinical trials of molecular targeting agents.