P-148 Molecular characterisation of gastric tumours in a South Indian cohort and their clinical correlation

Abstract

Gastric cancer is amongst the leading causes of cancer and cancer-related mortality worldwide. The prevalence of gastric cancer is particularly high in Eastern Asia with nearly 70 % incidence in developing countries. In this population, there remains a wide heterogenicity observed in the cancer aggressiveness and treatment outcomes, the plausible explanation being the molecular heterogenicity in gene expression and oncogenic pathways. The developments in microarray and next-generation sequencing platforms enabled the genome-scale identification of molecular dysregulations at the level of mRNA and the possible genomics-guided stratification of tumors for eventual targeted therapeutics. Numerous whole-genome profiles decoding the landscape of molecular determinants of gastric tumors have been established from Japan, South Korea, and China. However, such larger profiles of gastric tumors are largely lacking from India. This is the first comprehensive genome-wide expression landscape covering mRNA, miRNA, lncRNA, and splice variants from India. Gastric tumor samples were collected from our hospital in both liquid nitrogen and RNA later. Quality analysis was performed in Agilent Bioanalyzer 2100 using RNA 6000 Nano chips. Genome-wide expression profiling covering mRNA, miRNA, alternate splicing and lncRNA was performed using Affymetrix HTA 2.0 arrays. The dysregulated genes were compared with the profiles established from other countries. Integrative genomic analysis revealed the transcription factors OCT, NFkB, NFAT, STAT, LEF, AP, PAX, SP1, ELK1, NFYA to be up-regulated and HNF, GATA to be down-regulated in tumors. The tumour samples were then subtyped into eight different groups based on the differential activation of 21 oncogenic signaling pathways. The clinical features, intraoperative details and histopathological characteristics of the samples of the patients were collected. Patients were then followed up for a period of 2 years for recurrence and DFS was calculated. Correlation between genomic subtypes of the tumour samples and clinical outcomes/recurrence patterns were analysed. 108 patients who underwent distal gastrectomy were included in the study and genome-wide expression profiling were performed with their tumour samples. Based on the differential activation of 21 oncogenic signaling pathways, gastric cancer in this South Indian cohort could be subdivided into eight different subtypes. Among the eight subtypes, the subtype GC-1 that showed activation of differentiation-related pathways (Notch, NFAT, SP1-D, RXRA, ECM, HIF1A) was the most prevalent (n=36). The subgroup GC-7 that demonstrated no significant activation of any pathway had propensity for extranodal extension from lymph nodes when compared to the other types (p-value 0.17). Subgroup 5 and subgroup 7 had greater propensity for spread to regional lymph nodes (p-value 0.45 and 0.42, respectively). The patients belonging to subgroup 7 developed metastasis at an earlier date (p-value 0.08). Among the eight different subtypes of gastric cancer differentiated on the basis of oncogenic signaling pathway activation, the subtype GC-1 was the most prevalent, GC-7 had propensity for extranodal extension in lymph nodes and developed earlier metastasis, and subgroups GC-5 and GC-7 had a greater chance of involving regional nodes.