Immunogenomic classification of gastric cancer based on the tumor microenvironments expression of PD-L1 and CD8+ T-cell infiltration.

e16578 Background: Previous data has shown that a positive response to immunotherapy usually relies on active interactions between tumor cells and immunomodulators inside the tumor microenvironment (TME). The aim of this study was to classify gastric cancer (GC) subsets based on the TME immune status according to the expression of PD-L1 and infiltration of CD8+ T cells. Methods: One hundred and eighty-six tumor tissue from gastric cancer patients with a curative D2 gastrectomy were examined for evaluating PD-L1 and CD8+ T cells status using histopathologic analysis. The molecular characteristics of 289 GC samples in TCGA network were further analyzed to distinguish the genetic features of four immune subtypes depending on the presence of PD-L1/CD8+T cell. Results: GC samples were categorized into four types, type I (CD8+/PD-L1+, 60.3%), II (CD8-/PD-L1-, 11.8%), III (CD8-/PD-L1+, 0%), and IV (CD8+/PD-L1-, 27.9%), basing on PD-L1/CD8 expression. The PD-L1 expressing level was geographically associated with the intensity of CD8+ T cell infiltration which was significantly associated with disease-free survival (DFS) and overall survival (OS) (p = 0.003 and p = 0.006). Distinct patterns of genetic profile were described in four types of GC from TCGA database. Type I and III which PD-L1 were overly expressed had comparatively higher MSI and TMB, with EBV mainly enriched in Type I, whereas CIN was more likely to occur in PD-L1 aberrant types II and IV. SNV analysis illustrated higher gene mutations in oncogenes (PIK3CA and ERBB2), and in DNA damage repair related pathway, such as PRKDC, ATM, and SWI/SNF complexes (e.g. ARID1A) in Type I. However, TP53 mutations tend to enrich in Type II and IV. Similar results were obtained by transcriptome analysis. Conclusions: The genetic features of four immune subtypes proof that PD-L1 and CD8+ T cells status are reasonable immunogenomic classification of gastric cancer. SNV analysis prompts a potential mechanism for effectiveness of immunotherapy in Type I patients. Overall, the results may be useful for the development of clinical treatments for the blockade of immune checkpoints.