Abstract
The omentum is a common site of peritoneal metastasis in various cancers, including gastric cancer. It contains immune cell aggregates known as milky spots, which provide a microenvironment for peritoneal immunity by regulating innate and adaptive immune responses. In this study, we investigated gene expression profiles in cells from omental milky spots of patients with gastric cancer (n = 37) by RNA sequencing analysis and classified the patients into four groups (G1-4). Notably, significant differences were observed between the groups in terms of macroscopic type, lymphatic invasion, venous invasion, and pathological stage (pStage). G3, which was enriched in genes related to acquired immunity, showed earlier tumor stages (macroscopic type 0, Ly0, V0, and pStage I) and a better prognosis. In contrast, G4 showed enrichment of genes related to neutrophils and innate immunity; G1 and G2 showed no enrichment of innate or adaptive immune-related genes, suggesting an immune desert microenvironment. Cytometric analysis revealed significantly more T and B cells and fewer neutrophils in G3. Accordingly, the immune microenvironment in omental milky spots may vary depending on the stage of gastric cancer progression. When univariate Cox proportional hazards regression models were used to search for prognostically relevant genes specific to G3, 23 potential prognostic genes were identified as common genes associated with relapse-free survival and overall survival. In addition, the multivariate Cox proportional hazards model using these prognostic genes and clinicopathological information showed that combining the B cell marker CD19 and Ly had a high predictive accuracy for prognosis. Based on this study's results, it is possible that tumor progression, such as lymphatic and/or venous infiltration of tumor cells, may affect the immune cell composition and proportions in omental milky spots of patients with gastric cancer and analysis of gene expression in omental milky spots may help to predict gastric cancer prognosis.