Molecular typing and prognostic model of gastric cancer based on immunological signature gene sets

BACKGROUND The high incidence and mortality rates of gastric cancer posed a great challenge in its treatment. The principles of precision medicine called for improved staging diagnosis and prognosis assessment for better patient outcomes. Molecular staging of gastric cancer, based on the TCGA database, had gained significant attention. In this paper, we presented a novel approach to gastric cancer staging, utilizing the analysis of immunological signature gene sets expression across different samples, to construct a reliable prognostic model. Our findings provided a new tool for clinical management and treatment of gastric cancer. METHODS RNA expression data for gastric cancer tissue samples and normal samples were downloaded from the official websites of The Cancer Genome Atlas (TCGA) and the Gene Expression Omnibus (GEO). The immunological signature gene sets were obtained from the Gene Set Enrichment Analysis (GSEA) database. The RNA expression data was analyzed using various computational methods, including unsupervised clustering analysis, principal component analysis (PCA), and survival analysis. The unsupervised clustering analysis was used to divide the samples into three subtypes based on the expression of the immunological signature gene sets. PCA was performed to reduce the dimensionality of the data and identify the most informative genes. Survival analysis was used to identify the genes that were most strongly associated with patient prognosis. Based on the results of the analysis, we identified two gene sets that were strongly associated with patient prognosis. These gene sets were used to construct a prognostic model. The model was evaluated using various statistical methods, including receiver operating characteristic (ROC) curve analysis and Kaplan-Meier survival analysis. Finally, the biologically important functional pathways were analyzed using Gene Ontology (GO), Kyoto Encyclopedia of Genes and Genomes (KEGG), and other databases. These analyses were used to identify the biological processes and pathways that were most strongly associated with the different subtypes of gastric cancer and the genes that were most important for patient prognosis. RESULTS Based on the expression of immunological signature gene sets in different samples, we classified the samples into three subtypes and validated them in the GEO downloaded samples. By analyzing clinical information, we identified 34 prognosis-related gene sets. After screening, we selected two gene sets (GSE21546_UNSTIM_VS_ANTI_CD3_STIM_DP_THYMOCYTES_UP, HALLMARK_ANGIOGENESIS) for model construction. The models were able to accurately predict the prognosis of the samples. Functional analysis revealed that the PI3K-Akt pathway might play a key role. Finally, we performed drug sensitivity analysis on the five core genes (SPP1, COL3A1, FGFR1, APOH and THBD) and identified piperlongumine and cerulenin as highly sensitive drugs.