Abstract
BACKGROUND: Gastric cancer (GC), a major global health challenge, originates from the gastric mucosal epithelium and is characterized by high morbidity and mortality rates. Sphingosine-1-phosphate (S1P), a bioactive lipid mediator, has been implicated in tumor progression, metastasis, and immune regulation across multiple cancers; however, its specific role and associated molecular mechanisms in GC remain unclear. This study aims to explore the prognostic value of S1P-related genes (SRGs) and to construct a reliable risk model for GC. METHODS: The Cancer Genome Atlas (TCGA)-GC and GSE62254 datasets and the SRGs were derived from public databases. Differentially expressed genes (DEGs) were identified between the normal and GC groups, using the TCGA-GC dataset. Key module genes that are associated with SRGs were detected by a weighted gene co-expression network analysis (WGCNA), and candidate genes were derived from the intersection between DEGs and key module genes. Prognostic genes were determined via a univariate Cox regression analysis and the least absolute shrinkage and selection operator (LASSO) algorithm. A risk model was then established and assessed via a survival curve and receiver operating characteristic (ROC) curve. An independent prognostic analysis was performed and a nomogram was constructed using data from the TCGA-GC dataset. We also performed an enrichment analysis, an immune microenvironment analysis and a drug sensitivity analysis. RESULTS: A total of 260 candidate genes were detected using the intersection between 4,484 DEGs and 1,398 key module genes. Seven prognostic genes were screened. A risk model was established and verified, and ROC curves showed good efficiency. The risk score, age and cancer stage were selected as independent prognostic factors. A predictive nomogram, which contained independent prognostic factors, was established. The estimate score and stromal score were significantly higher in the high-risk group than the low-risk group (P<0.05). Moreover, 'DNA replication' and 'ribosome' were involved in the low-risk group, whilst 'neuroactive ligand receptor interaction', 'extracellular matrix (ECM) receptor interaction' and 'focal adhesion' were involved in the high-risk group. Finally, the half-maximal inhibitory concentrations (IC50) values of AZD8055 and docetaxel drugs were found to be lower in the high-risk group, when compared with the low-risk group. CONCLUSIONS: Seven prognostic genes that were associated with S1P, in GC, were investigated by constructing a risk model, which may provide clinical significance for the treatment of GC.