Prognostic value and risk model construction of hypoxic stress-related features in predicting gastric cancer

Hypoxia promotes tumor progression from multiple aspects, including metabolism, proliferation, migration and angiogenesis. Therefore, a thorough understanding of the impact of hypoxia on gastric adenocarcinoma (STAD) is warranted. The aim of the present study was to find a prognostic model of hypoxia in gastric cancer (GC) and its relationship with the immune microenvironment.

Distinct hypoxia-related patterns were identified with an unsupervised consensus clustering algorithm in STAD patients from the Gene Expression Omnibus (GEO) and the cancer genome atlas (TCGA) cohorts. The different biological processes among different hypoxia-related clusters were then explored with the algorithm of single sample gene set enrichment analysis. Then hypoxia-related Hub genes were selected by weighted gene co-expression network analysis (WGCNA) prior to the construction of a hypoxia-related gene prognostic model. The model was constructed using multivariate Cox regression analysis, least absolute shrinkage and selection operator (LASSO) regression and univariate Cox regression analysis. The relationship between immune infiltration and hypoxia-related features was analyzed.

We identified a hypoxia-related cluster (magenta) by WGCNA and found that different prognosis can be evidently induced by various hypoxia response patterns. LASSO analysis found seven hypoxia-related genes CPZ, LBH, NOX4, NRP1, NOS3, C3orf36 and CDH6, which were then used for the construction of hypoxia-related gene prognostic model. The model was verified by TCGA database and GEO dataset and showing good prognostic value.

A novel hypoxia-related prognostic signature was constructed to predict prognosis and correlate with immune infiltration in STAD. Hypoxia-related prognostic features are expected to be a new prognostic tool for GC.