Abstract
Osteosarcoma accounts for a frequently occurring cancer of the primary skeletal system. In osteosarcoma cells, a hypoxic microenvironment is commonly observed that drives tumor growth, progression, and heterogeneity. Hypoxia and tumor-infiltrating immune cells might be closely related to the prognosis of osteosarcoma. In this study, we aimed to determine the biomarkers and therapeutic targets related to hypoxia and immunity through bioinformatics methods to improve the clinical prognosis of patients. We downloaded the gene expression data of osteosarcoma samples and normal samples in the UCSC Xena database and GTEx database, respectively, and downloaded the validation dataset (GSE21257) in the GEO database. Subsequently, we performed GO enrichment analysis and KEGG pathway enrichment analysis on the data of the extracted osteosarcoma hypoxia-related genes. Through univariate COX regression analysis, lasso regression analysis, multivariate COX regression analysis, etc., we established a predictive model for the prognosis of osteosarcoma. Five genes, including ST3GAL4, TRIM8, STC2, TRPS1, and FAM207A, were found by screening. In particular, we analyzed the immune cell composition of each gene based on the five genes through the CIBERSORT algorithm and verified each gene at the cell and tissue level. Our findings are valuable for the clinical diagnosis and treatment of this disease.