An immune-related gene-based signature as prognostic tool in ovarian serous cystadenocarcinoma

Abstract

Background: Ovarian serous cystadenocarcinoma (OSCC) is a life-threatening malignancy with poor prognosis. Therefore, the identification of immune-related genes associated with OSCC prognosis may reveal new targets of immunotherapy for OSCC. Patients and Methods: The gene expression profiles of overlapped genes were extracted by weighted gene co-expression network analysis (WGCNA) to identify immune-related modules. Significant genes were identified by univariate Cox regression analysis of model genes. Model characteristic genes were obtained by least absolute shrinkage and selection operator (LASSO) analysis and used to calculate a “signature index”. The model’s ability to predict prognosis in OSCC patients was assessed using time-dependent receiver operator characteristic curves. Differences in the biological processes and Kyoto Encyclopedia of Genes and Genomes pathways between groups with high or low signature index were assessed using gene set enrichment analysis (GSEA). The types of immune cells and their abundance in the two index groups were explored by single-sample GSEA. Results: The expression profiles of 3517 overlapped genes were extracted by WGCNA, and nine modules related to the immune system of OSCC were obtained. The expression profiles of 114 hub genes were then subjected to LASSO analysis. Among them, 10 immune-related genes were significant, of which six were identified as model characteristic genes and were used to calculate the signature index. Moreover, 24 types of immune cells were identified in the tumor microenvironment, and their abundance was explored in high-and low-signature index groups of two datasets. Conclusion: ARHGEF18, PLEKHA7, MTOR, VPS45, BRCA1, and HINT2 were identified as characteristic genes and used to develop a new immune-related gene-based signature as a promising prognostic biomarker for OSCC.