Open AccessA four‐gene signature associated with clinical features can better predict prognosis in prostate cancer
Author(s) -
Yuan Penghui,
Ling Le,
Fan Qing,
Gao Xintao,
Sun Taotao,
Miao Jianping,
Yuan Xianglin,
Liu Jihong,
Liu Bo
Publication year - 2020
Publication title -
cancer medicine
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.403
H-Index - 53
ISSN - 2045-7634
DOI - 10.1002/cam4.3453
Subject(s) - prostate cancer , univariate , proportional hazards model , gene signature , oncology , gene , medicine , multivariate statistics , survival analysis , multivariate analysis , cancer , computational biology , bioinformatics , gene expression , biology , computer science , genetics , machine learning
Prostate cancer (PCa) is one of the most deadly urinary tumors in men globally, and the 5‐year over survival is poor due to metastasis of tumor. It is significant to explore potential biomarkers for early diagnosis and personalized therapy of PCa. In the present study, we performed an integrated analysis based on multiple microarrays in the Gene Expression Omnibus (GEO) dataset and obtained differentially expressed genes (DEGs) between 510 PCa and 259 benign issues. The weighted correlation network analysis indicated that prognostic profile was the most relevant to DEGs. Then, univariate and multivariate COX regression analyses were conducted and four prognostic genes were obtained to establish a four‐gene prognostic model. And the predictive effect and expression profiles of the four genes were well validated in another GEO dataset, The Cancer Genome Atlas and the Human Protein Atlas datasets. Furthermore, combination of four‐gene model and clinical features was analyzed systematically to guide the prognosis of patients with PCa to a largest extent. In summary, our findings indicate that four genes had important prognostic significance in PCa and combination of four‐gene model and clinical features could achieve a better prediction to guide the prognosis of patients with PCa.