Abstract
Given the high morbidity and the trend of younger individuals being affected observed in cervical cancer, it is important to identify sensitive and effective biomarkers for predicting the survival outcome of patients. Based on data from 307 cervical cancer cases acquired from The Cancer Genome Atlas portal, 1920 differentially expressed mRNAs, 70 microRNAs(miRNAs), and 493 long non-coding(lncRNAs) were screened by comparing cervical cancer tissues with paracancerous tissues. A competing endogenous (ceRNA) network containing 50 lncRNAs, 16 miRNAs, and 81 mRNAs was constructed. Eighteen RNAs, comprising 13 mRNAs, 2 miRNAs, and 3 lncRNAs, were identified as significant prognostic factors by univariate Cox proportional hazards regression. ETS-related gene and fatty acid synthase signatures were discovered using a multivariate Cox regression model built to identify independent prognostic factors in cervical cancer patients. Receiver operating characteristic (ROC) analysis was used to determine the optimal cut-off value for distinguishing the risk level of cervical cancer patients. High-risk patients exhibited a poorer prognosis than low-risk patients did. This study focused on ceRNA networks to provide a novel perspective and insight into cervical cancer and suggested that the identified signature can serve as an independent prognostic biomarker in cervical cancer.