Qin S, et al. - Recurrence is a major challenge for cervical cancer treatment, so researchers sought novel molecular markers of cervical cancer to help predict recurrence. Recurrent and nonrecurrent cervical cancer samples, obtained from the Cancer Genome Atlas, were assessed for prognosis-associated long non-coding RNAs (lncRNAs), microRNAs (miRNAs), and mRNAs using expression analysis. A competing endogenous RNA (ceRNA) network was constructed using the predicted regulatory relationships among these prognosis-associated RNAs. In total, they identified 15 lncRNAs, 10 miRNAs, and 348 mRNAs as significant prognosis-associated molecules. Thirteen prognosis-associated lncRNAs, five prognosis-associated miRNAs, and 120 prognosis-associated mRNAs were included in the cervical cancer-related ceRNA network. The key prognostic lncRNAs included H19 and HOTAIR, those for miRNAs included hsa-miR-133b, hsa-miR-138, and hsa-miR-301b, and those for mRNAs included Wnt family member 2, fibroblast growth factor 7, fibronectin 1, synaptic vesicle glycoprotein 2A, and bone morphogenetic protein 7. For recurrence prediction, the key prognostic lncRNAs, miRNAs, and mRNAs seem to have utility as potential molecular markers and may contribute to the treatment of cervical cancer.