The pathophysiology of ovarian cancer (OV) is complex and depends on multiple biological processes and pathways. Therefore, there is an urgent need to identify reliable prognostic biomarkers for predicting clinical outcomes and helping personalize treatment of OV. A long non-coding RNA (lncRNA)-based risk score model was constructed to infer the prognostic efficacy of transcription factors (TFs) based on the OV dataset from The Cancer Genome Atlas. The risk score model was further validated in other independent cohorts from Gene Expression Omnibus. Time-dependent receiver operating characteristic curves were used to evaluate the survival prediction performance in comparison with other clinical and molecular variables. Our results revealed that the top-ranked TF-associating lncRNAs were significantly associated with overall survival, progression-free survival and disease-free survival. Stratification analysis according to clinical variables indicated the prognostic independence of POLR2A-associating lncRNAs. In comparison, the signature of POLR2A-associating lncRNAs was more sensitive and specific than existing clinical and molecular signatures. Functional and experimental analysis suggested that POLR2A-associating lncRNAs may be involved in known biological processes and pathways of OV. Our findings revealed that the lncRNA-based risk score model can provide helpful information on OV prognosis stratification and discovery of therapeutic biomarkers.
Keywords: lncRNAs; ovarian cancer; prognostic signature; survival analysis; transcription factors.
© 2018 The Authors. Journal of Cellular and Molecular Medicine published by John Wiley & Sons Ltd and Foundation for Cellular and Molecular Medicine.