The Construction and Analysis of a Novel Four-Long Non-Coding RNA Signature Predicting Survival of Breast Cancer Patients

Background: As the critical regulators for tumorigenesis and progression, long-noncoding RNAs (lncRNAs) are becoming novel prognostic biomarkers for tumor patients. By the levels of lncRNAs expression, the patients with breast carcinoma may be divided into subgroups with different risk scores. Nevertheless, there is limited evidence to evaluate the role of lncRNAs in the prognosis of breast carcinoma. The present study aimed to construct lncRNA signatures for prognostic analysis and assist clinicians in choosing optimal therapies. Methods Abnormal expression profiles of breast cancer-associated lncRNAs were analyzed based on the TCGA datasets. Univariate and multivariate Cox regression analysis was used to build a prognostic risk signature according to the lncRNAs expression. The prognostic ability of this signature was verified in various subgroups. Functional enrichment analysis was employed to reveal the potential roles of these predictive lncRNAs in cancer-related biological processes and pathways. Results Compared with normal breast tissues, the differential analysis demonstrated that 286 lncRNAs were abnormally expressed in breast carcinoma. A four-lncRNA signature (RP1-193H18.2, AL022341.3, WDR86-AS1, LINC00511) was found to be closely related to the prognosis of breast carcinoma. The four-lncRNA signature could also qualify the magnitude of treatment benefits for different breast cancer subtypes. Additionally, it was an independent risk factor out of other clinicopathological parameters based on the multivariate Cox analysis. We also uncovered that the four predictive lncRNAs are involved in multiple cellular progression and pathways of breast cancer. Conclusions The four-lncRNA signature could be an essential reference for prognostic prediction and making therapeutic strategies.