Open AccessA fourteen-lncRNA risk score system for prognostic prediction of patients with non-small cell lung cancer
Author(s) -
Jiayi Song,
Xiaoping Li,
Xiujiao Qin,
Jingdong Zhang,
Jianyu Zhao,
Rui Wang
Publication year - 2020
Publication title -
disease markers. section a, cancer biomarkers
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.959
H-Index - 41
eISSN - 1875-8592
pISSN - 1574-0153
DOI - 10.3233/cbm-190505
Subject(s) - oncology , univariate , proportional hazards model , receiver operating characteristic , medicine , lung cancer , multivariate statistics , framingham risk score , multivariate analysis , long non coding rna , survival analysis , gene , disease , computer science , biology , machine learning , rna , biochemistry
Growing evidence has underscored long non-coding RNAs (lncRNAs) serving as potential biomarkers for cancer prognosis. However, systematic tracking of a lncRNA signature for prognosis prediction in non-small cell lung cancer (NSCLC) has not been accomplished yet. Here, comprehensive analysis with differential gene expression analysis, univariate and multivariate Cox regression analysis based on The Cancer Genome Atlas (TCGA) database was performed to identify the lncRNA signature for prediction of the overall survival of NSCLC patients. A risk-score model based on a 14-lncRNA signature was identified, which could classify patients into high-risk and low-risk groups and show poor and improved outcomes, respectively. The receiver operating characteristic (ROC) curve revealed that the risk-score model has good performance with high AUC value. Multivariate Cox’s regression model and stratified analysis indicated that the risk-score was independent of other clinicopathological prognostic factors. Furthermore, the risk-score model was competent for the prediction of metastasis-free survival in NSCLC patients. Moreover, the risk-score model was applicable for prediction of the overall survival in the other 30 caner types of TCGA. Our study highlighted the significant implications of lncRNAs as prognostic predictors in NSCLC. We hope the lncRNA signature could contribute to personalized therapy decisions in the future.