Identification of a Novel Prognostic Signature of Genome Instability-Related LncRNAs in Early-Stage Lung Adenocarcinoma

Background: Increasing evidence has demonstrated that long non-coding RNAs (lncRNAs) play a crucial part in maintaining genomic instability. We therefore identified genome instability-related lncRNAs and constructed a prediction signature for early-stage lung adenocarcinoma (LUAD) as well in order for classification of high-risk group of patients and improvement of individualized therapies. Methods: Early-stage LUAD RNA-seq and clinical data from The Cancer Genome Atlas (TCGA) were randomly divided into training set (n =177) and testing set (n =176). 146 genomic instability-associated lncRNAs were identified based on somatic mutation profiles combining lncRNA expression profiles from TCGA by the “limma R” package. We performed Cox regression analysis to develop this predictive indicator. We validated the prognostic signature by external independent LUAD cohort with microarray platform acquired from Gene Expression Omnibus (GEO). Findings: A genome instability-related six lncRNA-based gene signature (GILncSig) was established to divide subjects into high-risk and low-risk group with different outcomes at statistically significant levels. According to the multivariate Cox regression and stratification analysis, the GILncSig was an independent predictive factor. Furthermore, the six-lncRNA signature respectively achieved AUC values of 0.745, 0.659 and 0.708 in training set, testing set and TCGA set. When compared with other prognostic lncRNA signatures, the GILncSig also exhibited better prediction performance. Interpretation: The prognostic lncRNA signature is a potent tool for risk stratification of early-stage LUAD patients. Our study also provided new insights for identifying genome instability-related cancer biomarkers. Funding: None. Declaration of Interest: The authors have no conflicts of interest to declare.