A long non-coding RNA signature to improve prognosis prediction of gastric cancer

Abstract

Background: Increasing evidence suggests long non-coding RNAs (lncRNAs) are frequently aberrantly expressed in cancers, however, few related lncRNA signatures have been established for prediction of cancer prognosis. We aimed at developing alncRNA signature to improve prognosis prediction of gastric cancer (GC).

Methods: Using a lncRNA-mining approach, we performed lncRNA expression profiling in large GC cohorts from Gene Expression Ominus (GEO), including GSE62254 data set (N = 300) and GSE15459 data set (N = 192). We established a set of 24-lncRNAs that were significantly associated with the disease free survival (DFS) in the test series.

Results: Based on this 24-lncRNA signature, the test series patients could be classified into high-risk or low-risk subgroup with significantly different DFS (HR = 1.19, 95 % CI = 1.13-1.25, P < 0.0001). The prognostic value of this 24-lncRNA signature was confirmed in the internal validation series and another external validation series, respectively. Further analysis revealed that the prognostic value of this signature was independent of lymph node ratio (LNR) and postoperative chemotherapy. Gene set enrichment analysis (GSEA) indicated that high risk score group was associated with several cancer recurrence and metastasis associated pathways.

Conclusions: The identification of the prognostic lncRNAs indicates the potential roles of lncRNAs in GC biogenesis. Our results may provide an efficient classification tool for clinical prognosis evaluation of GC.

Keywords: GSEA; Gastric cancer; LNR; LncRNAs; Survival.

Fig. 1

Kaplan-Meier estimates of the disease free survival (DFS) or overall survival (OS) of GEO patients using the 24-lncRNA signature. The Kaplan-Meier plots were used to visualize the DFS probabilities for the low-risk versus high-risk group of patients based on the median risk score from corresponding GEO datasets patents. a Kaplan-Meier curves for GSE62254 test series patients (N = 180); (b) Kaplan-Meier curves for GSE62254 validation series patients (N = 120); (c) Kaplan-Meier curves for the entire GSE62254 series patients (combined test and validation series patients, N = 300). d Kaplan-Meier curves for GSE15459 patients (N = 192). The tick marks on the Kaplan-Meier curves represent the censored subjects. The differences between the two curves were determined by the two-side log-rank test

Fig. 2

Comparison of the score with prognostic clinical covariates. Multivariable Cox regression proportional hazards regression analyses incorporating the risk score and known prognostic clinical factors, including age at diagnosis, TNM stage (I, II, III, IV) and gender; risk score and age as continuous variables, TNM stage and gender as categorical variables. Solid tetragonums represent the HR of death and open-ended horizontal lines represent the 95 % confidence intervals (CIs). All P values were calculated using Cox proportional hazards analysis. a Multivariable analysis was performed using Cox proportional hazards regression analysis in patients of GSE62254 test series. b Multivariable analysis was performed using Cox proportional hazards regression analysis in patients of GSE62254 validation series. c Multivariable analysis was performed using Cox proportional hazards regression analysis in patients of entireGSE62254 series. d Multivariable analysis was performed using Cox proportional hazards regression analysis in patients of GSE15459 series. All of these were adjusted for the same categorical or continuous variables. Missing: HR (95 % CI) could not be calculated out

Fig. 3

LncRNA risk score analysis of GSE62254 test series. The distribution of 24-lncRNA risk score, patients’ survival status and lncRNA expression signature were analyzed in the GSE62254 test series patients (N = 180). a lncRNA signature risk score distribution; (b) patients’ survival status and time; (c) heatmap of the lncRNA expression profiles. Rows represent lncRNAs, and columns represent patients. The black dotted line represents the median lncRNA risk score cutoff dividing patients into low-risk and high-risk groups

Fig. 4

Kaplan-Meier estimates of the disease free survival (DFS) of GEO patients using the 24-lncRNA signature, stratified by lymph node ratio (LNR). Entire GSE62254 set (N = 300) were first stratified by LNR (LNR ≥ 16.7 % or LNR < 16.7 %). Kaplan-Meier plots were then used to visualize the survival probabilities for the high-risk versus low-risk group of patients determined on the basis of the median risk score from the entire GSE62254 set patients within each LNR stratum. a Kaplan-Meier curves for the entire GSE62254 set patients (N = 300); (b) Kaplan-Meier curves for patients LNR ≥16.7 % (N = 139); (c) Kaplan-Meier curves for patients LNR < 16.7 % (N = 161). The tick marks on the Kaplan-Meier curves represent the censored subjects. The differences between the two curves were determined by the two-sided log-rank test

Fig. 5

Kaplan-Meier estimates of the disease free survival (DFS) of GEO patients using the 24-lncRNA signature, stratified by postoperative chemotherapy. Entire GSE62254 set (N = 299) were first stratified by postoperative chemotherapy (with or without postoperative chemotherapy). Kaplan-Meier plots were then used to visualize the survival probabilities for the high-risk versus low-risk group of patients determined on the basis of the median risk score from the GSE62254 set patients within each postoperative chemotherapy stratum. a Kaplan-Meier curves for the entire GSE62254 set patients (N = 299); (b) Kaplan-Meier curves for patients with postoperative chemotherapy (N = 80); (c) Kaplan-Meier curves for patients without postoperative chemotherapy (N = 219). The tick marks on the Kaplan-Meier curves represent the censored subjects. The differences between the two curves were determined by the two-sided log-rank test

Fig. 6

Kaplan-Meier estimates of the disease free survival (DFS) of GEO patients using the 24-lncRNA signature, stratified by TNM stage (I, II, III & IV). Kaplan-Meier plots were then used to visualize the survival probabilities for the high-risk versus low-risk group of patients determined on the basis of the median risk score from the entire GSE62254 set patients within each TNM stage. a Kaplan-Meier curves for patients with TNM stage I (N = 30); (b) Kaplan-Meier curves for patients with TNM stage II (N = 97); (c) Kaplan-Meier curves for patients with TNM stage III (N = 96); (d) Kaplan-Meier curves for patients with TNM stage IV (N = 77). The tick marks on the Kaplan-Meier curves represent the censored subjects. The differences between the two curves were determined by the two-sided log-rank test

Fig. 7

Receiver operating characteristic(ROC) analysis of the sensitivity and specificity of the disease free survival (DFS) prediction by the 24-lncRNA risk score, AJCC stage, lymph node ratio (LNR) and postoperative chemotherapy in GSE62254 set patients with known chemotherapy information (N = 202). P values were from the comparisons of the area under the ROC (AUROC) of 24-lncRNA risk score versus those of AJCC stage, 24-lncRNA risk score combined with AJCC stage, LNR and postoperative chemotherapy, respectively. As can be seen, the 24-lncRNA risk score combined with AJCC stage showed a better prediction of DFS than AJCC stage (P = 0.0002). The predictive ability of risk score was equivalent to AJCC stage alone (P = 0.1861), but better than both LNR (P = 0.0297) and postoperative chemotherapy (P < 0.0001)

Fig. 8

a Gene set enrichment analysis delineates biological pathways and processes correlated with risk score. Cytoscape was used for visualization of the GESA results. Nodes represent enriched gene sets that are grouped and annotated by their similarity according to related gene sets. Enrichment results were mapped as a network of gene sets (nodes). Node size is proportional to the total number of genes within each gene set. Proportion of shared genes between gene sets is represented as the thickness of the green line between nodes. b Box plot of risk score of patients with or without recurrence in entire GSE62254 series excluding patients without available information (N = 283, P < 0.0001).T-test was used to determine the significance of the comparisons