Fig 1.
The overall analysis design of our study.
Table 1.
PCa patients from TCGA database clinicopathological features.
Table 2.
PCa patients from ICGC database clinicopathological features.
Fig 2.
DElncRNAs between prostate cancers and non-carcinoma tissue samples.
Red dots represent differentially upregulated lncRNAs, green dots represent differentially downregulated lncRNAs and then dark dots represent no differentially expressed genes.
Fig 3.
Four lncRNAs Forest map formed by multivariate Cox regression analysis.
Table 3.
Overview for the four prognostic lncRNAs related to PCa patient OS.
Fig 4.
Ability of the 4 lncRNAs-based signature in predicting the prognosis for PCa cases from TCGA database.
(A) Risk score distribution for patients. (B) PCa patient survival time. (C) Expression heat map for those four lncRNAs incorporated into the prognosis model. The vertical black dotted line stands for the optimum threshold to divide cases to high- or low-risk group.
Fig 5.
Predictive power of the model for prognosis in PCa patients.
(A) Distribution of patient risk score. (B) Survival time of PCa patients.(C) Heat maps of the four lncRNAs included in the prognostic model. The vertical dotted black line represents the optimal threshold for classifying cases into high—and low-risk groups.
Fig 6.
Prediction ability of the four-lncRNA nomogram evaluated by Kaplan-Meier analysis, log rank P and C-index based on TCGA data.
OS time between low- and high-risk groups were visualized and compared by plotting the Kaplan-Meier curves.
Fig 7.
Kaplan-Meier analysis, Log rank P, and C-index were used to evaluate the predictive power of four lncRNA nomogram based on ICGC data.
Fig 8.
Prognosis prediction ability of the four-lncRNA nomogram evaluated by ROC curves and dynamic AUC lines using TCGA data.
The time-dependent ROC curves at 1 (A), 3 (B) and 5 (C) years during the follow-up period, together with dynamic AUC lines were drawn for the cases.
Fig 9.
Using ICGC data, the ROC curve and dynamic AUC line were used to evaluate the prognostic prediction ability of four lncRNA histograms.
The ROC curve and the area under the dynamic curve (AUC) during 1 (A), 3 (B) and 5 (C) years of follow-up were plotted.
Fig 10.
TCGA multivariate analysis results demonstrated the possibility to use the 4-lncRNA signature as the potent predicting factor for PCa OS rate from other clinicopathological factors.
Red represents high risk indicators and green represents low risk ones.
Fig 11.
The results of ICGC multivariate analysis showed that 4-lncRNA characteristics could be used as an effective predictor of PCa OS rate for other clinicopathological factors.
Table 4.
TCGA OS-related univariate as well as multivariate Cox regression analysis.
Table 5.
ICGC OS-related univariate as well as multivariate Cox regression analysis.
Fig 12.
TCGA Kaplan-Meier survival analysis for PCa cases having the Gleason score > 7.
All cases were classified as high- or low-risk group using the four-lncRNAs model.
Fig 13.
ICGC Kaplan-Meier survival analysis of PCa patients with Gleason score > 7.
Fig 14.
TCGA ROC curves for the prediction of the 3 years overall survival among the 4-lncRNA signature model, Gleason score and our lncRNA nomogram combined with Gleason score.
Fig 15.
ICGC ROC curves for the prediction of the 3 years Overall survival among the 4-lncRNA signature model, Gleason score and our lncRNA nomogram combined with Gleason score.
Fig 16.
Functional enrichment analysis for the related biological processes and pathways related to the 4 lncRNAs used in that model.
GO biological process enrichment results (A). KEGG signaling pathways analysis (B).