Identification of microRNA-92a and the related combination biomarkers as promising substrates in predicting risk, recurrence and poor survival of colorectal cancer

Abstract

Background: Previous studies demonstrated that microRNA-92a (miR-92a) may serve as a novel promising biomarker in colorectal cancer (CRC) patients. However, a comprehensive analysis of the contribution of miR-92a in CRC is lacking. We aimed to systematically summarize the diagnostic and prognostic values of miR-92a in CRC. Methods: The diagnostic and prognostic roles of individual miR-92a and the combination biomarkers based on miR-92a were evaluated through comprehensive meta-analyses. Meanwhile, the function and potential mechanisms of miR-92a were assessed by an integrative bioinformatics analysis. Results: According to the results, we found that miR-92a yielded a pooled area under ROC curve (AUC) of 0.82 (sensitivity: 76%, specificity: 75%) in discriminating CRC from controls. Notably, the combination biomarkers based on miR-92a increased the diagnostic performance, yielding an AUC of 0.91, with a sensitivity of 83% and a specificity of 87%. For the prognostic meta-analysis, patients with higher expression of miR-92a had significant shorter overall survival (pooled HR: 2.30; 95% CI: 1.03-5.12). In addition, the regulated genes of miR-92a were retrieved and enriched through gene ontology and pathway analysis, indicating their correlations with the initiation and progression of CRC. Furthermore, protein-protein interaction network was set up with miR-92a targets and screened for hub nodes and significant modules, which were confirmed strongly involved in the occurrence and development of CRC again. Conclusions: Current evidences suggest miR-92a is a promising biomarker for early detection and prognosis of CRC while miRNA combination biomarkers may be considered as the right way for clinical practice. However, more prospective studies are required to highlight the theoretical strengths.

Keywords: Biomarker; Colorectal cancer; Meta-analysis; System biological analysis.

Figure 1

Flow chart of study selection process

Figure 2

Forest plots of sensitivities and specificities from test accuracy studies in the diagnosis of CRC. A Forest plots of sensitivities and specificities for miR-92a alone; B forest plots of sensitivities and specificities for miR-92a-related combination markers

Figure 3

The SROC curves of miR-92a in the diagnosis of CRC. A SROC curve overall including the outliers for miR-92a; B SROC curve for miR-92a in plasma samples; C SROC curve for miR-92a in serum samples; D SROC curve of outliers excluded for miR-92a. SROC summary receiver operator characteristic, CRC colorectal cancer

Figure 4

Sensitivity analysis results for miR-92a alone. a Goodness of fit; b bivariate normality; c influence analysis; d outlier detection

Figure 5

funnel plots for the assessment of potential bias in the meta-analysis for diagnosis. A Funnel plot of the studies on miR-92a alone; B funnel plot of the studies on miR-92a-related combination markers

Figure 6

The SROC curves of miR-92a-related combination biomarkers in the diagnosis of CRC. A SROC curve overall including the outliers for miRNA combination biomarkers; B SROC curve of outliers excluded for miRNA combination biomarkers; C SROC curve for miRNA combination biomarkers in plasma samples; D SROC curve for miRNA combination biomarkers in serum samples; E SROC curve for miRNA combination biomarkers (combinations=2); F SROC curve for miRNA combination biomarkers (combinations>2). SROC summary receiver operator characteristic, CRC colorectal cancer

Figure 7

Sensitivity analysis results for miR-92a-related combination markers. a Goodness of fit; b bivariate normality; c influence analysis; d outlier detection

Figure 8

Functional enrichment results of miR-92a target genes. A Top 10 GO items for target genes of miR-92a target genes; B Significantly enriched pathways for target genes of miR-92a. GO gene ontology, BP biological processes, CC cell component, MF molecular function

Figure 9

PPI network construction results. A Degree distributions of nodes for network set up with miR-92a targets; B-D top 10, 20 and 30 hub genes of network for miR-92a targets, respectively; E-G pathway enrichment results for the top 10, 20 and 30 hub genes of miR-92a targets network, respectively; PPI protein-protein interaction

Figure 10

The top three significant modules from the PPI network. A-C The top three significant modules in the PPI network for miR-92a targets; D-F pathways enriched by all the nodes involved in the identified three modules, respectively. PPI protein-protein interaction.