MiR-486 as an effective biomarker in cancer diagnosis and prognosis: a systematic review and meta-analysis

Purpose MiR-486 was found to be associated with cancer’s diagnosis and prognosis. This meta-analysis aimed to investigate the potential effect of miR-486 on cancer detection and prognosis. Materials and Methods We searched PubMed, Cochrane library, Embase, Chinese National Knowledge Infrastructure (CNKI) and Wanfang databases to find all correlated articles. The STATA 11.0 was applied to estimate the pooled effects, heterogeneity and publication bias. Results The pooled sensitivity (SEN), specificity (SPE) and Area under the curve (AUC) were 82% (95% CI: 78–85%), 88% (95% CI: 83–92%) and 0.91 (95% CI: 0.88–0.93). Subgroup analysis indicated miR-486 from circulating samples exhibited higher diagnostic accuracy with the AUC was 0.90 (95% CI: 0.87–0.92) than miR-486 from other specimen with the AUC of 0.78 (95% CI: 0.75–0.82) and miR-486 obtained a better diagnostic value in the Asian population with the AUC of 0.94 (95% CI: 0.91–0.95) than the Caucasian and Caucasian/African population with the AUC of 0.80 (95% CI: 0.76–0.83) and 0.89 (95% CI: 0.86–0.91) respectively. MiR-486 obtained high value for the diagnosis of non-small cell lung cancer with SEN, SPE and AUC were 0.82 (95% CI: 0.0.77–0.87), 0.90 (95% CI: 0.84–0.94) as well as 0.92 (95% CI: 0.89–0.94) respectively. For the 7 prognostic tests, the pooled hazard ratio (HR) was 0.48 (95% CI: –0.13–1.08) for low versus high miR-486 expression. Conclusions This meta-analysis indicated that miR-486 can be used as ideal biomarkers in the cancer’s diagnosis. However, Low miR-486 expression did not increase the risk of poor outcome.


INTRODUCTION
MicroRNA is a group of 19-22 nucleotide, small, single-stranded and conserved non-coding RNA that acts as a regulator of gene expression at both the posttranscriptional and the translational levels through acting on the 3'-untranslated region (UTR) of messenger RNA (mRNA) [1]. MicroRNAs involve in various biological processes associated with the tumorigenesis such as the cellular proliferation, differentiation, metabolism as well as apoptosis [2,3]. It is available to isolate the microRNAs from the clinical specimens including the plasma, serum, sputum and tissue. Meanwhile, it has a high stability. Due to these advantages, the microRNAs are increasingly becoming an ideal tool for the detection of human cancer.
Aberrant expression of miR-486 (miR-486-5p) has been reported to be associated with different types of human cancer such as hepatocellular carcinoma (HCC) [4,5], lung cancer [6], breast cancer [7], esophageal squamous cell carcinoma (ESCC) [8] and pancreatic cancer (PC) [9, Meta-Analysis 10]. It can act as both the tumor suppressor and oncogene to participant in the development and progression of tumors. The down-regulation of miR-486 can promote the progression of lung cancer [6], HCC [4,5], breast cancer [7] and osteosarcoma [11], while it is usually up-regulated in PC [9,10], chronic myeloid leukemia [12] and gliomas [13]. Recently, a series of articles have identified that miR-486 might be applied as a biomarker for cancer detection and prognosis. However, as a result of the small sample sizes, different microRNA profiling and the differences of the specimen and ethnicity, many articles showed conflicting conclusions and no meta-analysis was conducted to explore the association between miR-486 and diagnosis as well as the prognosis of human cancer. Therefore, this meta-analysis was performed to assess the performance of miR-486 in the detection and prognosis for human cancer.

Studies' characteristics and quality assessment
A total of 15 articles with 22 studies involving in 1315 cases and 1013 controls were analyzed. The main characteristics of the 22 studies were shown in Table 1. Types of cancer included lung cancer, non-small cell lung cancer (NSCLC), gastric cancer (GC), renal cell cancer (RCC) and pancreatic cancer (PC).The quantitative realtime polymerase chain reaction (qRT-PCR) was used in these studies to test the expression level of miR-486, and the most common reference miRNAs used as the endogenous controls for normalization were RNU6B (U6), miR-39 and miR-16. The quality of the included studies turned out to be generally good and was summarized in Figure 1B.

Publication bias and sensitivity analysis
In this meta-analysis, Deek's funnel plot asymmetry test was applied to test the probability of publication bias. The funnel plot was symmetry ( Figure 3B) and P value equaled 0.39, which demonstrated the publication bias didn't exist in these studies. Sensitivity analysis was also conducted but failed to find the sources of heterogeneity.

Studies' characteristics and quality assessment
A total of 761 participants with cancer from 6 articles on 7 studies were included. The main characteristics of the 7 studies were shown in Table 3 Most studies investigated miR-486 by qRT-PCR. The overall survival (OS), progression-free survival (PFS) and relapse free survival (RFS) were used to evaluate the outcome of the cohorts. Types of the cancer included NSCLC, ESCC, GC and HCC. The results of the studies' quality assessment were also included in Table 3.

Association between miR-486 and outcomes
A random-effects model was used since the heterogeneity among studies existed (I 2 = 89.9, P = 0.000). The pooled HR (hazard ratio) was 0.48 (95% CI: -0.13-1.08) for low versus high miR-486 expression as shown in Figure 7A. Low miR-486 expression did not increase the risk of poor outcome compared with the high miR-486 expression. And the studies were so few that we did not conduct subgroup analysis for the prognostic metaanalysis.

Publication bias and sensitivity analysis
Publication bias was checked by Begg's funnel plot and Egger's test under the random-effects model ( Figure 7B). Although the Begg's funnel plot seemed asymmetric, the P value of the Egger's regression intercept was 0.676, indicating that there was no obvious publication bias among these studies. The sensitivity analysis was also conducted but failed to find the sources of heterogeneity.

DISCUSSION
Cancer biomarkers are critical for cancer detection and predicting the outcome as well as choosing the suitable treatment methods. As involving in various biological processes in cancer, miRNA was considered to play a crucial role in cancer diagnosis and prognosis surveillance.
MiR-486 has been reported to be involved in different types of cancer. Research in the mechanism of miR-486 found that the down-regulation of miR-486 could target genes such as ARHGAP5 to inhibit cell migration and invasion in vitro and metastasis in vivo in lung cancer [6] and it could inhibit cancer cell proliferation, migration and invasive in vitro and suppress HCC growth in vivo by targeting PIK3R1 [4]. Furthermore, miR-486 might inhibit cell growth of papillary thyroid carcinoma by targeting fibrillin-1 [34] and estrogen receptor-mediated miR-486 could regulate expression of OLFM4 in ovarian cancer [35].
This present meta-analysis aimed to estimate the pooled effect of miR-486 expression on diagnosis and To our best knowledge, this meta-analysis was the first one to explore the effect of miR-486 on cancer diagnosis and prognosis. Although we perform this metaanalysis strictly according to the PRISMA guidelines, there were still some limitations that could not be neglected in this meta-analysis. First, the types of cancer included and the studies of each cancer as well as the samples of cases were so few that we could not conduct the subgroup analysis for the prognostic meta-analysis and these limitations might partly contribute to the negative result. Second, the heterogeneity among these studies could not be neglected and some articles might be missed or not be published online that did not be included in this meta-analysis. Third, most studies were from China in the prognostic metaanalysis and the results might just represent the prognostic value of miR-486 on Chinese cancer. Therefore, Studies on the large samples are still demanded to verify our results.

Search strategy
We based our meta-analysis on the Preferred Reporting Items for meta-analyses (PRISMA). We searched PubMed, Cochrane library, Embase, Chinese National Knowledge Infrastructure (CNKI) and Wanfang databases  to find all associated articles in order to investigate the potential utility of miR-486 as a diagnostic and prognostic surveillance tool for human cancer. The combination of the Medical Subject Headings (MeSH) and the keywords: (miR-486 or hsa-miR-486 or microRNA-486 or miR486) and (cancer or tumor or carcinoma or neoplasm) was used (updated to September 13, 2017). We also searched reference lists of the reviews aiming at obtaining other acceptable articles.

Study selection
There was a series of criteria for records inclusion as well as exclusion. For inclusion, records needed to meet the following criteria: 1) Patients of the cases were with cancer; 2) The controls were healthy controls (HC) or with benign diseases (BD); 3) Assess the diagnostic or prognostic value of miR-486 (miR-486-5p); 4) The TP, FP, FN, TN for the diagnosis and HR (hazard ratio) and its  95% CI for the prognosis can be extracted or calculated from the articles. For the exclusion, the criteria as follows: 1) Records that were review, meta-analysis and duplicate publications as well as the records unrelated; 2) Records without complete data or with the same population; 3) Records were about the miR-486-3p.

Data collection and quality assessment
The data was collected independently by two authors as follows: the first author, year of publication, subject's demographic characteristics (ethnicity, mean or median age, sample size, testing method of miR-486 and the types of the controls and cancer); types of the specimen; followup time and the outcomes; miRNA profiling and the data used for this meta-analysis (SEN, SPE, TP, FP, FN, TN, HR and its 95% CI). All HRs were reformatted as low miR-486 expression versus high miR-486 expression. We assessed the quality of these articles with the Quality Assessment of Diagnostic Accuracy Studies 2 (QUADAS-2) guidelines for the diagnostic records and followed the guidelines of the Newcastle-Ottawa Scale (NOS) for the prognostic publications [36,37].

Statistical analysis
We performed the statistical analysis using the and the pooled HR with its 95% Confidence Interval (95% CI) respectively. We also constructed the SROC curve and calculated the AUC value. Simultaneously, we assessed the heterogeneity among the selected studies through the Q test and the I 2 value [38]. The P value for the Q test less than 0.05 or the I 2 ≥ 50% demonstrated that there was heterogeneity among the included studies. For the diagnostic meta-analyses, meta-regression analysis and subgroup analysis (grouped by miRNA profiling: single miR-486 and multiple miRNAs including miR-486 and other miRNAs; specimen: circulating and not circulating; ethnicity: Asian, Caucasian and Caucasian/African; control-type: benign disease and healthy controls; stage: early stage and not only early stage; cancer-type: lung cancer, gastric cancer, pancreatic cancer and renal cell carcinoma and sample size: >= 100 and < -100) were used to identify the potential sources of the heterogeneity and the Deek 's funnel plot asymmetry test was also applied to explore the publication bias, with the P value less than 0.01 considered significant [39]. As for the prognostic meta-analyses, Begg's and Egger's tests were selected to evaluate the included studies for the possibility of publication bias. Finally, the sensitivity analysis was conducted to explore the possible sources of heterogeneity for both the diagnosis and the prognosis meta-analysis.

CONCLUSIONS
This was the first meta-analysis to confirm the potential value of miR-486 on cancer diagnosis and prognosis. The expression of miR-486 might be an effective biomarker for detection of human cancer.