Background
Bladder cancer (BC) as the second common urinary system malignancy is a leading cause of global cancer-related deaths. The incidences of BC were 37.5 per 100,000 in men and 9.3 per 100,000 in women [
1]. Approximated 72,570 new cases of BC were diagnosed and resulted in 15,210 deaths in the USA in 2013 [
1]. The 5-year survival rate after endoscopic resection is high as 80% because BC in majority of the cases is low grade and non-muscle-invasive [
1‐
3]. However, 10–20% of non-muscle-invasive BC (NMIBC) will progress to muscle-invasive BC (MIBC), and in these cases, the survival rate dramatically drops to 56% [
4,
5]. In addition, approximately 50–70% of NMIBC cases show cancer recurrence within the first 2 years of treatment, with up to 90% of MIBC cases showing recurrence overall [
3,
6,
7]. Due to the high mortality (44%) related to MIBC and the frequent recurrence (70%) of cancer, early diagnosis and monitoring of disease progression is particularly important [
4,
6].
At present, there are a number of conventional diagnostic methods for BC detection. These include cystoscopy, voided urine cytology, and detection of several urine-based markers such as NMP22 (nuclear matrix protein 22) and BTA (bladder tumor antigen). The most common way is cystoscopy coupled with voided urine cytology, which is highly sensitive [
8,
9]. However, cystoscopy is expensive, invasive, and uncomfortable for patients, and requires an experienced technician, which hinders its widespread application in BC diagnosis [
10]. Voided urine cytology only works well diagnosing BC in high grade and high stage, and thus hampers its utility in the detection of low-grade BC [
11]. Currently used urine-based biomarkers NMP22 and BTA are simple, quick, and non-invasive methods for BC screening [
12]. Unfortunately, these two biomarkers offer only low sensitivity and/or specificity detection. Hence, it is really urgent to find more sensitive but non-invasive biological markers to ensure the precise detection of early-stage BC.
miRNAs are a series of short (generally around 22 nucleotide long), single-stranded, and non-protein-coding RNA gene products that are vital for the regulation of gene expression. They work through binding their target mRNAs to cause degradation or translational silencing [
13]. Recent evidence has suggested that abnormal miRNA profiles are related to the development, progression, and prognosis of various human cancers [
14,
15]. miRNAs can present extensively in plasma, serum, and urine because they are protected from RNase degradation by some membrane-secreted vesicles and/or together with RNA-binding proteins [
16]. Thus non-invasive biomarkers for the diagnosis of BC may be developed from the urine-based miRNAs and circulating miRNAs [
17‐
24]. Unfortunately, the findings are inconclusive, which may be attributable to differences in sample size, ethnicity, miRNA profiling, and sample type. Therefore, a comprehensive analysis was administrated in this meta-analysis to further elucidate the diagnostic value of miRNAs in BC detection.
Methods
Documentation retrieval
The relevant literature in the ExcerptaMedica Database (EMBASE), PubMed, Chinese Biomedical Literature (CBM) database, and Chinese National Knowledge Infrastructure (CNKI) web database (updated until December 31, 2014) were searched using a combination of the terms “microRNAs” or “miRNA” or “miR”, “bladder cancer” or “bladder tumor” or “bladder urothelial cell carcinoma” and “sensitivity” or “specificity”, “ROC curve” or “diagnosis”. In order to retrieve the most relevant studies, manual filtration of the references was also conducted.
Inclusion and exclusion criteria
The adopted literature should meet the following criteria: (a) the diagnostic value of miRNAs for BC detection must have been evaluated, (b) the authors must have used the gold standards (histopathological examinations), and (c) the studies must contain detailed information for constructing two-by-two tables, which comprise true positives (TP), true negatives (TN), false positives (FP), and false negatives (FN). And exclusion criteria are as follows: (a) abstract, review, comment, editorial, and case reports, (b) overlapping data, (c) studies investigating survival or prognosis of BC, (d) deficient data, even on asking the corresponding authors for more information, and (e) sample size <100. If more than one study used some common samples, only the most precise one with the most samples was selected.
Data extraction and quality assessment
Data extraction from the selected publications was done using a standardized table by our two authors independently. And the extracted data contains first author’s surname, publication time, and country of origin, sample size, age, type of case, and control groups, sample specimen, studied miRNAs, detection method, TP, FP, FN, and TN, and information needed for quality assessment. For studies involving more than one type of miRNAs or sample specimen, the data were extracted, and the study on each miRNA was considered as an independent study. The quality of each article was evaluated by the revised quality assessment of diagnostic accuracy studies (QUADAS-2) checklists [
25]. All disagreements about the collected data were adequately debated by investigators and arrived at a final consensus.
Statistical analysis
For purpose of assessing the diagnostic value of miRNAs for BC, a bivariate meta-analysis was managed to obtain pooled negative likelihood ratio (NLR), positive likelihood ratio (PLR), sensitivity, diagnostic odds ratio (DOR), specificity, and corresponding 95% confidence intervals (CIs) [
26]. Meanwhile, the bivariate summary receiver operator characteristic (SROC) curve was drawn and the area under it (AUC) was calculated to display the sensitivity and specificity of each study. The heterogeneity assumption between the studies were also measured through chi-square based Cochran’s
Q and
I
2 tests. A
P value <0.10 and an
I
2 value >50% were considered as significant heterogeneity [
27,
28]. In order to examine the heterogeneity of origin, the sub-group analysis and meta-regression were performed according to the features (sample size, ethnicity, miRNA profiling, and specimen type). Finally, the potential publication bias of these literature was measured by Deeks’ funnel plot asymmetry test with the screening standard of
P value <0.05 [
29]. All
P values were two sided and all statistical analyses in this meta-analysis were performed by using the STATA statistical software (version 11.0; Stata Corp, College Station, Texas, USA).
Discussion
This meta-analysis is the first evidence-based analysis to assess the potential role of miRNAs for BC detection. This meta-analysis includes 31 studies from 10 researches covering 1556 cancer cases and 1347 controls in all, and the results indicate that miRNAs are potential novel and useful BC biomarkers with relatively high sensitivity and specificity that can be used for BC diagnosis.
Due to the high morbidity in men, high mortality for MIBC, and frequent recurrence, aggressive surveillance and timely intervention can dramatically improve the prognosis of BC. Currently, the method of the initial clinical diagnosis of BC—cystoscopy coupled with voided urine cytology—[
8,
9] is expensive, invasive, uncomfortable for patients, and requires technical expertise. A meta-analysis including 36 studies with 14,260 cases was administrated to appraise the diagnostic value of cytology, indicating that the pooled specificity for the method was up to 96% (95%CI 94–98%). However, the pooled sensitivity was only about 44% (95%CI 38–51%) [
30]. Several studies have also demonstrated that cytology has poor sensitivity for low-grade BC [
31,
32]. Although urine-based NMP22 and BTA are also being used as simple, quick, and non-invasive methods for BC screening, their diagnostic accuracy is not satisfactory and shows high variability. One study analyzed the function of NMP22 in the BC diagnosis in 1331 participants and showed that the sensitivity was only 55.7% [
33]. Other studies have reported sensitivities of NMP22 for BC diagnosis were ranging from 49.5 to 92.1%, whereas the specificities range from 66.0 to 87.3% [
11]. Similarly, urinary BTA tests have diagnostic sensitivities ranging from 29 to 91% and specificities from 56 to 85% [
11]. In comparison, miRNAs seem to have relatively high diagnostic accuracy with an overall pooled sensitivity of 0.72, specificity of 0.76, and an AUC of 0.80 in this meta-analysis.
It should be noted that available heterogeneity existed between these studies used in this paper; therefore, the effects of these confounding factors were explored via sub-group analyses. When stratified by miRNA profiling, the interesting finding was that single-miRNA assays showed relatively poor diagnostic performance with an AUC value of 0.74, while the diagnostic accuracy of multiple-miRNA assays is significantly higher with an AUC of 0.92. The results indicate that combination of multiple miRNAs might be more useful for the diagnosis of BC than single miRNAs. In addition, numerous studies have validated that single miRNAs can be used as accurate biomarkers for cancers. For example, Yang et al. performed a meta-analysis to evaluate the diagnostic performance of miRNA-21 in serum for lung carcinoma detection and suggested that miRNA-21 has diagnostic potential with an AUC value of 0.86 [
34]. Another meta-analysis indicated miRNA-21 was a reliable biomarker for diagnosing breast cancer [
35]. Wu et al. also proved miRNA-21 might be a potential biomarker for various early-stage cancers [
36]. However, the diagnostic accuracy estimated in these meta-analysis studies is still lower than that of multiple-miRNA assays. In addition, no specific miRNA was reported to be a reliable candidate biomarker for BC. Furthermore, other meta-analysis studies that focused on various cancers have also indicated that multiple miRNAs were more promising as diagnostic biomarkers than single miRNAs. The reason could be that the occurrence and development of a severe malignancy involves complicated, multiple epigenetic and genomic abnormalities.
When stratified by sample types, the results indicate that comparing to urine-based assays, blood-based assays could be a better choice as a diagnostic tool. Ding et al. carried out a meta-analysis to evaluate the diagnostic value of miRNAs for pancreatic cancer and made a similar observation that blood-based assays were significantly better than non-blood-based assays [
37]. However, Wei et al. performed a meta-analysis to measure the diagnostic value of miRNAs for cancers of the central nervous system and found that cerebrospinal fluid-based assays had more precise results than blood-based assays [
38]. These results imply that different sample types may result in different diagnostic accuracies for miRNA detection. MiRNAs in the urine are derived from tumor cells, which slough off; therefore, they may detect only high-grade cancers possibly in the final stages. However, miRNAs in the blood are secreted from tumor cells by microvesicles, and different expression profiles can be detected in the initial stages of cancer. Thus, detecting miRNAs in the blood may have more diagnostic value than detecting the same in the urine.
Furthermore, the sub-group analysis by ethnicity suggests that the diagnostic accuracy of miRNA assays is significantly better in Asian populations than that in Caucasian populations, with a pooled DOR of 25.0 vs. 6.0 and AUC of 0.90 vs. 0.76, respectively. Li et al. conducted a meta-analysis to evaluate the diagnostic value of miRNAs for hematologic malignancies and found that Asian population-based miRNA tests yield an overall higher accuracy than tests in Caucasian populations [
39]. However, other researchers did not observe any difference between Asian and Caucasian populations in their meta-analyses [
40]. Moreover, Cui et al. obtained the opposite results indicating that miRNA assays might be more accurate in Caucasian populations when used for the diagnosis of breast cancer [
41]. A potential explanation is that different ethnicities live in multiple environments and hold differing genetic backgrounds, lifestyles, and dietary habits, and therefore yield different miRNA expression profiles. The other potential explanation is that Caucasian populations come from Western developed countries, the medical services in Western developed countries are much better than developing countries in Asia, and they could accurately detect BC at an early stage; however, the change of miRNA expression profiles in early stage of tumor may not be obvious.
This study has several strengths notwithstanding a few limitations. This is the first meta-analysis to use as many as 1556 cancer cases and 1347 controls to summarize the diagnostic value of miRNAs in BC, which gives improved statistical power to the findings. The sub-group analysis confirmed that Asian population-based studies, multiple-miRNA profiling, and blood-based assays might yield increased diagnostic accuracy. Finally, this meta-regression analysis aimed to explore the sources of heterogeneity indicated that ethnicity, miRNA profiling, and sample type were independent confounding factors. However, the limitations of this study should also be acknowledged. Firstly, the cut-off parameters for miRNAs were different in the studies used and could be a potential source of heterogeneity. Secondly, this meta-analysis did not evaluate differences in the diagnostic accuracy of miRNAs presented in BCs with different clinicopathological features.