Diseases of the Esophagus (2015) ••, ••–•• DOI: 10.1111/dote.12338
Original article
Insights into the potential use of microRNAs as a novel class of biomarkers in esophageal cancer J. Wan,1 W. Wu,2 Y. Che,1 N. Kang,1 R. Zhang1 Departments of 1Thoracic Surgery and 2Hematology, The First Affiliated Hospital of Anhui Medical University, Hefei, China
SUMMARY. MicroRNAs (abbreviated miRNAs) have been demonstrated to be involved in tumorigenesis and cancer development and proposed as promising biomarkers in cancer diagnosis. Numerous studies have observed the aberrant expression of miRNAs in esophageal cancer. However, there are some discrepant results. Thus, we conducted this meta-analysis to identify the overall accuracy of miRNAs in the diagnosis of esophageal cancer. A comprehensive literature search was conducted in PubMed and other databases using combinations of key words. The summary receiver operator characteristic curves were plotted to assess the overall diagnostic performance of miRNAs. Chi-squared and I2 tests were used to assess the heterogeneity between studies. Additionally, we conducted subgroup and sensitivity analyses to analyze the potential sources of heterogeneity. In total, 33 studies from 12 articles were available in this meta-analysis. The pooled sensitivity, specificity, positive and negative likelihood ratio (PLR, NLR) diagnostic odds ratio, and area under the curve were 0.80, 0.80, 4.0, 0.25, 16, and 0.87, respectively. Subgroup analyses based on the sample types (saliva-, serum- and plasma-based) showed no differences in the diagnostic accuracy of each subgroup. An independent meta-analysis of eight articles was conducted to evaluate the diagnostic accuracy of miRNAs in patients with esophageal squamous cell carcinoma, with a pooled sensitivity of 0.77, specificity of 0.83, PLR of 4.4, NLR of 0.27, diagnostic odds ratio of 16, and area under the curve of 0.87. In conclusion, this meta-analysis demonstrates the feasibility of using miRNAs as non-invasive biomarkers to discriminate esophageal cancer from healthy controls. However, further high-quality studies on more clearly defined esophageal cancer patient are needed to confirm our conclusion. KEY WORDS: diagnosis, esophageal cancer, meta-analysis, microRNA.
INTRODUCTION Esophageal cancer represents the eighth most common lethal malignancies and the sixth leading cause of death from cancer worldwide, affecting three to four times more men than women.1 It shows a significant difference in geographic distribution, with high morbidity appearing in northern China, southeastern Africa, and Japan, and relatively low morbidity in western Africa and Central America.2 Based on the pathological characteristics, esophageal cancer is mainly composed of esophageal squamous cell carcinoma (ESCC) and adenocarcinoma.3 As the dominant subtype of esophageal cancer, ESCC accounts for ∼90% of esophageal cancer in Asian countries and Address correspondence to: Dr Renquan Zhang, M.M., Department of Thoracic Surgery, The First Affiliated Hospital of Anhui Medical University, No. 218 Jixi Road, Hefei 230022, China. Email: [email protected] Conflicts of interest: None. © 2015 International Society for Diseases of the Esophagus
is the second most common cancer in China.1,4 In spite of the improvements in surgical techniques and perioperative management, the overall survival rate is still poor, ranging from 3% to 5% in advanced stage or metastasis.5 However, the 5-year survival rate could increase to 90% if the tumors could be detected and treated at an early stage.6 Thus, the early detection of esophageal cancer may greatly improve clinical outcomes, and novel biomarkers that can facilitate disease detecting and staging are urgently needed to improve survival rates. At present, endoscopic biopsy and histopathological examinations are the golden standard methods for detecting esophageal cancer.7,8 However, there are many inherent deficiencies, including the use of invasive tool and the lack of a sufficient sensitivity and specificity.9 Besides, due to their tiring approaches, it is hard to be applied in high throughput of large-scale studies. Several studies have reported the biological behavior in esophageal 1
2
Diseases of the Esophagus
Cochrane Library, Chinese National Knowledge Infrastructure, and other sources to seek out the articles that assessed the diagnostic value of miRNAs in esophageal cancer up to August 6, 2014. The key words used for literature retrieval are as follows: ‘microRNAs, miRNA, miR’ and ‘esophageal cancer, esophageal neoplasm, esophageal carcinoma, ESCC, esophageal adenocarcinoma’ and ‘sensitivity, specificity, ROC curve, diagnosis’. No language restrictions were imposed on the search criteria.
cancer, and the certain molecules, including p53, cyclin D1, and FAS, are involved in tumorigenesis and the progress of ESCC.10–12 Studies by Kosugi et al. and Mroczko et al. have reported the use of carcinoembryonic antigen and squamous cell carcinoma antigen for early diagnosis of ESCC.13,14 However, these molecules showed poor diagnostic sensitivity, resulting in the diagnosis still at an advanced stage. Hence, the significance of developing novel and less invasive biomarkers for esophageal cancer diagnosis should be emphasized. MicroRNAs (miRNAs), a class of small (18–24 nucleotide) non-coding RNAs, have attracted novel attention for their involvement in cell differentiation, cycling, and apoptosis, through post-transcriptionally regulating gene expression by targeting specific miRNAs.15,16 Previous studies have demonstrated that the aberrant expression of miRNAs contributes to tumorigenesis and cancer development.17 MiRNA expression profiles are cell- or tissue- specific, which may help diagnose different cancer type and predict the progress of malignancies.18,19 Furthermore, tumorderived miRNAs are protected from the endogenous ribonuclease activity by joining with the secretory particles, such as apoptotic bodies and exosomes.20 Therefore, miRNAs existed in a stable form in the blood or other tissue fluid. Taken together, these findings open up an interesting field in the detecting and monitoring of esophageal cancer using miRNAs. A growing number of studies have reported the differential expression of miRNAs in esophageal cancer. For instance, a study by Feber et al. found that miR-194 and miR-192 were upregulated in esophageal adenocarcinoma compared with normal squamous epithelium, while miR-342 was upregulated in ESCC.21 Kurashige et al. found that miR-223 expression was markedly elevated in cancerous ESCC tissues.22 However, the published data on miRNAs in esophageal cancer have reached an inconsistent conclusion. Komatsu et al. reported a poor diagnostic accuracy with 60% sensitivity for miR-21, whereas Ye et al. reported a significantly higher sensitivity (97%).23 Interestingly, Wu et al. conducted a study of seven single miRNAs and its panel (miR-25, -100, -193-3p, -194, -223, -337-5p, -483-5p), and they found that both the single miRNA and the panel miRNAs yielded moderate diagnostic accuracy, with an area under the curve (AUC) ranging from 0.739 to 0.851.24 Thus, we performed this meta-analysis of all the available data to get an overall diagnostic characteristic of miRNAs as novel biomarkers in esophageal cancer.
This meta-analysis was performed based on the standard methods recommended for diagnostic accuracy. Stata software (version 12.0, College Station, TX, USA) was applied in the whole statistical analyses of this study. We used the bivariate meta-analysis model to calculate the summary parameters. The sensitivity and specificity of included studies were pooled to plot the summary receiver operator characteristic (sROC) curve and to calculate the AUC. Additionally, the heterogeneity among studies was assessed using the chi-squared and I2 tests. A value of P < 0.1 or I2 > 50% indicated significant heterogeneity. Subgroup and sensitivity analyses were also conducted to evaluate the potential sources of between-study heterogeneity. Finally, the publication bias was evaluated using Deeks et al. funnel plots. A two-sided P < 0.05 was considered statistically significant.
METHODS
RESULTS
Literature search
Search results and characteristics of studies
A comprehensive search strategy was carried out in numerous databases, including Medline, Embase, the
The systematic databases searches and manual review totally identified 346 records for initial search, and
Study selection and information extraction All relevant publications were included based on the following criteria: patients with esophageal cancer were definitely diagnosed based on endoscopic biopsy and histopathological examinations; studies must be about the use of miRNAs for esophageal cancer diagnosis; sufficient data were available for the generation of 2 × 2 tables, by which we can calculate the value of sensitivity and specificity. Additionally, the exclusion criteria are the following: studies with fewer than 20 patients; studies with duplicate subjects; conference abstracts, letters, editorials, or reviews were excluded because of the limited data. Information was retrieved from the included studies independently by two authors. More details are shown in Table 1. If studies contained both a training group and a validation group, each group was regarded as an independent study in this meta-analysis. Statistical analysis
© 2015 International Society for Diseases of the Esophagus
MicroRNAs in esophageal cancer
3
Table 1 Characteristics of included studies in this meta-analysis Included studies
Year
Location, race
Hirajima et al. Komatsu et al. Liu et al. Takeshita et al. Wang et al. Wu et al.
2013 2011 2012 2013 2012 2014
Japan, Asian Japan, Asian China, Asian Japan, Asian China, Asian China, Asian
Wu et al. Xie et al. Ye et al.
2013 2013 2014
China, Asian China, Asian China, Asian
Zhang et al.
2010
China, Asian
Zhang et al.
2011
China, Asian
Zhang et al.
2013
China, Asian
Case/control
Patient spectrum
Stage
Expression of microRNA
Sample
106/54 50/20 60/60 101/46 31/39 63/63
ESCC ESCC ESCC ESCC Esophageal cancer ESCC
I–IV I–IV I–IV I–IV I–IV I–IV
Plasma Plasma Plasma Serum Serum Serum
67/50 39/19 100/50 100/50 149/100
Esophageal cancer Esophageal cancer Esophageal cancer Esophageal cancer ESCC
I–IV I–IV I–II I–II I–IV
120/121 81/81 120/120 81/81
ESCC ESCC ESCC ESCC
I–IV I–IV I–IV I–IV
Up: miR-18a Up: miR-21; down: miR-375 Down: miR-155 Up: miR-1246 Up: miR-21 Up: miR-25, -100, -193-3p, -194, -223, -337-5p, -483-5p Up: miR-144 Up: miR-10b, -144, -451 Up: miR-21 Up: miR-21 Up: miR-10a, -22, -100, -148b, -223, -133a, -127-3p Up: miR-31 Up: miR-31 Up: miR-1322 Up: miR-1322
Saliva Saliva Plasma Saliva Serum Serum Serum Serum Serum
ESCC, esophageal squamous cell carcinoma.
after removing duplicate records (n = 138), 278 unique abstracts remained. After checking the titles and abstracts, 36 articles were remained for further full-text review. Of these, 12 articles were considered appropriate for this meta-analysis according to the inclusion criteria. Table 1 shows the general characteristics of these included publications. In total, 33 studies from 12 articles covering 2930 patients and 2173 healthy controls were available in this meta-analysis.9,23–33 Of all included studies, 26 studies from eight publications focused on miRNAs as diagnostic biomarkers for ESCC. All these studies reported a total of 22 types of miRNAs. The miRNA expression profiling of the included studies is also summarized in Table 1, and only miR-375 and miR-155 were downregulated in ESCC patients compared with control individuals. Thirty studies discussed a single miRNA, and the remaining three studies focused on a panel miRNAs. The specimens were composed of saliva (5), serum (22), and plasma (6). All the included studies were conducted in Asian populations, of which five studies were performed in Japan, while 28 studies were conducted in China, which may cause a discrepancy of country.
95% confidence intervals (CIs) are as follows: sensitivity, 0.80 (95% CI: 0.76–0.83); specificity, 0.80 (95% CI: 0.76–0.84); PLR, 4.0 (95% CI: 3.3–4.8); NLR, 0.25 (95% CI: 0.21–0.30); and diagnostic odds ratio (DOR), 16 (95% CI: 12–21). Additionally, the sROC curve was plotted in Figure 2a, and the AUC was estimated to be 0.87 (95% CI: 0.84–0.90). Subgroup analyses based on the sample types (saliva-, serum-, and plasma-based) were performed to explore the potential sources of heterogeneity between studies. The sROC curves for the three subgroups are plotted in Figure 2b–d. The pooled diagnostic parameters for each subgroup are listed in Table 2. For saliva specimens, the pooled sensitivity was 0.87, specificity was 0.67, PLR was 2.7, NLR was 0.19, DOR was 14, and the AUC was 0.88; for serum specimens, the pooled sensitivity was 0.77, specificity was 0.83, PLR was 4.5, NLR was 0.27, DOR was 16, and the AUC was 0.87; and for plasma specimens, the pooled sensitivity was 0.82, specificity was 0.79, PLR was 3.9, NLR was 0.22, DOR was 18, and the AUC was 0.88. Noticeable heterogeneity was also observed among studies of each subgroup. Therefore, specimens may not be possible sources of heterogeneity in this meta-analysis.
MiRNAs in diagnosis of esophageal cancer
MiRNAs in diagnosis of ESCC
Figure 1 shows the forest plot of 33 studies with its corresponding sensitivity and specificity for miRNAs assays in the diagnosis of esophageal cancer. Chisquared and I2 tests values of sensitivity and specificity were 150.03 (P < 0.001), and 78.67, 127.59 (P < 0.001), and 74.92, respectively, suggesting significant heterogeneity between studies. Thus, we used the random effect model to calculate the pool estimates. Pooled results for diagnostic accuracy are summarized in Table 2. The results with their corresponding
It has been widely known that ESCC accounts for ∼90% of esophageal cancer in Asian countries. In addition, of the 12 articles included in the metaanalysis of miRNAs associated with esophageal cancer, there are eight articles covering 26 studies that discussed about the relationships between miRNAs and ESCC. Therefore, in this study, we conducted an independent meta-analysis of diagnostic accuracy of miRNAs in discriminating ESCC from healthy controls. The sROC curve is graphed in Figure 3, with an
© 2015 International Society for Diseases of the Esophagus
4
Diseases of the Esophagus
(a)
(b)
Included study
Sensitivity (95% CI)
Included study
Specificity (95% CI)
Zhang et al./2013 Zhang et al./2013 Zhang et al./2011 Zhang et al./2011 Zhang et al./2010 Zhang et al./2010 Zhang et al./2010 Zhang et al./2010 Zhang et al./2010 Zhang et al./2010 Zhang et al./2010 Zhang et al./2010 Ye et al./2014 Ye et al./2014 Xie et al./2013 Xie et al./2013 Xie et al./2013 Wu et al./2013 Wu et al./2014 Wu et al./2014 Wu et al./2014 Wu et al./2014 Wu et al./2014 Wu et al./2014 Wu et al./2014 Wu et al./2014 Wang et al./2012 Takeshita et al./2013 Liu et al./2012 Komatsu et al./2011 Komatsu et al./2011 Komatsu et al./2011 Hirajima et al./2013
0.84 [0.74 − 0.91] 0.82 [0.74 − 0.88] 0.86 [0.77 − 0.93] 0.87 [0.79 − 0.92] 0.79 [0.71 − 0.85] 0.65 [0.57 − 0.73] 0.83 [0.76 − 0.89] 0.66 [0.58 − 0.74] 0.64 [0.55 − 0.71] 0.89 [0.82 − 0.93] 0.81 [0.74 − 0.87] 0.79 [0.71 − 0.85] 0.89 [0.81 − 0.94] 0.97 [0.91 − 0.99] 0.85 [0.69 − 0.94] 0.92 [0.79 − 0.98] 0.90 [0.76 − 0.97] 0.75 [0.63 − 0.84] 0.81 [0.69 − 0.90] 0.75 [0.62 − 0.85] 0.78 [0.66 − 0.87] 0.65 [0.52 − 0.77] 0.65 [0.52 − 0.77] 0.76 [0.64 − 0.86] 0.78 [0.66 − 0.87] 0.83 [0.71 − 0.91] 0.71 [0.52 − 0.86] 0.71 [0.61 − 0.80] 0.62 [0.48 − 0.74] 0.76 [0.62 − 0.87] 0.60 [0.45 − 0.74] 0.88 [0.76 − 0.95] 0.87 [0.79 − 0.93]
Zhang et al./2013 Zhang et al./2013 Zhang et al./2011 Zhang et al./2011 Zhang et al./2010 Zhang et al./2010 Zhang et al./2010 Zhang et al./2010 Zhang et al./2010 Zhang et al./2010 Zhang et al./2010 Zhang et al./2010 Ye et al./2014 Ye et al./2014 Xie et al./2013 Xie et al./2013 Xie et al./2013 Wu et al./2013 Wu et al./2014 Wu et al./2014 Wu et al./2014 Wu et al./2014 Wu et al./2014 Wu et al./2014 Wu et al./2014 Wu et al./2014 Wang et al./2012 Takeshita et al./2013 Liu et al./2012 Komatsu et al./2011 Komatsu et al./2011 Komatsu et al./2011 Hirajima et al./2013
0.80 [0.70 − 0.88] 0.82 [0.75 − 0.89] 0.79 [0.69 − 0.87] 0.84 [0.77 − 0.90] 0.87 [0.79 − 0.93] 0.83 [0.74 − 0.90] 0.83 [0.74 − 0.90] 0.87 [0.79 − 0.93] 0.81 [0.72 − 0.88] 0.86 [0.78 − 0.92] 0.80 [0.71 − 0.87] 0.96 [0.90 − 0.99] 0.64 [0.49 − 0.77] 0.56 [0.41 − 0.70] 0.58 [0.33 − 0.80] 0.47 [0.24 − 0.71] 0.58 [0.33 − 0.80] 0.92 [0.81 − 0.98] 0.81 [0.69 − 0.90] 0.75 [0.62 − 0.85] 0.86 [0.75 − 0.93] 0.79 [0.67 − 0.89] 0.90 [0.80 − 0.96] 0.86 [0.75 − 0.93] 0.75 [0.62 − 0.85] 0.75 [0.62 − 0.85] 0.69 [0.52 − 0.83] 0.74 [0.59 − 0.86] 0.65 [0.52 − 0.77] 0.85 [0.62 − 0.97] 0.60 [0.36 − 0.81] 0.70 [0.46 − 0.88] 1.00 [0.93 − 1.00]
COMBINED
0.80 [0.76 − 0.83] Q = 150.03, df = 32.00, P < 0.01 2 I = 78.67 [71.79 − 85.55]
COMBINED
0.80 [0.76 − 0.84] Q = 127.59, df = 32.00, P < 0.01 2 I = 74.92 [66.46 − 83.38]
0.5
1.0
0.2
1.0
Fig. 1 The forest plots of microRNA (a) sensitivity and (b) specificity in esophageal cancer diagnosis with the corresponding heterogeneity. The sensitivity and specificity from each study are represented by square, and the confidence interval (CI) is indicated by error bars: the pooled sensitivity, 0.80 (95% CI: 0.76–0.83); the pooled specificity, 0.80 (95% CI: 0.76–0.84).
AUC of 0.87 (95% CI: 0.84–0.90). The other pooled diagnostic parameters are listed in Table 2. The results are as follows: sensitivity, 0.77 (95% CI: 0.74– 0.81); specificity, 0.83 (95% CI: 0.79–0.85); PLR, 4.4 (95% CI: 3.7–5.4); NLR, 0.27 (95% CI: 0.23–0.32); and DOR, 16 (95% CI: 12–22). The diagnostic values of miRNAs in ESCC diagnosis were similar to that in esophageal cancer diagnosis, which suggested that miRNAs are promising biomarkers for ESCC detection, no matter whether the subtypes of esophageal cancer were identified in Asian people.
Sensitivity and outlier detection analyses We conducted the goodness of fit and bivariate normality analyses to assess the reliability of using the random effects bivariate model for the calculation of the pooled estimates (Fig. 4). The results suggested that the selected model was robust for the calculation of the pooled estimates. Additionally, in order to affirm that our findings were not driven by any individual study, we further performed the outlier detection analyses. After excluding the four
Table 2 Comparisons of assays based on sample for diagnosing esophageal cancer
Analyses
Sensitivity (95% CI)
Specificity (95% CI)
Positive LR (95% CI)
Negative LR (95% CI)
DOR (95% CI)
AUC (95% CI)
Combined Saliva Serum Plasma ESCC Excluding outliers
0.80 (0.76–0.83) 0.87 (0.80–0.92) 0.77 (0.74–0.81) 0.82 (0.67–0.91) 0.77 (0.74–0.81) 0.78 (0.74–0.81)
0.80 (0.76–0.84) 0.67 (0.49–0.82) 0.83 (0.80–0.85) 0.79 (0.55–0.92) 0.83 (0.79–0.85) 0.80 (0.77–0.82)
4.0 (3.3–4.8) 2.7 (1.6–4.3) 4.5 (3.9–5.1) 3.9 (1.6–10.0) 4.4 (3.7–5.4) 3.9 (3.4–4.4)
0.25 (0.21–0.30) 0.19 (0.14–0.27) 0.27 (0.23–0.32) 0.22 (0.11–0.46) 0.27 (0.23–0.32) 0.28 (0.24–0.32)
16 (12–21) 14 (8–25) 16 (13–21) 18 (4–75) 16 (12–22) 14 (11–18)
0.87 (0.84–0.90) 0.88 (0.85–0.90) 0.87 (0.84–0.90) 0.88 (0.85–0.90) 0.87 (0.84–0.90) 0.86 (0.82–0.89)
AUC, area under the curve; CI, confidence interval; DOR, diagnostic odds ratio; ESCC, esophageal squamous cell carcinoma; LR, likelihood ratio. © 2015 International Society for Diseases of the Esophagus
MicroRNAs in esophageal cancer
(a)
(b)
1.0
1.0
20
3
18 24 30
31 27 33 3215 23
2913 104
22 16
2
Sensitivity
26
2
5
19
4
8 9 1
14 6
11
17
21
7
28 12 25
Sensitivity
1
5 3
0.5
0.0 1.0
0.5 Specificity
0.5
0.0 1.0
0.0
(c)
0.5 Specificity
0.0
(d)
1.0
1.0 13 19
188 5
4
15
4
9 1
6
17
14
2
7
0.5
0.0 1.0
2
3
Sensitivity
11
6
1
20
16 22 2110 12
Sensitivity
5
0.5 Specificity
0.0
5 3
0.5
0.0 1.0
0.5 Specificity
0.0
Fig. 2 The summary receiver operator characteristic (sROC) curves based on subgroup analyses. (a) overall studies [(○) observed data; (◆) summary operating point, SENS = 0.80 (0.76–0.83), SPEC = 0.80 (0.76–0.84); (—) sROC curve, AUC = 0.87 (0.84–0.90); (---) 95% confidence contour; (. . .) 95% prediction contour)]; (b) saliva-based miRNA assay [(○) observed data; (◆) summary operating point, SENS = 0.87 (0.80–0.92), SPEC = 0.67 (0.49–0.82); (---) sROC curve, AUC = 0.88 (0.85–0.90); (—) 95% confidence contour; (. . .) 95% prediction contour)]; (c) serum-based miRNA assay [(○) observed data; (◆) summary operating point, SENS = 0.77 (0.74–0.81), SPEC = 0.83 (0.80–0.85); (—) sROC curve, AUC = 0.87 (0.84–0.90); (---) 95% confidence contour; (. . .) 95% prediction contour)]; (d) plasma-based miRNA assay [(○) observed data; (◆) summary operating point, SENS = 0.82 (0.67–0.91), SPEC = 0.79 (0.55–0.92); (—) sROC curve, AUC = 0.88 (0.85–0.90); (---) 95% confidence contour; (. . .) 95% prediction contour)].
outlier studies, the I2 value for heterogeneity decreased, from 78.67% to 75.29% for sensitivity, and from 74.92% to 57.06% for specificity. Compared with the overall analysis, after excluding the four outlier studies, the pooled diagnostic estimates were only minimally changed. The data before and after excluding four outliers are listed in Table 2, with sensitivity of 0.80 versus 0.78, specificity of © 2015 International Society for Diseases of the Esophagus
0.80 versus 0.80, PLR of 4.0 versus 3.9, NLR of 0.25 versus 0.28, DOR of 16 versus 14, and AUC of 0.87 versus 0.86, respectively. Publication bias Considering that publication bias is a key factor that may influence the results of meta-analysis, we
6
Diseases of the Esophagus
ESCC studies, which suggested symmetry in the included articles and a relatively low likelihood of publication bias.
1.0 17 23
1
15
2
24
20 26 2514 16 2212 94
7 8 13 6
19
21
11 18
DISCUSSION
5 3
0.5
0.0 1.0
0.5
0.0
Specificity
Fig. 3 The summary receiver operator characteristic (sROC) curve of the diagnostic studies of miRNAs in discriminating esophageal squamous cell carcinoma from healthy subjects (○) observed data; (◆) summary operating point, SENS = 0.77 (0.74–0.81), SPEC = 0.83 (0.79–0.85); (—) sROC curve, AUC = 0.87 (0.84–0.90); (—) 95% confidence contour; (. . .) 95% prediction contour.
(a)
(b)
1.00
1.00
Mahalanobis D−square
Deviance residual
performed the Deeks et al.’s funnel plot with asymmetry test to explore publication bias in this study. The slope coefficient reflected a P-value of 0.11 in esophageal cancer studies and a P-value of 0.21 in
0.75 0.50 0.25 0.00 0.00
0.25 0.50 0.75 Normal quantile
3.00 20 1
2.00 22
1.00 3
5
18
0.00 0
10
20 Study
30
0.75 0.50 0.25 0.00
1.00
(c)
Cook’s distance
A non-invasive assay using miRNAs opens up a novel and interesting field in the screening and diagnosing of cancer patients. Experimental evidence has demonstrated that an abundance of human miRNAs genes is located at fragile sites and genomic regions, which may function as oncogenes or tumor suppressor genes.34 Tumor-specific alterations of miRNAs in the blood or other body fluid have been identified in patients with several malignancies, such as prostate cancer,35 oral cancer,36 lung cancer,37 breast cancer,38 pancreatic cancer,39 colorectal cancer,40 ovarian cancer,41 and esophageal cancer.42 In addition, studies have demonstrated that miRNAs are presented in plasma/serum in a remarkably stable form, and the expression level of miRNAs is reproducible and consistent among individuals.43 Thus, miRNAs may provide new therapeutic strategies such as biomarkers for cancers detection. Esophageal cancer is one of the most common types of malignant tumors worldwide.1 Despite the improvements in surgical techniques, esophageal
40
0.00
Standardized residual(diseased)
Sensitivity
10
0.25 0.50 0.75 Chi-squared quantile
1.00
(d)
3.0
20
2.0
18 17 21
1.0
24
2 19 8
0.0 7
−1.0 5 3
9 14 6
30 31 33 27 2315 32 25 12 19 26 17 28 18 3 5 20 23 15 32 33 13 27 31 10 29 14 22 30 16 4 8 9 1 6 2 7 21 24 11
1
13 29 4 10
12 25 28
26
22 16 11
−2.0 −3.0 −3.0
−2.0
−1.0 0.0 1.0 2.0 Standardized residual
3.0
Fig. 4 Graphical depiction of residual-based (a) goodness of fit, (b) bivariate normality, (c) influence, and (d) outlier detection analyses, respectively. © 2015 International Society for Diseases of the Esophagus
MicroRNAs in esophageal cancer
cancer is regarded as one of the most lethal carcinomas of the gastrointestinal tract, with a dismal overall 5-year survival rate. At present, radiology and endoscopic biopsy are the most common diagnostic methods.8 However, these tests are invasive and time-consuming, and they may cause discomfort to patients. Patients with esophageal cancer are usually diagnosed at an advanced stage or at a stage with the generation of lymph nodes, which remains the main cause of poor survival rate.3 Therefore, developing a biomarker of early-stage esophageal cancer is essential and continuous. In our meta-analysis, after a comprehensive literature research, we finally included 12 articles covering 33 studies to calculate the pooled diagnostic estimates. The overall analysis yield relatively high diagnostic performances of miRNAs with an AUC of 0.87, a sensitivity of 0.80, and a specificity of 0.80. A DOR of 1.0 shows no ability to discriminate between patients with the disease and those without it. The mean DOR in this meta-analysis was 16, suggesting that miRNAs are powerful biomarkers for esophageal cancer patient diagnosis. PLR and NLR are identified as more clinically meaningful estimates of diagnostic accuracy in clinical practice. A value of PLR greater than 10 and NLR lower than 0.1 are considered the threshold for reliability.44 In this metaanalysis, the pooled PLR and NLR were 4.0 and 0.25, respectively. Although the PLR and NLR are less encouraging, miRNAs are still relatively accurate in diagnosing esophageal cancer in terms of the pooled values of AUC and DOR. As the dominant subtype of esophageal cancer, ESCC is the second most common cancer and the fourth leading cause of cancer-related mortality. Accumulating studies have showed that aberrant expression of miRNAs may be involved in the initiation and progression of ESCC. Therefore, an independent meta-analysis of 26 included studies was conducted to evaluate the diagnostic efficacy of miRNAs in discriminating ESCC from healthy controls. MiRNAs achieved a pooled sensitivity of 0.77, specificity of 0.83, PLR of 4.4, NLR of 0.27, DOR of 16, and AUC of 0.87. In contrast to miRNAs in esophageal cancer diagnosis, the diagnostic values of miRNAs in ESCC diagnosis showed no significant differences. The few esophageal adenocarcinoma or other types of esophageal cancer cases in the included studies may contribute to the results. The roles of miRNAs in carcinogenesis, progression, invasion, and metastasis of esophageal cancer have been widely studied, and the mechanisms are gradually excluded. MiRNAs act as onco-miRNA or tumor suppressors by targeting to the tumor suppressor genes or ontogenesis.45 A study by Ma et al. has proven that miR-21 promotes cell proliferation by targeting chromosome-10 (PTEN) at posttranscriptional level, and inhibits cell growth and © 2015 International Society for Diseases of the Esophagus
7
invasion by targeting FASL, TIMP3, and RECK genes.46,47 MiR-10a was reported to be downregulated in esophageal cancer and to affect cell migration and invasion through targets in homeobox genes, whereas miR-10b was overexpressed and affected the cell motility and invasiveness by suppressing the tumor suppressor gene KLF4.48,49 Zhang et al. demonstrated that upregulation of miR-31 promoted the tumor formation through the EMP1, KSR2, and RGS4 genes.32 As a growing number of researchers studied the relationships between miRNAs and esophageal cancer, the exact role of miRNA in oncogenesis of esophageal cancer could be clearly elucidated in the future. It has been reported that the specific miRNA profiles were closely related to prognosis and histopathological features of malignancies, such as stages and types. Iorio et al. found that miR-145 and miR-21 were differentially expressed in breast cancer with different tumor stage, suggesting that these two miRNAs were involved in tumor progression and could be acquired in the identification of tumor progression.50 Since esophageal cancers diagnosed at stage I or II have a good chance of improving survival rate, we expect to identify specific miRNAs unique to esophageal cancer in different stages. However, only one article investigated the early stage (I–II) of esophageal cancer patients in this meta-analysis. Therefore, in order to clarify whether miRNA expressed differentially in esophageal cancers with different tumor stage, further original studies have been performed. When getting a reasonable conclusion of any metaanalyses, it is indispensable to consider the interstudy heterogeneity. In our meta-analysis, subgroup analyses based on the sample types (saliva-, serum-, and plasma-based) were performed to explore the potential sources of heterogeneity between studies. The pooled results showed no obvious differences in each subgroup, and broad heterogeneities also existed in the three subgroups, which suggest that specimens may not be possible sources of heterogeneity in this meta-analysis. In view of our sensitivity analysis, after excluding the four outlier studies, the pooled results did not change much, indicating that our findings were not driven by any single study. A previous meta-analysis by Fu et al. shows that miR-21 and miR-375 can be used as prognostic biomarkers in esophageal cancer.51 However, no meta-analyses on the overall accuracy of miRNAs in diagnosing esophageal cancer are available at present. Thus, we conduct this meta-analysis to assess the diagnostic performances of miRNAs in esophageal cancer diagnosis. Furthermore, we also identified the diagnostic accuracy of miRNAs in ESCC diagnosis by performing an independent meta-analysis of 26 studies. However, there are still several shortcomings in the research. First, of the 12 articles included, only
8
Diseases of the Esophagus
three publications were conducted in Japan, and most of the studies were in performed in China, which may lead a country selection bias. Second, patients in some studies were not clearly classified, and thus we did not know the certain subtypes of the esophageal cancer patients. Last, because of a limited number of articles, subgroups analysis on ethnicity, miRNAs profiling and stages could not be conducted to explore the potential sources of heterogeneity between studies. In conclusion, this meta-analysis demonstrates the feasibility of using miRNAs as non-invasive biomarkers to discriminate esophageal cancer from healthy controls. However, further high-quality studies are needed to confirm our conclusion, and in order to efficiently apply these findings in clinic, studies on more clearly defined esophageal cancer patient are warranted.
14
15 16 17 18 19 20 21
Acknowledgments
22
This work was supported by a grant from the Natural Science Foundation of China (No. 81302028): The biological significance of HIF-1a expression in transition zone of small cell lung cancer after radiofrequency ablation.
23
References
25
1 Enzinger P C, Mayer R J. Esophageal cancer. NEJM 2003; 349: 2241–52. 2 Jemal A, Bray F, Center M M, Ferlay J, Ward E, Forman D. Global cancer statistics. CA Cancer J Clin 2011; 61: 69–90. 3 He B, Yin B, Wang B, Xia Z, Chen C, Tang J. MicroRNAs in esophageal cancer (review). Mol Med Rep 2012; 6: 459–65. 4 Hiyama T, Yoshihara M, Tanaka S, Chayama K. Genetic polymorphisms and esophageal cancer risk. Int J Cancer 2007; 121: 1643–58. 5 Kim T, Grobmyer S R, Smith R et al. Esophageal cancer – the five year survivors. J Surg Oncol 2011; 103: 179–83. 6 Daly J M, Fry W A, Little A G et al. Esophageal cancer: results of an American College of Surgeons Patient Care Evaluation Study. J Am Coll Surg 2000; 190: 562–72, discussion 72–3. 7 Barber T W, Duong C P, Leong T, Bressel M, Drummond E G, Hicks R J. 18F-FDG PET/CT has a high impact on patient management and provides powerful prognostic stratification in the primary staging of esophageal cancer: a prospective study with mature survival data. J Nucl Med 2012; 53: 864–71. 8 Tachimori Y, Kanamori N, Uemura N, Hokamura N, Igaki H, Kato H. Salvage esophagectomy after high-dose chemoradiotherapy for esophageal squamous cell carcinoma. J Thorac Cardiovasc Surg 2009; 137: 49–54. 9 Hirajima S, Komatsu S, Ichikawa D et al. Clinical impact of circulating miR-18a in plasma of patients with oesophageal squamous cell carcinoma. Br J Cancer 2013; 108: 1822–9. 10 Hollstein M C, Metcalf R A, Welsh J A, Montesano R, Harris C C. Frequent mutation of the p53 gene in human esophageal cancer. Proc Natl Acad Sci U S A 1990; 87: 9958–61. 11 Adelaide J, Monges G, Derderian C, Seitz J F, Birnbaum D. Oesophageal cancer and amplification of the human cyclin D gene CCND1/PRAD1. Br J Cancer 1995; 71: 64–8. 12 Gratas C, Tohma Y, Barnas C, Taniere P, Hainaut P, Ohgaki H. Up-regulation of Fas (APO-1/CD95) ligand and downregulation of Fas expression in human esophageal cancer. Cancer Res 1998; 58: 2057–62. 13 Kosugi S, Nishimaki T, Kanda T, Nakagawa S, Ohashi M, Hatakeyama K. Clinical significance of serum carcinoem-
24
26
27 28
29
30 31
32
33 34
35
bryonic antigen, carbohydrate antigen 19-9, and squamous cell carcinoma antigen levels in esophageal cancer patients. World J Surg 2004; 28: 680–5. Mroczko B, Kozlowski M, Groblewska M, et al. The diagnostic value of the measurement of matrix metalloproteinase 9 (MMP-9), squamous cell cancer antigen (SCC) and carcinoembryonic antigen (CEA) in the sera of esophageal cancer patients. Clinica chimica acta; International journal of clinical chemistry 2008; 389(1–2): 61–6. Krol J, Loedige I, Filipowicz W. The widespread regulation of microRNA biogenesis, function and decay. Nat Rev Genet 2010; 11: 597–610. Farazi T A, Hoell J I, Morozov P, Tuschl T. MicroRNAs in human cancer. Adv Exp Med Biol 2013; 774: 1–20. Paranjape T, Slack F J, Weidhaas J B. MicroRNAs: tools for cancer diagnostics. Gut 2009; 58: 1546–54. Madhavan D, Cuk K, Burwinkel B, Yang R. Cancer diagnosis and prognosis decoded by blood-based circulating microRNA signatures. Front Genet 2013; 4: doi:10.3389/fgene.2013.00116. Calin G A, Croce C M. MicroRNA signatures in human cancers. Nat Rev Cancer 2006; 6: 857–66. Mitchell P S, Parkin R K, Kroh E M et al. Circulating microRNAs as stable blood-based markers for cancer detection. Proc Natl Acad Sci U S A 2008; 105: 10513–18. Feber A, Xi L, Luketich J D et al. MicroRNA expression profiles of esophageal cancer. J Thorac Cardiovasc Surg 2008; 135: 255–60, discussion 60. Kurashige J, Watanabe M, Iwatsuki M et al. Overexpression of microRNA-223 regulates the ubiquitin ligase FBXW7 in oesophageal squamous cell carcinoma. Br J Cancer 2012; 106: 182–8. Komatsu S, Ichikawa D, Takeshita H et al. Circulating microRNAs in plasma of patients with oesophageal squamous cell carcinoma. Br J Cancer 2011; 105: 104–11. Wu C, Wang C, Guan X et al. Diagnostic and prognostic implications of a serum miRNA panel in oesophageal squamous cell carcinoma. PLoS ONE 2014; 9: e92292. Ye M, Ye P, Zhang W, Rao J, Xie Z. Diagnostic values of salivary versus and plasma microRNA-21 for early esophageal cancer. Nan Fang Yi Ke Da Xue Xue Bao 2014; 34: 885– 9. Wu W, Hou W, Wu Z, Wang Y, Yi Y, Lin W. miRNA-144 in the saliva is a genetic marker for early diagnosis of esophageal cancer. Nan Fang Yi Ke Da Xue Xue Bao 2013; 33: 1783– 6. Xie Z, Chen G, Zhang X et al. Salivary microRNAs as promising biomarkers for detection of esophageal cancer. PLoS ONE 2013; 8: e57502. Takeshita N, Hoshino I, Mori M et al. Serum microRNA expression profile: miR-1246 as a novel diagnostic and prognostic biomarker for oesophageal squamous cell carcinoma. Br J Cancer 2013; 108: 644–52. Liu R, Liao J, Yang M et al. Circulating miR-155 expression in plasma: a potential biomarker for early diagnosis of esophageal cancer in humans. J Toxicol Environ Health A 2012; 75: 1154– 62. Wang B, Zhang Q. The expression and clinical significance of circulating microRNA-21 in serum of five solid tumors. J Cancer Res Clin Oncol 2012; 138: 1659–66. Lou X, Qi X, Zhang Y, Long H, Yang J. Decreased expression of microRNA-625 is associated with tumor metastasis and poor prognosis in patients with colorectal cancer. J Surg Oncol 2013; 108: 230–5. Zhang T, Wang Q, Zhao D et al. The oncogenetic role of microRNA-31 as a potential biomarker in oesophageal squamous cell carcinoma. Clin Sci (Lond) 2011; 121: 437– 47. Zhang C, Wang C, Chen X et al. Expression profile of microRNAs in serum: a fingerprint for esophageal squamous cell carcinoma. Clin Chem 2010; 56: 1871–9. Calin G A, Sevignani C, Dumitru C D et al. Human microRNA genes are frequently located at fragile sites and genomic regions involved in cancers. Proc Natl Acad Sci U S A 2004; 101: 2999–3004. Bryant R J, Pawlowski T, Catto J W et al. Changes in circulating microRNA levels associated with prostate cancer. Br J Cancer 2012; 106: 768–74. © 2015 International Society for Diseases of the Esophagus
MicroRNAs in esophageal cancer 36 Ries J, Vairaktaris E, Agaimy A et al. miR-186, miR-3651 and miR-494: potential biomarkers for oral squamous cell carcinoma extracted from whole blood. Oncol Rep 2014; 31: 1429– 36. 37 Sozzi G, Boeri M, Rossi M et al. Clinical utility of a plasmabased miRNA signature classifier within computed tomography lung cancer screening: a correlative MILD trial study. J Clin Oncol 2014; 32: 768–73. 38 McDermott A M, Miller N, Wall D et al. Identification and validation of oncologic miRNA biomarkers for luminal A-like breast cancer. PLoS ONE 2014; 9: e87032. 39 Zhang J, Zhao C Y, Zhang S H et al. Upregulation of miR-194 contributes to tumor growth and progression in pancreatic ductal adenocarcinoma. Oncol Rep 2014; 31: 1157–64. 40 Zhang L, Meng L, Fan Z, Liu B, Pei Y, Zhao Z. Expression of plasma miR-106a in colorectal cancer and its clinical significance. Nan Fang Yi Ke Da Xue Xue Bao 2014; 34: 354–7. 41 Zheng H, Zhang L, Zhao Y et al. Plasma miRNAs as diagnostic and prognostic biomarkers for ovarian cancer. PLoS ONE 2013; 8: e77853. 42 Javidi M A, Ahmadi A H, Bakhshinejad B, Nouraee N, Babashah S, Sadeghizadeh M. Cell-free microRNAs as cancer biomarkers: the odyssey of miRNAs through body fluids. Med Oncol 2014; 31: doi:10.1007/s12032-014-0295-y. 43 Kosaka N, Iguchi H, Yoshioka Y, Takeshita F, Matsuki Y, Ochiya T. Secretory mechanisms and intercellular transfer of
© 2015 International Society for Diseases of the Esophagus
44 45 46
47 48 49 50 51
9
microRNAs in living cells. J Biol Chem 2010; 285: 17442– 52. Deeks J J, Altman D G. Diagnostic tests 4: likelihood ratios. BMJ 2004; 329: 168–9. Esquela-Kerscher A, Slack F J. Oncomirs – microRNAs with a role in cancer. Nat Rev Cancer 2006; 6: 259–69. Wang N, Zhang C Q, He J H et al. MiR-21 down-regulation suppresses cell growth, invasion and induces cell apoptosis by targeting FASL, TIMP3, and RECK genes in esophageal carcinoma. Dig Dis Sci 2013; 58: 1863–70. Ma W J, Lv G D, Tuersun A et al. Role of microRNA-21 and effect on PTEN in Kazakh’s esophageal squamous cell carcinoma. Mol Biol Rep 2011; 38: 3253–60. Matsushima K, Isomoto H, Kohno S, Nakao K. MicroRNAs and esophageal squamous cell carcinoma. Digestion 2010; 82: 138–44. Tian Y, Luo A, Cai Y et al. MicroRNA-10b promotes migration and invasion through KLF4 in human esophageal cancer cell lines. J Biol Chem 2010; 285: 7986–94. Iorio M V, Ferracin M, Liu C G et al. MicroRNA gene expression deregulation in human breast cancer. Cancer Res 2005; 65: 7065–70. Fu W, Pang L, Chen Y, Yang L, Zhu J, Wei Y. The microRNAs as prognostic biomarkers for survival in esophageal cancer: a meta-analysis. Scientificworldjournal 2014; 2014: 523979.
Original article
Insights into the potential use of microRNAs as a novel class of biomarkers in esophageal cancer J. Wan,1 W. Wu,2 Y. Che,1 N. Kang,1 R. Zhang1 Departments of 1Thoracic Surgery and 2Hematology, The First Affiliated Hospital of Anhui Medical University, Hefei, China
SUMMARY. MicroRNAs (abbreviated miRNAs) have been demonstrated to be involved in tumorigenesis and cancer development and proposed as promising biomarkers in cancer diagnosis. Numerous studies have observed the aberrant expression of miRNAs in esophageal cancer. However, there are some discrepant results. Thus, we conducted this meta-analysis to identify the overall accuracy of miRNAs in the diagnosis of esophageal cancer. A comprehensive literature search was conducted in PubMed and other databases using combinations of key words. The summary receiver operator characteristic curves were plotted to assess the overall diagnostic performance of miRNAs. Chi-squared and I2 tests were used to assess the heterogeneity between studies. Additionally, we conducted subgroup and sensitivity analyses to analyze the potential sources of heterogeneity. In total, 33 studies from 12 articles were available in this meta-analysis. The pooled sensitivity, specificity, positive and negative likelihood ratio (PLR, NLR) diagnostic odds ratio, and area under the curve were 0.80, 0.80, 4.0, 0.25, 16, and 0.87, respectively. Subgroup analyses based on the sample types (saliva-, serum- and plasma-based) showed no differences in the diagnostic accuracy of each subgroup. An independent meta-analysis of eight articles was conducted to evaluate the diagnostic accuracy of miRNAs in patients with esophageal squamous cell carcinoma, with a pooled sensitivity of 0.77, specificity of 0.83, PLR of 4.4, NLR of 0.27, diagnostic odds ratio of 16, and area under the curve of 0.87. In conclusion, this meta-analysis demonstrates the feasibility of using miRNAs as non-invasive biomarkers to discriminate esophageal cancer from healthy controls. However, further high-quality studies on more clearly defined esophageal cancer patient are needed to confirm our conclusion. KEY WORDS: diagnosis, esophageal cancer, meta-analysis, microRNA.
INTRODUCTION Esophageal cancer represents the eighth most common lethal malignancies and the sixth leading cause of death from cancer worldwide, affecting three to four times more men than women.1 It shows a significant difference in geographic distribution, with high morbidity appearing in northern China, southeastern Africa, and Japan, and relatively low morbidity in western Africa and Central America.2 Based on the pathological characteristics, esophageal cancer is mainly composed of esophageal squamous cell carcinoma (ESCC) and adenocarcinoma.3 As the dominant subtype of esophageal cancer, ESCC accounts for ∼90% of esophageal cancer in Asian countries and Address correspondence to: Dr Renquan Zhang, M.M., Department of Thoracic Surgery, The First Affiliated Hospital of Anhui Medical University, No. 218 Jixi Road, Hefei 230022, China. Email: [email protected] Conflicts of interest: None. © 2015 International Society for Diseases of the Esophagus
is the second most common cancer in China.1,4 In spite of the improvements in surgical techniques and perioperative management, the overall survival rate is still poor, ranging from 3% to 5% in advanced stage or metastasis.5 However, the 5-year survival rate could increase to 90% if the tumors could be detected and treated at an early stage.6 Thus, the early detection of esophageal cancer may greatly improve clinical outcomes, and novel biomarkers that can facilitate disease detecting and staging are urgently needed to improve survival rates. At present, endoscopic biopsy and histopathological examinations are the golden standard methods for detecting esophageal cancer.7,8 However, there are many inherent deficiencies, including the use of invasive tool and the lack of a sufficient sensitivity and specificity.9 Besides, due to their tiring approaches, it is hard to be applied in high throughput of large-scale studies. Several studies have reported the biological behavior in esophageal 1
2
Diseases of the Esophagus
Cochrane Library, Chinese National Knowledge Infrastructure, and other sources to seek out the articles that assessed the diagnostic value of miRNAs in esophageal cancer up to August 6, 2014. The key words used for literature retrieval are as follows: ‘microRNAs, miRNA, miR’ and ‘esophageal cancer, esophageal neoplasm, esophageal carcinoma, ESCC, esophageal adenocarcinoma’ and ‘sensitivity, specificity, ROC curve, diagnosis’. No language restrictions were imposed on the search criteria.
cancer, and the certain molecules, including p53, cyclin D1, and FAS, are involved in tumorigenesis and the progress of ESCC.10–12 Studies by Kosugi et al. and Mroczko et al. have reported the use of carcinoembryonic antigen and squamous cell carcinoma antigen for early diagnosis of ESCC.13,14 However, these molecules showed poor diagnostic sensitivity, resulting in the diagnosis still at an advanced stage. Hence, the significance of developing novel and less invasive biomarkers for esophageal cancer diagnosis should be emphasized. MicroRNAs (miRNAs), a class of small (18–24 nucleotide) non-coding RNAs, have attracted novel attention for their involvement in cell differentiation, cycling, and apoptosis, through post-transcriptionally regulating gene expression by targeting specific miRNAs.15,16 Previous studies have demonstrated that the aberrant expression of miRNAs contributes to tumorigenesis and cancer development.17 MiRNA expression profiles are cell- or tissue- specific, which may help diagnose different cancer type and predict the progress of malignancies.18,19 Furthermore, tumorderived miRNAs are protected from the endogenous ribonuclease activity by joining with the secretory particles, such as apoptotic bodies and exosomes.20 Therefore, miRNAs existed in a stable form in the blood or other tissue fluid. Taken together, these findings open up an interesting field in the detecting and monitoring of esophageal cancer using miRNAs. A growing number of studies have reported the differential expression of miRNAs in esophageal cancer. For instance, a study by Feber et al. found that miR-194 and miR-192 were upregulated in esophageal adenocarcinoma compared with normal squamous epithelium, while miR-342 was upregulated in ESCC.21 Kurashige et al. found that miR-223 expression was markedly elevated in cancerous ESCC tissues.22 However, the published data on miRNAs in esophageal cancer have reached an inconsistent conclusion. Komatsu et al. reported a poor diagnostic accuracy with 60% sensitivity for miR-21, whereas Ye et al. reported a significantly higher sensitivity (97%).23 Interestingly, Wu et al. conducted a study of seven single miRNAs and its panel (miR-25, -100, -193-3p, -194, -223, -337-5p, -483-5p), and they found that both the single miRNA and the panel miRNAs yielded moderate diagnostic accuracy, with an area under the curve (AUC) ranging from 0.739 to 0.851.24 Thus, we performed this meta-analysis of all the available data to get an overall diagnostic characteristic of miRNAs as novel biomarkers in esophageal cancer.
This meta-analysis was performed based on the standard methods recommended for diagnostic accuracy. Stata software (version 12.0, College Station, TX, USA) was applied in the whole statistical analyses of this study. We used the bivariate meta-analysis model to calculate the summary parameters. The sensitivity and specificity of included studies were pooled to plot the summary receiver operator characteristic (sROC) curve and to calculate the AUC. Additionally, the heterogeneity among studies was assessed using the chi-squared and I2 tests. A value of P < 0.1 or I2 > 50% indicated significant heterogeneity. Subgroup and sensitivity analyses were also conducted to evaluate the potential sources of between-study heterogeneity. Finally, the publication bias was evaluated using Deeks et al. funnel plots. A two-sided P < 0.05 was considered statistically significant.
METHODS
RESULTS
Literature search
Search results and characteristics of studies
A comprehensive search strategy was carried out in numerous databases, including Medline, Embase, the
The systematic databases searches and manual review totally identified 346 records for initial search, and
Study selection and information extraction All relevant publications were included based on the following criteria: patients with esophageal cancer were definitely diagnosed based on endoscopic biopsy and histopathological examinations; studies must be about the use of miRNAs for esophageal cancer diagnosis; sufficient data were available for the generation of 2 × 2 tables, by which we can calculate the value of sensitivity and specificity. Additionally, the exclusion criteria are the following: studies with fewer than 20 patients; studies with duplicate subjects; conference abstracts, letters, editorials, or reviews were excluded because of the limited data. Information was retrieved from the included studies independently by two authors. More details are shown in Table 1. If studies contained both a training group and a validation group, each group was regarded as an independent study in this meta-analysis. Statistical analysis
© 2015 International Society for Diseases of the Esophagus
MicroRNAs in esophageal cancer
3
Table 1 Characteristics of included studies in this meta-analysis Included studies
Year
Location, race
Hirajima et al. Komatsu et al. Liu et al. Takeshita et al. Wang et al. Wu et al.
2013 2011 2012 2013 2012 2014
Japan, Asian Japan, Asian China, Asian Japan, Asian China, Asian China, Asian
Wu et al. Xie et al. Ye et al.
2013 2013 2014
China, Asian China, Asian China, Asian
Zhang et al.
2010
China, Asian
Zhang et al.
2011
China, Asian
Zhang et al.
2013
China, Asian
Case/control
Patient spectrum
Stage
Expression of microRNA
Sample
106/54 50/20 60/60 101/46 31/39 63/63
ESCC ESCC ESCC ESCC Esophageal cancer ESCC
I–IV I–IV I–IV I–IV I–IV I–IV
Plasma Plasma Plasma Serum Serum Serum
67/50 39/19 100/50 100/50 149/100
Esophageal cancer Esophageal cancer Esophageal cancer Esophageal cancer ESCC
I–IV I–IV I–II I–II I–IV
120/121 81/81 120/120 81/81
ESCC ESCC ESCC ESCC
I–IV I–IV I–IV I–IV
Up: miR-18a Up: miR-21; down: miR-375 Down: miR-155 Up: miR-1246 Up: miR-21 Up: miR-25, -100, -193-3p, -194, -223, -337-5p, -483-5p Up: miR-144 Up: miR-10b, -144, -451 Up: miR-21 Up: miR-21 Up: miR-10a, -22, -100, -148b, -223, -133a, -127-3p Up: miR-31 Up: miR-31 Up: miR-1322 Up: miR-1322
Saliva Saliva Plasma Saliva Serum Serum Serum Serum Serum
ESCC, esophageal squamous cell carcinoma.
after removing duplicate records (n = 138), 278 unique abstracts remained. After checking the titles and abstracts, 36 articles were remained for further full-text review. Of these, 12 articles were considered appropriate for this meta-analysis according to the inclusion criteria. Table 1 shows the general characteristics of these included publications. In total, 33 studies from 12 articles covering 2930 patients and 2173 healthy controls were available in this meta-analysis.9,23–33 Of all included studies, 26 studies from eight publications focused on miRNAs as diagnostic biomarkers for ESCC. All these studies reported a total of 22 types of miRNAs. The miRNA expression profiling of the included studies is also summarized in Table 1, and only miR-375 and miR-155 were downregulated in ESCC patients compared with control individuals. Thirty studies discussed a single miRNA, and the remaining three studies focused on a panel miRNAs. The specimens were composed of saliva (5), serum (22), and plasma (6). All the included studies were conducted in Asian populations, of which five studies were performed in Japan, while 28 studies were conducted in China, which may cause a discrepancy of country.
95% confidence intervals (CIs) are as follows: sensitivity, 0.80 (95% CI: 0.76–0.83); specificity, 0.80 (95% CI: 0.76–0.84); PLR, 4.0 (95% CI: 3.3–4.8); NLR, 0.25 (95% CI: 0.21–0.30); and diagnostic odds ratio (DOR), 16 (95% CI: 12–21). Additionally, the sROC curve was plotted in Figure 2a, and the AUC was estimated to be 0.87 (95% CI: 0.84–0.90). Subgroup analyses based on the sample types (saliva-, serum-, and plasma-based) were performed to explore the potential sources of heterogeneity between studies. The sROC curves for the three subgroups are plotted in Figure 2b–d. The pooled diagnostic parameters for each subgroup are listed in Table 2. For saliva specimens, the pooled sensitivity was 0.87, specificity was 0.67, PLR was 2.7, NLR was 0.19, DOR was 14, and the AUC was 0.88; for serum specimens, the pooled sensitivity was 0.77, specificity was 0.83, PLR was 4.5, NLR was 0.27, DOR was 16, and the AUC was 0.87; and for plasma specimens, the pooled sensitivity was 0.82, specificity was 0.79, PLR was 3.9, NLR was 0.22, DOR was 18, and the AUC was 0.88. Noticeable heterogeneity was also observed among studies of each subgroup. Therefore, specimens may not be possible sources of heterogeneity in this meta-analysis.
MiRNAs in diagnosis of esophageal cancer
MiRNAs in diagnosis of ESCC
Figure 1 shows the forest plot of 33 studies with its corresponding sensitivity and specificity for miRNAs assays in the diagnosis of esophageal cancer. Chisquared and I2 tests values of sensitivity and specificity were 150.03 (P < 0.001), and 78.67, 127.59 (P < 0.001), and 74.92, respectively, suggesting significant heterogeneity between studies. Thus, we used the random effect model to calculate the pool estimates. Pooled results for diagnostic accuracy are summarized in Table 2. The results with their corresponding
It has been widely known that ESCC accounts for ∼90% of esophageal cancer in Asian countries. In addition, of the 12 articles included in the metaanalysis of miRNAs associated with esophageal cancer, there are eight articles covering 26 studies that discussed about the relationships between miRNAs and ESCC. Therefore, in this study, we conducted an independent meta-analysis of diagnostic accuracy of miRNAs in discriminating ESCC from healthy controls. The sROC curve is graphed in Figure 3, with an
© 2015 International Society for Diseases of the Esophagus
4
Diseases of the Esophagus
(a)
(b)
Included study
Sensitivity (95% CI)
Included study
Specificity (95% CI)
Zhang et al./2013 Zhang et al./2013 Zhang et al./2011 Zhang et al./2011 Zhang et al./2010 Zhang et al./2010 Zhang et al./2010 Zhang et al./2010 Zhang et al./2010 Zhang et al./2010 Zhang et al./2010 Zhang et al./2010 Ye et al./2014 Ye et al./2014 Xie et al./2013 Xie et al./2013 Xie et al./2013 Wu et al./2013 Wu et al./2014 Wu et al./2014 Wu et al./2014 Wu et al./2014 Wu et al./2014 Wu et al./2014 Wu et al./2014 Wu et al./2014 Wang et al./2012 Takeshita et al./2013 Liu et al./2012 Komatsu et al./2011 Komatsu et al./2011 Komatsu et al./2011 Hirajima et al./2013
0.84 [0.74 − 0.91] 0.82 [0.74 − 0.88] 0.86 [0.77 − 0.93] 0.87 [0.79 − 0.92] 0.79 [0.71 − 0.85] 0.65 [0.57 − 0.73] 0.83 [0.76 − 0.89] 0.66 [0.58 − 0.74] 0.64 [0.55 − 0.71] 0.89 [0.82 − 0.93] 0.81 [0.74 − 0.87] 0.79 [0.71 − 0.85] 0.89 [0.81 − 0.94] 0.97 [0.91 − 0.99] 0.85 [0.69 − 0.94] 0.92 [0.79 − 0.98] 0.90 [0.76 − 0.97] 0.75 [0.63 − 0.84] 0.81 [0.69 − 0.90] 0.75 [0.62 − 0.85] 0.78 [0.66 − 0.87] 0.65 [0.52 − 0.77] 0.65 [0.52 − 0.77] 0.76 [0.64 − 0.86] 0.78 [0.66 − 0.87] 0.83 [0.71 − 0.91] 0.71 [0.52 − 0.86] 0.71 [0.61 − 0.80] 0.62 [0.48 − 0.74] 0.76 [0.62 − 0.87] 0.60 [0.45 − 0.74] 0.88 [0.76 − 0.95] 0.87 [0.79 − 0.93]
Zhang et al./2013 Zhang et al./2013 Zhang et al./2011 Zhang et al./2011 Zhang et al./2010 Zhang et al./2010 Zhang et al./2010 Zhang et al./2010 Zhang et al./2010 Zhang et al./2010 Zhang et al./2010 Zhang et al./2010 Ye et al./2014 Ye et al./2014 Xie et al./2013 Xie et al./2013 Xie et al./2013 Wu et al./2013 Wu et al./2014 Wu et al./2014 Wu et al./2014 Wu et al./2014 Wu et al./2014 Wu et al./2014 Wu et al./2014 Wu et al./2014 Wang et al./2012 Takeshita et al./2013 Liu et al./2012 Komatsu et al./2011 Komatsu et al./2011 Komatsu et al./2011 Hirajima et al./2013
0.80 [0.70 − 0.88] 0.82 [0.75 − 0.89] 0.79 [0.69 − 0.87] 0.84 [0.77 − 0.90] 0.87 [0.79 − 0.93] 0.83 [0.74 − 0.90] 0.83 [0.74 − 0.90] 0.87 [0.79 − 0.93] 0.81 [0.72 − 0.88] 0.86 [0.78 − 0.92] 0.80 [0.71 − 0.87] 0.96 [0.90 − 0.99] 0.64 [0.49 − 0.77] 0.56 [0.41 − 0.70] 0.58 [0.33 − 0.80] 0.47 [0.24 − 0.71] 0.58 [0.33 − 0.80] 0.92 [0.81 − 0.98] 0.81 [0.69 − 0.90] 0.75 [0.62 − 0.85] 0.86 [0.75 − 0.93] 0.79 [0.67 − 0.89] 0.90 [0.80 − 0.96] 0.86 [0.75 − 0.93] 0.75 [0.62 − 0.85] 0.75 [0.62 − 0.85] 0.69 [0.52 − 0.83] 0.74 [0.59 − 0.86] 0.65 [0.52 − 0.77] 0.85 [0.62 − 0.97] 0.60 [0.36 − 0.81] 0.70 [0.46 − 0.88] 1.00 [0.93 − 1.00]
COMBINED
0.80 [0.76 − 0.83] Q = 150.03, df = 32.00, P < 0.01 2 I = 78.67 [71.79 − 85.55]
COMBINED
0.80 [0.76 − 0.84] Q = 127.59, df = 32.00, P < 0.01 2 I = 74.92 [66.46 − 83.38]
0.5
1.0
0.2
1.0
Fig. 1 The forest plots of microRNA (a) sensitivity and (b) specificity in esophageal cancer diagnosis with the corresponding heterogeneity. The sensitivity and specificity from each study are represented by square, and the confidence interval (CI) is indicated by error bars: the pooled sensitivity, 0.80 (95% CI: 0.76–0.83); the pooled specificity, 0.80 (95% CI: 0.76–0.84).
AUC of 0.87 (95% CI: 0.84–0.90). The other pooled diagnostic parameters are listed in Table 2. The results are as follows: sensitivity, 0.77 (95% CI: 0.74– 0.81); specificity, 0.83 (95% CI: 0.79–0.85); PLR, 4.4 (95% CI: 3.7–5.4); NLR, 0.27 (95% CI: 0.23–0.32); and DOR, 16 (95% CI: 12–22). The diagnostic values of miRNAs in ESCC diagnosis were similar to that in esophageal cancer diagnosis, which suggested that miRNAs are promising biomarkers for ESCC detection, no matter whether the subtypes of esophageal cancer were identified in Asian people.
Sensitivity and outlier detection analyses We conducted the goodness of fit and bivariate normality analyses to assess the reliability of using the random effects bivariate model for the calculation of the pooled estimates (Fig. 4). The results suggested that the selected model was robust for the calculation of the pooled estimates. Additionally, in order to affirm that our findings were not driven by any individual study, we further performed the outlier detection analyses. After excluding the four
Table 2 Comparisons of assays based on sample for diagnosing esophageal cancer
Analyses
Sensitivity (95% CI)
Specificity (95% CI)
Positive LR (95% CI)
Negative LR (95% CI)
DOR (95% CI)
AUC (95% CI)
Combined Saliva Serum Plasma ESCC Excluding outliers
0.80 (0.76–0.83) 0.87 (0.80–0.92) 0.77 (0.74–0.81) 0.82 (0.67–0.91) 0.77 (0.74–0.81) 0.78 (0.74–0.81)
0.80 (0.76–0.84) 0.67 (0.49–0.82) 0.83 (0.80–0.85) 0.79 (0.55–0.92) 0.83 (0.79–0.85) 0.80 (0.77–0.82)
4.0 (3.3–4.8) 2.7 (1.6–4.3) 4.5 (3.9–5.1) 3.9 (1.6–10.0) 4.4 (3.7–5.4) 3.9 (3.4–4.4)
0.25 (0.21–0.30) 0.19 (0.14–0.27) 0.27 (0.23–0.32) 0.22 (0.11–0.46) 0.27 (0.23–0.32) 0.28 (0.24–0.32)
16 (12–21) 14 (8–25) 16 (13–21) 18 (4–75) 16 (12–22) 14 (11–18)
0.87 (0.84–0.90) 0.88 (0.85–0.90) 0.87 (0.84–0.90) 0.88 (0.85–0.90) 0.87 (0.84–0.90) 0.86 (0.82–0.89)
AUC, area under the curve; CI, confidence interval; DOR, diagnostic odds ratio; ESCC, esophageal squamous cell carcinoma; LR, likelihood ratio. © 2015 International Society for Diseases of the Esophagus
MicroRNAs in esophageal cancer
(a)
(b)
1.0
1.0
20
3
18 24 30
31 27 33 3215 23
2913 104
22 16
2
Sensitivity
26
2
5
19
4
8 9 1
14 6
11
17
21
7
28 12 25
Sensitivity
1
5 3
0.5
0.0 1.0
0.5 Specificity
0.5
0.0 1.0
0.0
(c)
0.5 Specificity
0.0
(d)
1.0
1.0 13 19
188 5
4
15
4
9 1
6
17
14
2
7
0.5
0.0 1.0
2
3
Sensitivity
11
6
1
20
16 22 2110 12
Sensitivity
5
0.5 Specificity
0.0
5 3
0.5
0.0 1.0
0.5 Specificity
0.0
Fig. 2 The summary receiver operator characteristic (sROC) curves based on subgroup analyses. (a) overall studies [(○) observed data; (◆) summary operating point, SENS = 0.80 (0.76–0.83), SPEC = 0.80 (0.76–0.84); (—) sROC curve, AUC = 0.87 (0.84–0.90); (---) 95% confidence contour; (. . .) 95% prediction contour)]; (b) saliva-based miRNA assay [(○) observed data; (◆) summary operating point, SENS = 0.87 (0.80–0.92), SPEC = 0.67 (0.49–0.82); (---) sROC curve, AUC = 0.88 (0.85–0.90); (—) 95% confidence contour; (. . .) 95% prediction contour)]; (c) serum-based miRNA assay [(○) observed data; (◆) summary operating point, SENS = 0.77 (0.74–0.81), SPEC = 0.83 (0.80–0.85); (—) sROC curve, AUC = 0.87 (0.84–0.90); (---) 95% confidence contour; (. . .) 95% prediction contour)]; (d) plasma-based miRNA assay [(○) observed data; (◆) summary operating point, SENS = 0.82 (0.67–0.91), SPEC = 0.79 (0.55–0.92); (—) sROC curve, AUC = 0.88 (0.85–0.90); (---) 95% confidence contour; (. . .) 95% prediction contour)].
outlier studies, the I2 value for heterogeneity decreased, from 78.67% to 75.29% for sensitivity, and from 74.92% to 57.06% for specificity. Compared with the overall analysis, after excluding the four outlier studies, the pooled diagnostic estimates were only minimally changed. The data before and after excluding four outliers are listed in Table 2, with sensitivity of 0.80 versus 0.78, specificity of © 2015 International Society for Diseases of the Esophagus
0.80 versus 0.80, PLR of 4.0 versus 3.9, NLR of 0.25 versus 0.28, DOR of 16 versus 14, and AUC of 0.87 versus 0.86, respectively. Publication bias Considering that publication bias is a key factor that may influence the results of meta-analysis, we
6
Diseases of the Esophagus
ESCC studies, which suggested symmetry in the included articles and a relatively low likelihood of publication bias.
1.0 17 23
1
15
2
24
20 26 2514 16 2212 94
7 8 13 6
19
21
11 18
DISCUSSION
5 3
0.5
0.0 1.0
0.5
0.0
Specificity
Fig. 3 The summary receiver operator characteristic (sROC) curve of the diagnostic studies of miRNAs in discriminating esophageal squamous cell carcinoma from healthy subjects (○) observed data; (◆) summary operating point, SENS = 0.77 (0.74–0.81), SPEC = 0.83 (0.79–0.85); (—) sROC curve, AUC = 0.87 (0.84–0.90); (—) 95% confidence contour; (. . .) 95% prediction contour.
(a)
(b)
1.00
1.00
Mahalanobis D−square
Deviance residual
performed the Deeks et al.’s funnel plot with asymmetry test to explore publication bias in this study. The slope coefficient reflected a P-value of 0.11 in esophageal cancer studies and a P-value of 0.21 in
0.75 0.50 0.25 0.00 0.00
0.25 0.50 0.75 Normal quantile
3.00 20 1
2.00 22
1.00 3
5
18
0.00 0
10
20 Study
30
0.75 0.50 0.25 0.00
1.00
(c)
Cook’s distance
A non-invasive assay using miRNAs opens up a novel and interesting field in the screening and diagnosing of cancer patients. Experimental evidence has demonstrated that an abundance of human miRNAs genes is located at fragile sites and genomic regions, which may function as oncogenes or tumor suppressor genes.34 Tumor-specific alterations of miRNAs in the blood or other body fluid have been identified in patients with several malignancies, such as prostate cancer,35 oral cancer,36 lung cancer,37 breast cancer,38 pancreatic cancer,39 colorectal cancer,40 ovarian cancer,41 and esophageal cancer.42 In addition, studies have demonstrated that miRNAs are presented in plasma/serum in a remarkably stable form, and the expression level of miRNAs is reproducible and consistent among individuals.43 Thus, miRNAs may provide new therapeutic strategies such as biomarkers for cancers detection. Esophageal cancer is one of the most common types of malignant tumors worldwide.1 Despite the improvements in surgical techniques, esophageal
40
0.00
Standardized residual(diseased)
Sensitivity
10
0.25 0.50 0.75 Chi-squared quantile
1.00
(d)
3.0
20
2.0
18 17 21
1.0
24
2 19 8
0.0 7
−1.0 5 3
9 14 6
30 31 33 27 2315 32 25 12 19 26 17 28 18 3 5 20 23 15 32 33 13 27 31 10 29 14 22 30 16 4 8 9 1 6 2 7 21 24 11
1
13 29 4 10
12 25 28
26
22 16 11
−2.0 −3.0 −3.0
−2.0
−1.0 0.0 1.0 2.0 Standardized residual
3.0
Fig. 4 Graphical depiction of residual-based (a) goodness of fit, (b) bivariate normality, (c) influence, and (d) outlier detection analyses, respectively. © 2015 International Society for Diseases of the Esophagus
MicroRNAs in esophageal cancer
cancer is regarded as one of the most lethal carcinomas of the gastrointestinal tract, with a dismal overall 5-year survival rate. At present, radiology and endoscopic biopsy are the most common diagnostic methods.8 However, these tests are invasive and time-consuming, and they may cause discomfort to patients. Patients with esophageal cancer are usually diagnosed at an advanced stage or at a stage with the generation of lymph nodes, which remains the main cause of poor survival rate.3 Therefore, developing a biomarker of early-stage esophageal cancer is essential and continuous. In our meta-analysis, after a comprehensive literature research, we finally included 12 articles covering 33 studies to calculate the pooled diagnostic estimates. The overall analysis yield relatively high diagnostic performances of miRNAs with an AUC of 0.87, a sensitivity of 0.80, and a specificity of 0.80. A DOR of 1.0 shows no ability to discriminate between patients with the disease and those without it. The mean DOR in this meta-analysis was 16, suggesting that miRNAs are powerful biomarkers for esophageal cancer patient diagnosis. PLR and NLR are identified as more clinically meaningful estimates of diagnostic accuracy in clinical practice. A value of PLR greater than 10 and NLR lower than 0.1 are considered the threshold for reliability.44 In this metaanalysis, the pooled PLR and NLR were 4.0 and 0.25, respectively. Although the PLR and NLR are less encouraging, miRNAs are still relatively accurate in diagnosing esophageal cancer in terms of the pooled values of AUC and DOR. As the dominant subtype of esophageal cancer, ESCC is the second most common cancer and the fourth leading cause of cancer-related mortality. Accumulating studies have showed that aberrant expression of miRNAs may be involved in the initiation and progression of ESCC. Therefore, an independent meta-analysis of 26 included studies was conducted to evaluate the diagnostic efficacy of miRNAs in discriminating ESCC from healthy controls. MiRNAs achieved a pooled sensitivity of 0.77, specificity of 0.83, PLR of 4.4, NLR of 0.27, DOR of 16, and AUC of 0.87. In contrast to miRNAs in esophageal cancer diagnosis, the diagnostic values of miRNAs in ESCC diagnosis showed no significant differences. The few esophageal adenocarcinoma or other types of esophageal cancer cases in the included studies may contribute to the results. The roles of miRNAs in carcinogenesis, progression, invasion, and metastasis of esophageal cancer have been widely studied, and the mechanisms are gradually excluded. MiRNAs act as onco-miRNA or tumor suppressors by targeting to the tumor suppressor genes or ontogenesis.45 A study by Ma et al. has proven that miR-21 promotes cell proliferation by targeting chromosome-10 (PTEN) at posttranscriptional level, and inhibits cell growth and © 2015 International Society for Diseases of the Esophagus
7
invasion by targeting FASL, TIMP3, and RECK genes.46,47 MiR-10a was reported to be downregulated in esophageal cancer and to affect cell migration and invasion through targets in homeobox genes, whereas miR-10b was overexpressed and affected the cell motility and invasiveness by suppressing the tumor suppressor gene KLF4.48,49 Zhang et al. demonstrated that upregulation of miR-31 promoted the tumor formation through the EMP1, KSR2, and RGS4 genes.32 As a growing number of researchers studied the relationships between miRNAs and esophageal cancer, the exact role of miRNA in oncogenesis of esophageal cancer could be clearly elucidated in the future. It has been reported that the specific miRNA profiles were closely related to prognosis and histopathological features of malignancies, such as stages and types. Iorio et al. found that miR-145 and miR-21 were differentially expressed in breast cancer with different tumor stage, suggesting that these two miRNAs were involved in tumor progression and could be acquired in the identification of tumor progression.50 Since esophageal cancers diagnosed at stage I or II have a good chance of improving survival rate, we expect to identify specific miRNAs unique to esophageal cancer in different stages. However, only one article investigated the early stage (I–II) of esophageal cancer patients in this meta-analysis. Therefore, in order to clarify whether miRNA expressed differentially in esophageal cancers with different tumor stage, further original studies have been performed. When getting a reasonable conclusion of any metaanalyses, it is indispensable to consider the interstudy heterogeneity. In our meta-analysis, subgroup analyses based on the sample types (saliva-, serum-, and plasma-based) were performed to explore the potential sources of heterogeneity between studies. The pooled results showed no obvious differences in each subgroup, and broad heterogeneities also existed in the three subgroups, which suggest that specimens may not be possible sources of heterogeneity in this meta-analysis. In view of our sensitivity analysis, after excluding the four outlier studies, the pooled results did not change much, indicating that our findings were not driven by any single study. A previous meta-analysis by Fu et al. shows that miR-21 and miR-375 can be used as prognostic biomarkers in esophageal cancer.51 However, no meta-analyses on the overall accuracy of miRNAs in diagnosing esophageal cancer are available at present. Thus, we conduct this meta-analysis to assess the diagnostic performances of miRNAs in esophageal cancer diagnosis. Furthermore, we also identified the diagnostic accuracy of miRNAs in ESCC diagnosis by performing an independent meta-analysis of 26 studies. However, there are still several shortcomings in the research. First, of the 12 articles included, only
8
Diseases of the Esophagus
three publications were conducted in Japan, and most of the studies were in performed in China, which may lead a country selection bias. Second, patients in some studies were not clearly classified, and thus we did not know the certain subtypes of the esophageal cancer patients. Last, because of a limited number of articles, subgroups analysis on ethnicity, miRNAs profiling and stages could not be conducted to explore the potential sources of heterogeneity between studies. In conclusion, this meta-analysis demonstrates the feasibility of using miRNAs as non-invasive biomarkers to discriminate esophageal cancer from healthy controls. However, further high-quality studies are needed to confirm our conclusion, and in order to efficiently apply these findings in clinic, studies on more clearly defined esophageal cancer patient are warranted.
14
15 16 17 18 19 20 21
Acknowledgments
22
This work was supported by a grant from the Natural Science Foundation of China (No. 81302028): The biological significance of HIF-1a expression in transition zone of small cell lung cancer after radiofrequency ablation.
23
References
25
1 Enzinger P C, Mayer R J. Esophageal cancer. NEJM 2003; 349: 2241–52. 2 Jemal A, Bray F, Center M M, Ferlay J, Ward E, Forman D. Global cancer statistics. CA Cancer J Clin 2011; 61: 69–90. 3 He B, Yin B, Wang B, Xia Z, Chen C, Tang J. MicroRNAs in esophageal cancer (review). Mol Med Rep 2012; 6: 459–65. 4 Hiyama T, Yoshihara M, Tanaka S, Chayama K. Genetic polymorphisms and esophageal cancer risk. Int J Cancer 2007; 121: 1643–58. 5 Kim T, Grobmyer S R, Smith R et al. Esophageal cancer – the five year survivors. J Surg Oncol 2011; 103: 179–83. 6 Daly J M, Fry W A, Little A G et al. Esophageal cancer: results of an American College of Surgeons Patient Care Evaluation Study. J Am Coll Surg 2000; 190: 562–72, discussion 72–3. 7 Barber T W, Duong C P, Leong T, Bressel M, Drummond E G, Hicks R J. 18F-FDG PET/CT has a high impact on patient management and provides powerful prognostic stratification in the primary staging of esophageal cancer: a prospective study with mature survival data. J Nucl Med 2012; 53: 864–71. 8 Tachimori Y, Kanamori N, Uemura N, Hokamura N, Igaki H, Kato H. Salvage esophagectomy after high-dose chemoradiotherapy for esophageal squamous cell carcinoma. J Thorac Cardiovasc Surg 2009; 137: 49–54. 9 Hirajima S, Komatsu S, Ichikawa D et al. Clinical impact of circulating miR-18a in plasma of patients with oesophageal squamous cell carcinoma. Br J Cancer 2013; 108: 1822–9. 10 Hollstein M C, Metcalf R A, Welsh J A, Montesano R, Harris C C. Frequent mutation of the p53 gene in human esophageal cancer. Proc Natl Acad Sci U S A 1990; 87: 9958–61. 11 Adelaide J, Monges G, Derderian C, Seitz J F, Birnbaum D. Oesophageal cancer and amplification of the human cyclin D gene CCND1/PRAD1. Br J Cancer 1995; 71: 64–8. 12 Gratas C, Tohma Y, Barnas C, Taniere P, Hainaut P, Ohgaki H. Up-regulation of Fas (APO-1/CD95) ligand and downregulation of Fas expression in human esophageal cancer. Cancer Res 1998; 58: 2057–62. 13 Kosugi S, Nishimaki T, Kanda T, Nakagawa S, Ohashi M, Hatakeyama K. Clinical significance of serum carcinoem-
24
26
27 28
29
30 31
32
33 34
35
bryonic antigen, carbohydrate antigen 19-9, and squamous cell carcinoma antigen levels in esophageal cancer patients. World J Surg 2004; 28: 680–5. Mroczko B, Kozlowski M, Groblewska M, et al. The diagnostic value of the measurement of matrix metalloproteinase 9 (MMP-9), squamous cell cancer antigen (SCC) and carcinoembryonic antigen (CEA) in the sera of esophageal cancer patients. Clinica chimica acta; International journal of clinical chemistry 2008; 389(1–2): 61–6. Krol J, Loedige I, Filipowicz W. The widespread regulation of microRNA biogenesis, function and decay. Nat Rev Genet 2010; 11: 597–610. Farazi T A, Hoell J I, Morozov P, Tuschl T. MicroRNAs in human cancer. Adv Exp Med Biol 2013; 774: 1–20. Paranjape T, Slack F J, Weidhaas J B. MicroRNAs: tools for cancer diagnostics. Gut 2009; 58: 1546–54. Madhavan D, Cuk K, Burwinkel B, Yang R. Cancer diagnosis and prognosis decoded by blood-based circulating microRNA signatures. Front Genet 2013; 4: doi:10.3389/fgene.2013.00116. Calin G A, Croce C M. MicroRNA signatures in human cancers. Nat Rev Cancer 2006; 6: 857–66. Mitchell P S, Parkin R K, Kroh E M et al. Circulating microRNAs as stable blood-based markers for cancer detection. Proc Natl Acad Sci U S A 2008; 105: 10513–18. Feber A, Xi L, Luketich J D et al. MicroRNA expression profiles of esophageal cancer. J Thorac Cardiovasc Surg 2008; 135: 255–60, discussion 60. Kurashige J, Watanabe M, Iwatsuki M et al. Overexpression of microRNA-223 regulates the ubiquitin ligase FBXW7 in oesophageal squamous cell carcinoma. Br J Cancer 2012; 106: 182–8. Komatsu S, Ichikawa D, Takeshita H et al. Circulating microRNAs in plasma of patients with oesophageal squamous cell carcinoma. Br J Cancer 2011; 105: 104–11. Wu C, Wang C, Guan X et al. Diagnostic and prognostic implications of a serum miRNA panel in oesophageal squamous cell carcinoma. PLoS ONE 2014; 9: e92292. Ye M, Ye P, Zhang W, Rao J, Xie Z. Diagnostic values of salivary versus and plasma microRNA-21 for early esophageal cancer. Nan Fang Yi Ke Da Xue Xue Bao 2014; 34: 885– 9. Wu W, Hou W, Wu Z, Wang Y, Yi Y, Lin W. miRNA-144 in the saliva is a genetic marker for early diagnosis of esophageal cancer. Nan Fang Yi Ke Da Xue Xue Bao 2013; 33: 1783– 6. Xie Z, Chen G, Zhang X et al. Salivary microRNAs as promising biomarkers for detection of esophageal cancer. PLoS ONE 2013; 8: e57502. Takeshita N, Hoshino I, Mori M et al. Serum microRNA expression profile: miR-1246 as a novel diagnostic and prognostic biomarker for oesophageal squamous cell carcinoma. Br J Cancer 2013; 108: 644–52. Liu R, Liao J, Yang M et al. Circulating miR-155 expression in plasma: a potential biomarker for early diagnosis of esophageal cancer in humans. J Toxicol Environ Health A 2012; 75: 1154– 62. Wang B, Zhang Q. The expression and clinical significance of circulating microRNA-21 in serum of five solid tumors. J Cancer Res Clin Oncol 2012; 138: 1659–66. Lou X, Qi X, Zhang Y, Long H, Yang J. Decreased expression of microRNA-625 is associated with tumor metastasis and poor prognosis in patients with colorectal cancer. J Surg Oncol 2013; 108: 230–5. Zhang T, Wang Q, Zhao D et al. The oncogenetic role of microRNA-31 as a potential biomarker in oesophageal squamous cell carcinoma. Clin Sci (Lond) 2011; 121: 437– 47. Zhang C, Wang C, Chen X et al. Expression profile of microRNAs in serum: a fingerprint for esophageal squamous cell carcinoma. Clin Chem 2010; 56: 1871–9. Calin G A, Sevignani C, Dumitru C D et al. Human microRNA genes are frequently located at fragile sites and genomic regions involved in cancers. Proc Natl Acad Sci U S A 2004; 101: 2999–3004. Bryant R J, Pawlowski T, Catto J W et al. Changes in circulating microRNA levels associated with prostate cancer. Br J Cancer 2012; 106: 768–74. © 2015 International Society for Diseases of the Esophagus
MicroRNAs in esophageal cancer 36 Ries J, Vairaktaris E, Agaimy A et al. miR-186, miR-3651 and miR-494: potential biomarkers for oral squamous cell carcinoma extracted from whole blood. Oncol Rep 2014; 31: 1429– 36. 37 Sozzi G, Boeri M, Rossi M et al. Clinical utility of a plasmabased miRNA signature classifier within computed tomography lung cancer screening: a correlative MILD trial study. J Clin Oncol 2014; 32: 768–73. 38 McDermott A M, Miller N, Wall D et al. Identification and validation of oncologic miRNA biomarkers for luminal A-like breast cancer. PLoS ONE 2014; 9: e87032. 39 Zhang J, Zhao C Y, Zhang S H et al. Upregulation of miR-194 contributes to tumor growth and progression in pancreatic ductal adenocarcinoma. Oncol Rep 2014; 31: 1157–64. 40 Zhang L, Meng L, Fan Z, Liu B, Pei Y, Zhao Z. Expression of plasma miR-106a in colorectal cancer and its clinical significance. Nan Fang Yi Ke Da Xue Xue Bao 2014; 34: 354–7. 41 Zheng H, Zhang L, Zhao Y et al. Plasma miRNAs as diagnostic and prognostic biomarkers for ovarian cancer. PLoS ONE 2013; 8: e77853. 42 Javidi M A, Ahmadi A H, Bakhshinejad B, Nouraee N, Babashah S, Sadeghizadeh M. Cell-free microRNAs as cancer biomarkers: the odyssey of miRNAs through body fluids. Med Oncol 2014; 31: doi:10.1007/s12032-014-0295-y. 43 Kosaka N, Iguchi H, Yoshioka Y, Takeshita F, Matsuki Y, Ochiya T. Secretory mechanisms and intercellular transfer of
© 2015 International Society for Diseases of the Esophagus
44 45 46
47 48 49 50 51
9
microRNAs in living cells. J Biol Chem 2010; 285: 17442– 52. Deeks J J, Altman D G. Diagnostic tests 4: likelihood ratios. BMJ 2004; 329: 168–9. Esquela-Kerscher A, Slack F J. Oncomirs – microRNAs with a role in cancer. Nat Rev Cancer 2006; 6: 259–69. Wang N, Zhang C Q, He J H et al. MiR-21 down-regulation suppresses cell growth, invasion and induces cell apoptosis by targeting FASL, TIMP3, and RECK genes in esophageal carcinoma. Dig Dis Sci 2013; 58: 1863–70. Ma W J, Lv G D, Tuersun A et al. Role of microRNA-21 and effect on PTEN in Kazakh’s esophageal squamous cell carcinoma. Mol Biol Rep 2011; 38: 3253–60. Matsushima K, Isomoto H, Kohno S, Nakao K. MicroRNAs and esophageal squamous cell carcinoma. Digestion 2010; 82: 138–44. Tian Y, Luo A, Cai Y et al. MicroRNA-10b promotes migration and invasion through KLF4 in human esophageal cancer cell lines. J Biol Chem 2010; 285: 7986–94. Iorio M V, Ferracin M, Liu C G et al. MicroRNA gene expression deregulation in human breast cancer. Cancer Res 2005; 65: 7065–70. Fu W, Pang L, Chen Y, Yang L, Zhu J, Wei Y. The microRNAs as prognostic biomarkers for survival in esophageal cancer: a meta-analysis. Scientificworldjournal 2014; 2014: 523979.
