The Prognostic Value of mir-21 and mir-155 in Non-small-cell Lung Cancer: A Meta-analysis

Transcription

1 Jpn J Clin Oncol 2013;43(8) doi: /jjco/hyt084 Advance Access Publication 30 June 2013 The Prognostic Value of mir-21 and mir-155 in Non-small-cell Lung Cancer: A Meta-analysis Yan Wang 1,, Jinyuan Li 1,, Liping Tong 1, Jingwei Zhang 1, Aixia Zhai 1, Kun Xu 2, Lanlan Wei 1,* and Ming Chu 2,* 1 Department of Microbiology, Harbin Medical University, Immunity and Infection, Pathogenic Biology Key Laboratory, Harbin, Heilongjiang Province and 2 Department of Neurosurgery, The First Affiliated Hospital of Harbin Medical University, Harbin, Heilongjiang Province, China *For reprints and all correspondence: Lanlan Wei, Department of Microbiology, Harbin Medical University, Harbin , China. weilanlan1119@163.com; Ming Chu, Department of Neurosurgery, The First Affiliated Hospital of Harbin Medical University, Harbin , China. chuming120@yahoo.com.cn Contributed equally. Received February 28, 2013; accepted May 20, 2013 Objective: mir-21 and mir-155 have been implicated in the prognosis of non-small-cell lung cancer, but the results are controversial. To resolve this issue, we performed a meta-analysis on mir-21 and mir-155 and non-small-cell lung cancer prognosis and lymphoid metastasis. Methods: Eligible data were extracted and the correlation between mir-21 and mir-155 and non-small-cell lung cancer survival was analyzed by calculating a pooled hazard ratio and sensitivity analysis. The heterogeneity was detected by Q statistic and I-squared statistic, and the publication bias was tested by funnel plots and Egger s test. Results: Nineteen studies were included. High mir-21 level (hazard ratio ¼ 2.00, 95% confidence interval ¼ , P ¼ for heterogeneity test, I 2 ¼ 84.9%) and high mir-155 level (hazard ratio ¼ 1.65, 95% confidence interval ¼ , P ¼ for heterogeneity test, I 2 ¼ 68.3%) were significantly associated with worse non-small-cell lung cancer survival. Furthermore, a high mir-21 level was associated with an increased risk of lymphoid infiltration for non-small-cell lung cancer (odds ratio ¼ 1.93; 95% confidence interval ¼ ). Funnel plot and Egger s test suggested that there was no publication bias in the current meta-analysis. Conclusions: This meta-analysis provides evidence that mir-21 and mir-155 are predicting factors for non-small-cell lung cancer prognosis and lymphoid infiltration. Key words: meta-analysis mir-21 mir-155 NSCLC prognosis INTRODUCTION Lung cancer is the leading cause of cancer-related deaths worldwide, and non-small-cell lung cancer (NSCLC) accounts for 80% of all cases. Lung cancer is usually diagnosed at advanced stages when the prognosis is poor, resulting in an overall 5-year survival rate of 0 14% (1). However, the 5-year survival rate in patients with Stage I NSCLC can be as high as 83%. Therefore, the early detection of NSCLC can remarkably reduce the morality (2,3). However, NSCLC at an early stage appears to be overlooked or overdiagnosed, highlighting the need for refined diagnostic and prognostic strategies (4). Recently, micrornas (mirnas) have been shown to play a critical role in the development of NSCLC and proposed as potential biomarkers for NSCLC diagnosis and therapy. mirnas are small non-coding RNAs, nucleotides, which influence a variety of cellular functions, including proliferation, differentiation and apoptosis (5,6). Increasing data suggest that mirnas can behave as either tumor suppressor genes or oncogenes depending on the cellular context (7). Several studies have shown that mirnas could be the potential diagnostic biomarkers for many kinds of cancer, such as # The Author Published by Oxford University Press. All rights reserved. For Permissions, please journals.permissions@oup.com

2 814 mir-21 and mir-155 in NSCLC lymphoma, pancreatic ductal adenocarcinoma and breast cancer (8,9). In particular, the upregulation of mir-21 and mir-155 was reported in breast cancer, B-cell lymphoma, lung cancer, colon cancer and other solid cancer (10 13). Meanwhile, their deregulation could induce apoptosis and inhibit cell proliferation and invasion (14,15). The association between mir-21 or mir-155 expression and NSCLC prognosis was explored in several studies, but the results are controversial (16 24). While some studies showed that mir-21 or mir-155 overexpression was associated with poor NSCLC survival (16 22), other studies indicated the converse or insignificant association (23,24). Meta-analysis is regarded as the best means to summarize the results of a number of randomized controlled trials reporting similar questions (25). In addition, meta-analysis has been increasingly employed to investigate the association of gene polymorphorisms with cancer risk and prognosis (26 28). By extracting summary statistics of the published literature for survival endpoints, we conducted a systematic meta-analysis to analyze the predictive value of mir-21 and mir-155 for NSCLC prognosis and explore the relationship between mir-21 or mir-155 level and lymphoid infiltration of NSCLC. MATERIALS AND METHODS SEARCH STRATEGY The literature published from 1 January 1978 to 10 June 2012 were searched in PubMed, EBSCO and LWW with the key words including lung cancer, non-small-cell lung cancer or NSCLC, and mir-21 or microrna 21, mir-155 or microrna 155. A manual review of the references of relevant publications was also performed to obtain additional studies. STUDY SELECTION Three reviewers primarily examined the titles and abstracts of all literature. The full text articles were then screened separately by two reviewers to determine whether they met the inclusion criteria or not. We also contacted some authors when the crucial data were not reported in their original papers. Articles were independently read and selected according to the inclusion criteria. Disagreements were resolved through the consensus with a third reviewer. Inclusion criteria: The primary literature contained expression profiles of mir-21 or mir-155 together with (i) NSCLC patients survival analysis with OS (overall survival rate); (ii) hazard analysis with HR (hazard ratio), 95% CIs (confidence intervals) and P value, or relevant data that could be used to calculate the HR and 95%CI or (iii) study of NSCLC lymphatic node metastasis or lymphatic infiltration. Furthermore, the studies should give basic information and demographic data for investigated samples. Exclusion criteria: (i) non-english and (ii) studies not involving the association between mir-21 or mir-155 expression level and NSCLC prognosis. QUALITY ASSESSMENT Two reviewers critically assessed the quality of all the studies included according to a basic standard as follows: (i) basic information of the enrolled patients, population and ethnicity was given; (ii) information of the disease status of the samples, disease course and follow-up was detailed; (iii) a clear report of study design was provided; (iv) period of follow-up was sufficient; (v) definition of cut-off of mir-21 and mir-155 was clear and (vi) report of outcome investigation was also included (29). Otherwise, the study was removed in order to enhance the quality of the meta-analysis. DATA EXTRACTION Two investigators independently extracted the required information from all eligible studies. Pre-specified data parameters contained were as follows: (i) demographic data regarding population, ethnicity, the numbers of death and survival during follow-up; (ii) tumor data including the number of primary lesions, stage, histology and lymphoid node invasion; (iii) experimental data involving study design, materials, assays and cut-off definition of mir-21 and mir-155; (iv) statistical data including the HR of elevated mir-21 and mir-155 for OS, as well as their 95% CI and P value; (v) publication data containing first author s last name and publication year. STATISTICAL METHODS The high or low expression of mir-21 and mir-155 was defined according to the cut-off values provided by the authors. For the quantitative aggregation of the survival results, the mir-21 or mir-155 effect was measured by the HR between high and low expression survival distributions. For each trial, the HR was estimated depending on some results provided in a previous publication, and the most accurate method was conducted based on the directly reported HR of the original study. If the HR was not available directly, the total number of deaths, the number of patients at risk in each group and the log-rank statistic or its P value would be used to make an approximate calculation of the HR estimate. When the useful data were only in the form of graphical representations (Kaplan Meier curves) of the survival distributions, the survival rates at specified time points were extracted to reconstruct the HR estimate and its variance (25). Conventionally, an observed HR of.1 implied worse survival for the group with mir-21 or mir-155 overexpression. This impact of high expression of mir-21 or mir-155 on the survival rate was considered statistically significant if the 95% CI for the overall HR did not overlap in the forest plot. The fixed-effects or random-effects models were used to pool the effect size based on Mantel Haenszel and

3 Jpn J Clin Oncol 2013;43(8) 815 DerSimonian-Laird methods, respectively. These two models provide similar results when between-studies heterogeneity is absent; otherwise, a random-effects model is more appropriate (30). Q statistic was performed to test the heterogeneity among the studies included in meta-analysis, and it was considered statistically significant with a P value of,0.05. I-squared statistic was also used to assess the between-studies heterogeneity. I-squared values of 50% indicate large heterogeneity among studies, whereas values of 25 50% indicate moderate heterogeneity (30). Subgroup analysis was further performed to explore the source of existed heterogeneity. The inverted funnel plots and Egger s test (linear regression analysis) were used to evaluate the publication bias. All the P values were two-sided. The meta-analysis was carried out by using software STATA11. RESULTS CHARACTERISTICS OF THE STUDIES INCLUDED IN META-ANALYSIS As shown in Fig. 1, a total of 874 published records were retrieved by using the key words mentioned earlier in the Methods section. Among those publications, 10 publications were screened and others were excluded either because they did not provide available data to extract the HRs and 95% CI or not mention NSCLC prognosis. The remaining 10 publications contained 16 studies (4,16 24). Among them, nine studies are available to examine the association between mir-21 and NSCLC, while seven for mir-155 and NSCLC. The number of patients enrolled in all studies varied from 30 to 637, and the follow-up period ranged from 30 months to 8 years. The principal individual characteristics of these different trials are listed in Table 1. Among them, seven studies recruited participants from Asia (4,16 19) and ten from other ethnic groups (4,20 24). In addition, eight studies involved both adenocarcinoma and square cell carcinoma (16 18, 22 24), six studies involved adenocarcinoma (4,20), and two studies involved only square cell carcinoma (19,21). Although most of the studies investigated patients with stages I-III NSCLC, two studies included only patients with stages I-II (4). QRT PCR was used to detect mirna expression in all trials, except one (20). Cancerous tissues were usually applied to determine mirna expression levels, while serum samples were tested in two studies (16,17). Apparently, the cut-off values of mirna were different in each study, with median appliedin11studiesandmean,5-foldor2-foldusedinother studies. All the studies are retrospective in design. THE ASSOCIATION BETWEEN MIR-21 EXPRESSION AND NSCLC PROGNOSIS For studies evaluating the association between mir-21 expression and NSCLC prognosis, there was heterogeneity among the studies (P, 0.05). Hence, a random-effects model was applied to calculate a pooled HR and its 95% CI. When nine eligible studies were pooled into the meta-analysis, we found that higher mir-21 expression was significantly associated with worse NSCLC survival (HR ¼ 2.00, 95% CI ¼ , P ¼ for heterogeneity test, I 2 ¼ 84.9%) (Fig. 2A). Considering the substantial heterogeneity among the studies, the effect of mir-21 expression was further evaluated using subgroup analysis. The analyzed subgroups were Figure 1. Flow chart of searching the literature used in this meta-analysis.

4 Table 1. Characteristics of the studies included in this meta-analysis Study Country Ethnicity Histology N Stage Materials MicroRNA assay Cut-off Follow-up (month) Gao et al (18) China Asian AC SC 47 I II III Frozen tissues qrt-pcr Median 60 mir-21 Gao et al (19) China Asian SC 30 I II III Frozen tissues qrt-pcr Median 60 mir-21 Markou et al (23) America Non-Asian AC SC 48 I II III Frozen tissues qrt-pcr 2-fold 50 mir-21 IV Liu 2012 (17) China Asian AC SC 70 I II III Serum qrt-pcr 2-fold 30 mir-21 IV Liu 2012 (17) China Asian AC SC 70 I II III Frozen tissues qrt-pcr 2-fold 30 mir-21 IV Saito 2011 (4) Japan Asian AC 191 I II Frozen tissues qrt-pcr Median 80 mir-21 Saito 2011 (4) America/Norway Non-Asian AC 126 I II III Frozen tissues qrt-pcr Median 80 mir-21 Voortman 2010 (24) 14 countries Non-Asian AC SC 631 I II III Paraffin-embedded qrt-pcr Median 96 mir-21 tissues Wang et al (16) China Asian AC SC 88 I II III Serum qrt-pcr 5-fold 36 mir-21 Donnem 2011 (20) Norway Non-Asian SCC 94 I II IIIa Paraffin-embedded ISH Median 86 mir-155 tissues Raponi et al (21) University of Non-Asian SCC 54 I II III Frozen tissues qrt-pcr Median 36 mir-155 Michigan Saito 2011 (4) Maryland Non-Asian AC 89 I II III Frozen tissues qrt-pcr Median 80 mir-155 Saito 2011 (4) Norway Non-Asian AC 37 I II III Frozen tissues qrt-pcr Median 80 mir-155 Saito 2011 (4) Japan Asian AC 191 I II Frozen tissues qrt-pcr Median 80 mir-155 Voortman 2010 (24) 14 countries Non-Asian AC, SCC, other types Yanaihara et al (22) America Non-Asian AC, SCC 64 I II III IV 637 I II III Paraffin-embedded tissues qrt-pcr, ISH micrornas (mirna) Mean 96 mir-155 Tissues SH, qrt-pcr Mean 60 mir mir-21 and mir-155 in NSCLC AC, adenocarcinoma; SCC, squamous cell carcinoma; qrt-pcr, quantitative real-time PCR.

5 Jpn J Clin Oncol 2013;43(8) 817 Figure 2. The association between mir-21 expression and non-small-cell lung cancer (NSCLC) prognosis. (A) Forest plots of studies evaluating hazard ratios (HRs) of mir-21 expression to NSCLC survival rates. (B) Funnel plots of studies included in (A). Table 2. Associations between mir-21 expression and NSCLC prognosis grouped by selected factors Variables No. of studies High vs. low P for heterogeneity HR (95% CI) a HR (95% CI) b All ( ) 2.00 ( ) Cut-off Median ( ) 1.80 ( ) fold ( ) Ethnicity Asian ( ) 2.38 ( ) Non-Asian ( ) 1.57 ( ) Follow-up,60 months ( ) months ( ) 1.82 ( ) Source of RNA Tissue ( ) 1.91 ( ) Serum ( ) NSCLC, non-small-cell lung cancer. a Fixed-effects model. b Random-effects model. defined according to the main features of pooled studies including cut-off, ethnicity, follow-up and source of RNA (Table 2). In the cut-off subgroup, mir-21 expression measured by a median method showed the same result (HR ¼ 1.80, 95% CI ¼ , P ¼ for heterogeneity test, I 2 ¼ 89.4%) as the 2-fold method (HR ¼ 2.98, 95% CI ¼ , P ¼ for heterogeneity test, I 2 ¼ 0.0%). The association between higher mir-21 expression and worse NSCLC prognosis remained statistically significant in studies of Asian subjects (HR ¼ 2.38, 95% CI ¼ , P ¼ for heterogeneity test, I 2 ¼ 79.2%), tissue-source RNA (HR ¼ 1.91, 95% CI ¼ , P ¼ for heterogeneity test, I 2 ¼ 85.7%), serum-source RNA (HR ¼ 2.08, 95% CI ¼ , P ¼ for heterogeneity test, I 2 ¼ 13.9%), follow-up,60 months (HR ¼ 2.98, 95% CI ¼ , P ¼ for heterogeneity test, I 2 ¼ 0.0%) and follow-up.60 months (HR ¼ 1.82, 95% CI ¼ , P ¼ for heterogeneity test, I 2 ¼ 89.4%). The non-asian subgroup was an exception (HR ¼ 1.57, 95% CI ¼ , P ¼ for heterogeneity test, I 2 ¼ 88.2%). Finally, funnel plots and Egger s test were used to evaluate the publication bias (Fig. 2B). The P value of Egger s regression intercept was 0.096, indicating that there was no significant publication bias in the meta-analysis. THE ASSOCIATION BETWEEN MIR-155 EXPRESSION AND NSCLC PROGNOSIS For studies evaluating the association between mir-155 expression and NSCLC prognosis, a random model was applied because of the heterogeneity among the studies. mir-155

6 818 mir-21 and mir-155 in NSCLC Figure 3. The association between mir-155 expression and NSCLC prognosis. (A) Forest plots of studies evaluating HRs of mir-155 expression to NSCLC survival rates. (B) Funnel plots of studies included in (A). expression was significantly associated with NSCLC prognosis (HR ¼ 1.65, 95% CI ¼ , P ¼ for heterogeneity test, I 2 ¼ 68.3%) (Fig. 3A). The decreased survival rates of patients with high mir-155 expression were even more pronounced in the subgroup analysis of adenocarcinoma (HR ¼ 1.74, 95% CI ¼ , P ¼ for heterogeneity test, I 2 ¼ 0.0%). Likewise, funnel plots and Egger s test were used to evaluate the publication bias, the P value of Egger s regression intercept was (Fig. 3B). THE ASSOCIATION BETWEEN MIR-21 EXPRESSION AND LYMPHOID INFILTRATION OF NSCLC A total of seven publications included information about lymphoid infiltration of NSCLC and mir-21 expression, but only four of them reported statistical information including OR and 95% CI. We found that high mir-21 expression (relative mir-21 expression value. 1) conferred an increased risk of lymphoid infiltration for NSCLC (OR ¼ 1.93; 95% CI ¼ ; P ¼ for heterogeneity test, I 2 ¼ 0.0%) (Fig. 4). DISCUSSION In this meta-analysis of 1350 NSCLC cases from 10 independent publications, we found that both mir-21 and mir-155 overexpression were significantly associated with worse NSCLC prognosis. Furthermore, the subgroup analysis showed that the association was more prominent in studies of median cut-off, Asian subjects, tissue-source RNA and follow-up 60 months. In addition, NSCLC patients with high mir-21 expression (relative mir-21 expression value. 1) had more node metastatic potential than those with low mir-21 expression (relative mir-21 expression value, 1), indicating the need for randomized controlled trials with accurate cut-off values to set a standard mir-21 expression threshold with clinical implications. However, several limitations of our meta-analysis should be pointed out. First, marked heterogeneity of subjects existed when exploring the relationship between the two mirnas and NSCLC prognosis. The heterogeneity of the population was Figure 4. The pooled odds ratios (ORs) NSCLC lymphoid infiltration. High mir-21 expression was associated with higher risk of lymphoid infiltration than low mir-21 expression. probably due to the difference in the cut-off value of mirna, the duration of follow-up or the source of RNA. For example, for studies involving mir-21 and NSCLC prognosis, when we stratified them according to the cut-off value of mirna, heterogeneity disappeared in a 2-fold subgroup (P ¼ 0.807). Because such differences might have a residual confounding effect within these studies, we attempted to minimize the effect by using a random-effects model. Second, potential publication bias existed in the meta-analysis involving the association between mir-155 expression and NSCLC prognosis (P, 0.05), which could result in favor of publications with significant results. Third, the results of this meta-analysis were based on unadjusted estimates because some studies did not provide detailed information, for example, mir-21 expression levels among different NSCLC stages, to calculate adjusted estimates. To get a convincing conclusion on the value of mir-21 or mir-155 for the prognosis of NSCLC, other issues should also be addressed. First, an appropriate sample type should be determined to extract mirnas. Currently, different kinds of samples are used to detect mirnas including frozen tissues, formalin-fixed paraffin-embedded (FFPE) tissues, serum and sputum. Saito et al. found that the results from frozen tissues were in apparent contrast to those from FFPE tissues when exploringtheassociationofmir-21expressionandnsclc prognosis (4). Therefore, the type of samples could influence the outcome, and additional studies are needed to screen an ideal source of total RNA. Second, a standard cut-point for high vs. low mirna expression should be identified. Some

7 Jpn J Clin Oncol 2013;43(8) 819 studies have used two as the cut-off, whereas other studies have chosen mean as the cut-off. The inconsistency makes it difficult to directly compare or pool data from different studies. However, the lack of information on the average mir-21 or mir-155 expression level in global population limits the definition of a standard cut-point. Third, the prognostic impact of mirna differs depending on the histological subtype of NSCLC. Donnem et al. reported that high mir-155 expression was an independent negative prognostic factor in adenocarcinoma (ACs), while high mir-155 expression was an independent favorable prognosticator in SCC patients (20). Therefore, further investigations are necessary to explore the association between histotype and mirnas expression. At last, we wondered whether mir-21 or mir-155 may have more powerful prognostic effect when combined with other biomarkers. The expression of mir-21 or mir-155 in NSCLC is regulated by a complicated molecular network. Numerous regulatory factors contribute to the altered expression of mir-21, including the signal transducer and activator of transcription 3 (STAT3), activator protein 1, transforming growth factor-b and epidermal growth factor receptor (EGFR) (12,14). Likewise, mir-21 modulates thousands of downstream targets, such as programmed cell death 4 (11), phosphatase and tensin homolog (PTEN) (13), and reversion-inducing-cysteine-rich protein with kazal motifs (RECK) (14). Seike et al. showed that the changes in the expression of mature mir-21 were more remarkable in the presence of EGFR mutations in NSCLC (31). This study suggests that mir-21 may become a more powerful prognostic factor when combined with other factors. Furthermore, NSCLC is a heterogeneous disease which involves the aberration of multiple cells such as stromal, vascular and immune cells. However, altered expression of mirnas tends to be cell typespecific. Sempere et al. reported that altered expression of mir-21 and mir-34a was manifested within cancer cells, while altered expression of mir-126 and mir-155 was predominantly confined to endothelial cells and immune cells, respectively (32). Therefore, simultaneous assessment of a panel of tumor-specific mirnas may provide a more sensitive and specific diagnostic test than a single mirna. In conclusion, this meta-analysis provides statistical evidence for an association between high mir-21 or mir-155 expression and poor NSCLC survival outcomes. Meanwhile, we found that mir-21 expression was significantly associated with lymphoid infiltration status of NSCLC. However, due to the limitations of original studies included in the meta-analyses, larger, well-designed prospective studies are needed to confirm these findings, which may help unravel the underlying mechanisms of mirnas in NSCLC prognosis, lymphoid infiltration status and stage. Funding This work was supported by the National Natural Science Foundation of China (No , ), the Natural Science Foundation of Heilongjiang Province (No. ZD201020), the Scientific Research Foundation for ROCS, SEM. (No ) and Innovative Research Team Program in the University of Heilongjiang Province. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. Conflict of interest statement None declared. References 1. YuL,ToddNW,XingL,etal.Earlydetectionoflungadenocarcinoma in sputum by a panel of microrna markers. Int J Cancer 2010;127: Chansky K, Sculier JP, Crowley JJ, et al. The International Association for the Study of Lung Cancer Staging Project: prognostic factors and pathologic TNM stage in surgically managed non-small cell lung cancer. J Thorac Oncol 2009;4: van Rens MT, de la Rivière AB, Elbers HR, et al. Prognostic assessment of 2,361 patients who underwent pulmonary resection for non-small cell lung cancer, stage I, II, and IIIA. Chest 2000;117: Saito M, Schetter AJ, Mollerup S, et al. The association of microrna expression with prognosis and progression in early-stage, non-small cell lung adenocarcinoma: a retrospective analysis of three cohorts. Clin Cancer Res 2011;17: Bartel DP. MicroRNAs: genomics, biogenesis, mechanism, and function. Cell 2004;116: Esquela-Kerscher A, Slack FJ. Oncomirs micrornas with a role in cancer. Nat Rev Cancer 2006;6: Zhang B, Pan X, Cobb GP, Anderson TA. MicroRNAs as oncogenes and tumor suppressors. Dev Biol 2007;302: Calin GA, Croce CM. MicroRNA signatures in human cancers. Nat Rev Cancer 2006;6: Toyama T, Kondo N, Endo Y, et al. High expression of microrna-210 is an independent factor indicating a poor prognosis in Japanese triple-negative breast cancer patients. Jpn J Clin Oncol 2012;42: Yan LX, Huang XF, Shao Q, et al. MicroRNA mir-21 overexpression in human breast cancer is associated with advanced clinical stage, lymph node metastasis and patient poor prognosis. RNA 2008;14: Asangani IA, Rasheed SA, Nikolova DA, et al. MicroRNA-21 (mir-21) post-transcriptionally downregulates tumor suppressor Pdcd4 and stimulates invasion, intravasation and metastasis in colorectal cancer. Oncogene 2008;27: Iliopoulos D, Jaeger SA, Hirsch HA, et al. STAT3 activation of mir-21 and mir-181b-1 via PTEN and CYLD are part of the epigenetic switch linking inflammation to cancer. Mol Cell 2010;39: Zhang JG, Wang JJ, Zhao F, et al. MicroRNA-21 (mir-21) represses tumor suppressor PTEN and promotes growth and invasion in non-small cell lung cancer (NSCLC). Clin Chim Acta 2010;411: Krichevsky AM, Gabriely G. MiR-21: a small multi-faceted RNA. J Cell Mol Med 2009;13: Zhou X, Ren Y, Moore L, et al. Downregulation of mir-21 inhibits EGFR pathwayand suppressesthe growth of human glioblastoma cells independent of PTEN status. Lab Invest 2010;90: Wang ZX, Bian HB, Wang JR, et al. Prognostic Significance of Serum mirna-21 Expression in Human Non-Small Cell Lung Cancer. JSurg Oncol 2011;104: Liu XG, Zhu WY, Huang YY, Ma LN, Zhou SQ, et al. High expression of serum mir-21 and tumor mir-200c associated with poor prognosis in patients with lung cancer. Med Oncol 2012;29: Gao W, Yu Y, Cao H, et al. Deregulated expression of mir-21, mir-143 and mir-181a in non small cell lung cancer is related to clinicopathologic characteristics or patient prognosis. Biomed Pharmacother 2010;64:

8 820 mir-21 and mir-155 in NSCLC 19. Gao W, Shen H, Lui L, et al. MiR-21 overexpression in human primary squamous cell lung carcinoma is associated with poor patient prognosis. J Cancer Res Clin Oncol 2011;137: Donnem T, Eklo K, Berg T et al. Prognostic impact of MiR-155 in non-small cell lung cancer evaluated by in situ hybridization. J Transl Med 2011;9: Raponi M, Dossey L, Jatkoe T. MicroRNA classifiers for predicting prognosis of squamous cell lung cancer. Cancer Res 2009;69: Yanaihara N, Caplen N, Bowman E, et al. Unique microrna molecular profiles in lung cancer diagnosis and prognosis. Cancer Cell 2006;9: Markou A, Tsaroucha EG, Kaklamanis L, et al. Prognostic value of mature microrna-21 and microrna-205 overexpression in non-small cell lung cancer by quantitative real-time RT-PCR. Clin Chem 2008;54: Voortman J, Goto A, Mendiboure J, et al. MicroRNA expression and clinical outcomes in patients treated with adjuvant chemotherapy after complete resection of non-small cell lung carcinoma. Cancer Res 2010;70: Kim HS, Suh DH, Kim MK, Chung HH, Park NH, Song YS. Systematic lymphadenectomy for survival in patients with endometrial cancer: a meta-analysis. Jpn J Clin Oncol 2012;42: Cao C, Ying T, Fang JJ, et al. Polymorphism of vascular endothelial growth factor 2578C/A with cancer risk: evidence from subjects. Med Oncol 2011;28: Jiang DK, Yao L, Ren WH, Wang WZ, Peng B, Yu L. TP53 Arg72Pro polymorphism and endometrial cancer risk: a meta-analysis. Med Oncol 2011;28: FangF,WangJ,YaoL,YuXJ,YuL,YuL.Relationshipbetween XRCC3 T241M polymorphism and gastric cancer risk: a meta-analysis. Med Oncol 2011;28: Stroup DF, Berlin JA, Morton SC, et al. Meta-analysis of observational studies in epidemiology: a proposal for reporting. Meta-analysis Of Observational Studies in Epidemiology (MOOSE) group. JAMA 2000;283: Higgins JP, Thompson SG, Deeks JJ, et al. Measuring inconsistency in meta-analyses. BMJ 2003;327: Seike M, Goto A, Okano T, et al. MiR-21 is an EGFR-regulated anti-apoptotic factor in lung cancer in never-smokers. PNAS 2009;106: Sempere LF, Preis M, Yezefski T, et al. Fluorescence-based codetection with protein markers reveals distinct cellular compartments for altered MicroRNA expression in solid tumors. Clin Cancer Res 2010;16: