A Network Meta-Analysis on the Diagnostic Values of MiR-21/MiR-16/MiR-34a/let-7b-5p/let-7g/MiR-218-5p in Non-Small-Cell Lung Cancer

Yifei Yun ¹, Yutong Zhang ¹, Yao Ou ¹, Fengbin Zhang ^2,*, Zhonghua Lu ^1,*

¹Department of Radiotherapy, Changzhou Tumor Hospital, Soochow University, Changzhou 213000, China; 119424774@qq.com (Yun Y); 1309711393@qq.com (Zhang Y); ouyao@163.com (Ou Y)

²Department of Radiotherapy, Xijing Hospital of Air Force Military Medical University, Xi'an 710032, China

*Correspondence to:2732779813@qq.com (Zhang F);Luzhonghua8687@163.com (Lu Z)

Received: 29 October 2020; Accepted: 30 November 2020; Published: 11 December 2020

Abstract: Non-small-cell lung cancer (NSCLC) is a major subtype of lung cancer, accompanied by dismal prognosis. Recent studies have reported microRNAs (miRNAs) may act as diagnosis biomarkers in NSCLC. This network meta-analysis was performed toidentify miRNA and their potential diagnostic value for NSCLC systematically. Several databases were recruited for eligible studies, and the retrieval spectrum ranged from the inception to June 2018. Besides, 142 NSCLC patients and 168 patients with benign lung disease were enrolled. Direct and indirect evidence were incorporated to evaluate weighted mean difference (WMD) and ranking probability (RP) based on surface under the cumulative ranking curves (SUCRAs). The expressions of miRNAs were further validated by RT-qPCR. The sensitivity, specificity, and receiver operator characteristic (ROC) curve were used to evaluate diagnostic performance. The expression of miR-192 and miR-17 was lower than that of miR-21 in plasma samples of NSCLC. However, the expression of miR-16, miR-34a and miR-1-5p was lower while the levels of let-7g and mR-218-5p were higher than that of miR-21 in NSCLC tissues. miR-21 and let-7g levels were upregulated in plasma/tissues of NSCLS, but miR-16, miR-34a, let-7b-5p and miR-218-5p expression wasdownregulated. In the analysis of ROC curve, let-7g (sensitivity: 0.901; specificity: 0.893) and miR-218-5p (0.817, 0.964) in plasma as well as miR-16 (0.915, 0.964), miR-34a (0817, 0.964) and miR-218-5p (0.831, 0.964) in tissues revealed higher diagnostic value for NSCLC. It demonstrates that let-7g/miR-218-5p in plasma and miR-16/miR-34a/miR-218-5p in tissues can potentially be used as diagnostic biomarkers of NSCLC.

Keywords:non-small-cell lung cancer; microRNA;case-control study;network meta-analysis

1. Introduction

Lung cancer is one of the most common malignancies with high incidence and morbidity in male population globally [1]. Over 80% of all lung cancer cases were diagnosed as NSCLC, which mainly involves three major subtypes: squamous cell carcinoma, adenocarcinoma and large-cell carcinoma, with distinguishing pathological characteristics [2,3]. The high mortality rate of NSCLC is significantly correlated with the late detection at advanced stages, and good screening methods are rare so far, with an overall 5-year survival rate as low as 15% [4,5]. In contrast, patients who are diagnosed at stage I and receive effective treatment achieve a 5-year overall survival rate of about 80%, and a 2-year cancer-specific survival of about 90% [6,7]. Thus, effective and accurate diagnostic methods for early diagnosis of NSCLC are urgently needed to improve the prognosis as well as reduce the mortality of NSCLC.

Currently, clinical screening tools applied in NSCLC are based on spiral computed tomography (CT), chest X-rays, sputum cytology, bronchoscopy, anatomical pathology, or their combination [8]. However, radiation carcinogenesis from CT has been debated for decades [9]. The fluorescence bronchoscopy is also an invasive technique [10]. Furthermore, the great cost and validity of these screenings must be taken into consideration. Consequently, numerous tumor-specific molecular biomarkers have been identified for NSCLC detection. Although several lung-specific biomarkers may carry useful predictive information in patients with NSCLC, including the carcinoembryonic antigen, neuron-specific enolase, chromogranin A, cytokeratin-21 fragment, neuron-specific enolase, tissue polypeptide-specific antigen, and cancer antigen-125, however, their clinical application is limited due to the low specificity, sensitivity, and reproducibility [8,11]. MicroRNAs (miRNAs), particularly those in plasma and tissue samples, may be promising prognostic biomarkers for NSCLC [12,13].

MicroRNA (miRNAs or miRs), a class of small non-protein-coding RNAs with a length of 18–25 nucleotides, affects both the translation and the stability of messenger RNAs, therefore regulating the gene expression at a post-transcriptional level [14,15]. The dysregulation of miRNAs has been reported in many diseases, particularly, liver cancers [16,17]. MiRNAs, presented being highly stable even under abominable conditions, which strengthen their potentiality to serve as cancer biomarkers in clinical samples (e.g., plasma, serum or tissue) [18–21]. Recently, accumulating studies have suggested that miRNAs profiling may be a useful method in the initial screening process for NSCLC [22–24]. However, many of these studies procured unfavorable results, and various subject ages, small sample sizes and different miRNAs profiles enrolled in these studies might provide an obscure explanation [24–27]. Therefore, a network meta-analysis aims to determine the potential diagnostic value of miRNAs for NSCLC systematically is helpful and essential.

2. Methods

2.1. Retrieval Strategy

The electronic retrieval of databases, including PubMed, Cochrane Library and Embase, were adopted from their inceptions up to June 2018, supplemented with manual retrieval of relevant references. The electronic retrieval combined keywords and free words to search for references, and keywords included NSCLC, miRNAs and expression level.

2.2. Study Selection

Eligible studies met the inclusion criteria: 1) study type: case-control study; 2) interventions: miR-21, miR-126, miR-141, miR-200c, miR-15a, miR-16, miR-34a, miR-128, miR-210, miR-200b, miR-193b, miR-301, let-7g, miR-192, miR-17, miR-103, miR-93, miR-451, miR-1-5p, miR-7b-5p, miR-1290, miR-149-5p, miR-196a-5P and miR-218-5p; 3) study subject: plasma/tissue samples from patients with NSCLC; (4) outcome indicator: miRNA expression levels in plasma/tissues of NSCLC. Exclusion criteria were as follows: 1) studies without complete data (e.g. unpaired study); 2) no case-control study; 3) no randomized control trial; 4) letters, reviews or summaries; 5) non-English studies; 6) literatures on non-human research; 7) duplicated literatures.

2.3. Data Extraction and Quality Assessment

The data of included literatures were independently extracted by two researchers according to a unified data collection form. If there were disputes, the data extraction should be discussed and negotiated by several researchers until consensus was reached. Two investigators assessed the study quality using the Newcastle-Ottawa Scale (NOS) including 3 subscales: selection, comparability and outcome [28]: NOS1, a. representative exposed cohort or not; NOS2, the non-exposed cohort was drawn from the same community as the exposed cohort or not; NOS3, the study had credible record or structured interview; NOS4, the initial research had no outcome; NOS5, the study selected and analyzed the controls based on important factors; NOS6, the study controlled any other confounding factors; NOS7, independent blind assessment; NOS8, follow-up study was long enough for outcome to occur; NOS9, complete follow-up – all subjects accounted for or a few subjects lost to follow-up unlikely to introduce bias. A maximum of 9 points can be given for one study. Studies that got 5 points or more were enrolled in the network meta-analysis.

2.4. Study Subjects

A total of 142 NSCLC patients (as NSCLC group) were diagnosed by fiber bronchoscope the Cancer Hospital of China Medical University, Liaoning Cancer Hospital & Institute between June 2015 and June 2016. There were 92 males and 50 females, with a mean age of (56.80 ± 9.20) years. According to the pathological diagnosis, there were 84 cases of squamous cell carcinoma and 58 cases of adenocarcinoma. All patients underwent primary tumor, and received no radiotherapy or chemotherapy. According to the 7th edition Union for International Cancer Control (UICC)/American Joint Committee on Cancer (AJCC) Tumor-Node-Metastasis(TNM) classification for NSCLC [29], there were 18 cases with stage T1a, 34 cases with stage T1b, 52 cases with stage T2a, 24 cases of stage T2b and 14 cases with T3. We selected 168 cases (120 male and 48 female) of benign lung disease as control group. Subjects had a mean age of (57.3 5 ± 9.81) years. Among them, there are 60 cases of pulmonary tuberculosis, 42 cases of pulmonary sarcoidosis, 54 cases of community acquired pneumonia and 12 cases of inflammatory pseudotumor, and no one had a history of tumor. No differences were shown in age and sex among NSCLC and control groups. This study was approved by the Ethics Committee of Cancer Hospital of China Medical University, Liaoning Cancer Hospital & Institute.  Written informed consent was obtained from all subjects.

2.5. Sample Collection

Fresh biopsy samples were preserved in RNAlater TissueProtect tubes that five times larger than the sample, and tubes were stored at -80°C properly. All samples had ≤ 0.5 cm thickness. Besides, we collected peripheral plasma of fasting patients in the morning, preserved in sodium citrate tubes at 4°C. Within 1 hour after collection, the plasma samples were centrifuged at 3000 rpm for 10 min at 4°C. The collected supernatant was sub-packed in Eppendorf (EP) tubes (600 μL per tube) which had sterilized by treatment with 0.1% diethyl pyrocarbonate, and then frozen at -80°C.

2.6. Reverse Transcription-Quantitative Polymerase Chain Reaction (RT-qPCR)

Total RNA was extracted from plasma and tissues according to the manufacturer’s protocol of miRNAVana PARIS kit (P/N AM1556) and miRVana^TM miRNA isolation kit (P/N AM1560). Primers (Table 1) were designed by Primer Premier 5.0 according to the gene sequences published in GenBank and synthesized by Shanghai Sangon Biological Engineering Technology & Services Co., Ltd, (Shanghai, China). The cDNAs were synthesized by TaqMan miRNA reverse transcription (RT) kit (P/N 4366596, Applied Biosystems, Inc., Carlsbad, CA, USA). RT-qPCR was performed according to the manufacturer’s protocol of Applied Biosystems 7500 RT-PCR with three parallel wells set. U6 served as the internal reference. The 2^-ΔΔCt method was used to calculate the relative mRNA expression.

Table 1. Primer sequences of related genes for reverse transcription-quantitative polymerase chain reaction.

Note: miR, microRNA.

2.7. Statistical Analysis

Initially, we performed pair-wise meta-analyses of direct evidence using the fixed-effects model, with R version 3.2.1 and the meta package. The pooled estimates of weighted mean difference (WMD) and 95% confidence intervals (CIs) of three endpoint outcomes were shown. Chi-square test and I-square test were used for testing heterogeneity among the studies [30]. Second, R 3.2.1 software was used to draw a network diagram, in which each node represented an intervention measure, the size of node represented sample size and the thickness of line between nodes represented the number of included studies. Third, we did a random-effect network meta-analysis with the gemtc package. It models the relative effects (e.g. log-odds ratio) fitting a generalized linear model (GLM) under the Bayesian framework by linking to JAGS, OpenBUGS or WinBUGS as first described by Lu and Ades [31] and extended by others [32,33]. To assist the interpretation of WMDs, we calculated the probability of each intervention being the most effective treatment method based on a Bayesian approach using probability values summarized as surface under the cumulative ranking curve (SUCRA), the larger the SUCRA value, the better the rank of the intervention [34,35]. All computations were done using R (V.3.2.2) package gemtc (V.0.6), along with the Markov Chain Monte Carlo engine Open BUGS (V.3.4.0). Statistical analysis was performed using the SPSS 19.0 software (IBM Corp., Armonk, NY, USA). Data are presented as the mean ± standard deviation. Comparisons between two groups were analyzed using unpaired t test. Comparisons among multiple groups were assessed using one-way analysis of variance. Enumeration data are expressed as percentage and analyzed by chi-square test. Receiver operator characteristic (ROC) curve was applied for diagnostic value of miRNAs of plasma/tissues in NSCLC. All P values were two-sided, and apvalue of less than 0.05 was considered statistically significant.

3. Results

3.1. Baseline Characteristics of Included Literatures

In this study, we totally retrieved 1173 related literatures, among which 16 duplicated publications, 113 letters or reviews and 120 non-English literatures were eliminated. In the remained 924 literatures, we excluded 401 no case control studies, 513 literatures unrelated to miRNA expression in NSCLC and 1 literature without sufficient data. Finally, nine eligible case-controls published from 2012 to 2016 were enrolled into this network meta-analysis [36–44] (Figure 1). Among these literature, 6 studies were based on the subjects of Asians, 3 studies were based on the subjects of Caucasians. The baseline characteristics of included literatures are shown in Table 2and the study quality assessed by the Newcastle-Ottawa Scale (NOS) is shown in Appendix Figure 1. Among the enrolled studies, there were more studies focusing on tissues samples and concerning miR-21, miR-126, miR-141 and miR-200c (Figure 2).

Table 2. The baseline characteristics for included studies.

Note: RT-qPCR, Reverse transcription-quantitative polymerase chain reaction; miR, microRNA.

3.2. Pairwise Meta-Analysis of the Expression Levels of MiRNAs in Plasma/Tissues in NSCLC

Pairwise meta-analysis of improvements in expression level in plasma and tissues were shown in Table 3. In the plasma samples, the expression levels of miR-192 and miR-17 in NSCLC were lower than the miR-21; the expression levels of miR-93 and miR-451 were lower compared with the miR-192. In the tissue samples, miR-126, miR-200c, miR-15a, miR-16, miR-34a, miR-1-5p and let-7b-5p showed lower expression levels in NSCLC than miR-21; miR-141, miR-210, let-7g, miR-196a-5p and miR-218-5p revealed higher expression levels than miR-21; miR-200c, miR-128 and miR-210 indicated higher expression levels compared with miR-126; miR-34a showed lower expression level than miR-126; miR-200c presented higher expression level compared with miR-141. All in all, miR-21, miR-414, miR-210 and let-7g showed higher expression while miR-1-5p, let-7b-5p, miR-34a, miR-93 and miR-451 showed lower expression in plasma/tissues in NSCLC as compared to in benign pulmonary diseases.

Table 3. Pairwise meta-analysis of improvements in expression level in blood and tissues in NSCLC.

Note: NSCLC, non-small-cell lung cancer; WMD, weighted mean difference; CI, confidence interval; miR, microRNA.

3.3. The Main Results of Network Meta-Analysis for MiRNA Expression Level in NSCLC

There was no difference in miRNA expression level in NSCLC between the plasma and tissues samples (Supplementary Table 1).

3.4. Cumulative Probability Ranking of MiRNA Expression Levels in NSCLC

As shown in Figure 3, the miRNA expression levels in the plasma and tissues in NSCLC revealed that miR-16 ranked the highest SUCRA value (91.57%), while miR-21 ranked the lowest SUCRA value (21.57%). Expression levels in tissues (-ΔΔCt)-A showed that miR-16 and miR-34a ranked relatively high SUCRA value (83.31%; 86.23%), but let-7g ranked relatively low SUCRA value (17.54%). Expression levels in tissues (-ΔΔCt)-B indicated that miR-1-5P and let-7b-5p ranked comparatively high SUCRA value (84.00%; 85.57%), while miR-218-5p ranked comparatively low SUCRA value (20.57%). These results revealed that miR-16 showed relatively lower and miR-21 relatively higher expression levels in the plasma in NSCLC, whereas miR-16, miR-34a, miR-1-5P and let-7b-5p showed comparatively high and let-7g and

Figure 3. Cumulative probability ranking of miRNA expression level in NSCLC Note: NSCLC, non-small-cell lung cancer; miRNA, microRNA.A, SUCRA value of the miRNA expression levels in the plasma and tissues in NSCLC. B, SUCRA value of expression levels in tissues (-ΔΔCt)-A. C, SUCRA value of expression levels in tissues (-ΔΔCt)-B.

3.5.Comparison of MiRNA Expression Levels in Plasma and Tissue Samples between the NSCLC and Control Groups

Figure 4. Relative miRNA expression levels in plasma and tissue samples between the NSCLC and control groups Note: A, miRNA expression level in plasma between the NSCLC and control groups; B, miRNA expression level in tissues between the NSCLC and control groups; NSCLC, non-small-cell lung cancer; miRNA, microRNA; *, p < 0.05 compared with the control group.

3.6. Diagnostic Performance of Different MiRNAs in NSCLC

Figure 5. ROC curve concerning the expression level of miRNAs in plasma or tissues sample in NSCLC Note: A, ROC curve concerning the expression level of miRNAs in plasma in NSCLC; B, ROC curve concerning the expression level of miRNAs in tissues in NSCLC; NSCLC, non-small-cell lung cancer; miRNA, microRNA; ROC, receiver operator characteristic.

Table 4. Results of ROC curve concerning the expression level of microRNAs in blood or tissues samples in NSCLC.

Note: NSCLC, non-small-cell lung cancer; miR, microRNA; ROC, receiver operator characteristic; AUC, area under curve; BCP, best cutoff point.

4. Discussion

Highly accurate tests for the early detecting and monitoring of NSCLC patients have long been major concerns in cancer research. Unfortunately, current diagnostic biomarkers and tools have their own inherent deficiencies mentioned in the introduction of this study. Highly sensitive and specific, minimally invasive biomarkers that capable of detecting early neoplastic lesions are needed to increase the probability of detecting NSCLC early. In the present study, we perform a systematic meta-analysis to gain insight into the particular diagnostic performance of miRNAs for NSCLC.

Pairwise meta-analysis of miRNA expression levels in plasma and tissue samples of NSCLC intuitively manifested that in the plasma samples, the expression levels of miR-192 and miR-17 in NSCLC were lower than miR-21, while in the tissue samples, miR-126, miR-200c, miR-15a, miR-16 and miR-34a showed lower expression levels in NSCLC than miR-21, but miR-141, miR-210 and let-7g revealed higher expression levels. Regarding plasma can easily be sampled and subjected to analysis, plasma samples are relatively noninvasive specimen and have been widely used, in which miRNAs stably packaged in microvesicles and can be easily detected [21,45]. MiR-21, for example, one of cancer-related miRNAs studied widely, might play a key role in most cancers [46]. Upregulated expression level of miR-21 had been proved to be associated with poor outcomes in cancer patients [47]. Thus, it’s possible that miR-21 downregulate the pro-apoptotic phosphatase and tensin homolog, which in turn, may stimulate growth and invasion in NSCLC cells [48]. From the Cancer Genome Atlas database for lung adenocarcinoma, hsa-miR-192 was related to lung cancer [49]. Besides, miR-126 was found to be a cancer promotor of NSCLC [50]; however, it was determined to be expressed at low levels in serum of NSCLC patients [51]. A study suggested that high miR-200c expression was detected in more half of the 110 cases and associated with poor outcomes of NSCLC [52] while another study demonstrated that miR-200c was a suppressor in NSCLC [53]. A more recent study identified the miR-17 was also found to be poorly expressed in plasma in lung cancer [54]. A study documented that miR-15a expression was decreased in NSCLC tissues when compared with the adjacent tissues [55]. However, the above studies only compared the expression of the same miRNA between NSCLC and normal tissues/cells. Different from these studies, we compared the expression of miR-21 with that of other miRNAs in NSCLC plasma and tissue samples.

Therefore, we then detected the expression of those miRNAs in plasma and tissues of NSCLC patients and patients with benign pulmonary diseases. The results suggested that miR-21 and let-7g were upregulated while miR-16, miR-34a and miR-218-5p were downregulated in plasma and tissues of NSCLC patients. Consistent with our study, miR-16 was significantly downregulated in tissue samples of NSCLC patients [56]. Besides, miR-34a expression in plasma and tissues was negatively correlated with lymph node metastasis in NSCLC and its downregulation indicated a poor outcome for NSCLC patients [57]. Consistently, miR-21 was upregulated while miR-218-5p was downregulated in adenocarcinoma tissues in NSCLC [37,58,59]. Hsa-let-7g was found upregulated in adenocarcinoma specimens [60]. Those results predicted that miR-21, let-7g, miR-218-5p, miR-16 and miR-34a might be effective prognostic factors for NSCLC. To verify whether they could be used as prognostic factors or not, ROC curve analysis was conducted to detect the sensitivity and specificity in plasma and tissues. The sensitivity and specificity of let-7g and miR-218-5p in plasma, miR-16, miR-34a and miR-218-5p in tissues were relatively high (all sensitivity and specificity > 0.8), indicating better diagnostic value for NSCLC. Thereafter, these miRNAs can be used as diagnostic markers for NSCLC patients with high specificity and sensitivity.

5. Conclusions

This is a comprehensive, systematic study that focuses on the overall diagnostic performance of miRNAs for NSCLC patients, showing that assays based on miRNAs have a high accuracy which highlights their clinical application. We also conducted a subgroup analysis based on the expression levels of miR-21, miR-126, miR-141, miR-200c, miR-15a, miR-16, miR-34a, miR-128, miR-210, miR-200b, miR-193b, miR-301, let-7g, miR-192, miR-17, miR-103, miR-93, miR-451, miR-1-5p, miR-7b-5p, miR-1290, miR-149-5p, miR-196a-5P and miR-218-5p in plasma/tissue samples in NSCLC, demonstrating clinical significance to the diagnosis and precaution of NSCLC. Finally, there is conflict of interest claimed in this study, strengthening the reliance of the findings. Certain limitations should also be considered while clarifying the results of this network meta-analysis. Firstly, due to the differences in the samples used in detecting the25 miRNAs and the number of included studies for paired comparison among miRNAs, the results of this meta-analysis are restricted. Secondly, most of the included studies lack records of follow-up times which may influence the diagnostic performance of the miRNAs. Thirdly, some studies may be missed in the selection processes. Despite all the limitations, current study indicated that miRNAs had a high diagnostic accuracy, which may serve as a novel screening tool or be combined with other biomarkers in NSCLC diagnosis. Further large-scale studies are required to validate the clinical application of miRNA profiling.

Supplementary Materials:This section must be cited in the main text.The following are available online at http://mbn.techlandgroup.com.

Author Contributions:All authors have read and agreed to the published version of the manuscript. Conceptualization,Yun Y and ZhangY; methodology, Yun Y and ZhangY; software, ZhangF and Lu Z; validation, Yun Y, Zhang Y and OuY; formal analysis, writing—original draft preparation, Yun Y, Zhang Y and OuY.

Funding:None

Conflicts of Interest: All authors declare no conflict of interest.

Figure A1. The Newcastle-Ottawa Scale evaluation of included literatures. Note: NOS, Newcastle-Ottawa Scale.

Copyright Statement

References

1. Bray F, Ferlay J, Soerjomataram I, Siegel RL, Torre LA, et al. Global cancer statistics 2018: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA: A Cancer Journal for Clinicians, 2018, 68: 394–424.

2. Xiang J, Hang JB, Che JM, Li HC. MiR-25 is upregulated in non-small cell lung cancer and promotes cell proliferation and motility by targeting FBXW7. International Journal of Clinical and Experimental Pathology, 2015, 8: 9147–9153.

3. Zhang N, Jiang BY, Zhang XC, Xie Z, Su J, et al. The BCL11A-XL expression predicts relapse in squamous cell carcinoma and large cell carcinoma. Journal of Thoracic Disease, 2015, 7: 1630–1636.

4. Jezierski T, Walczak M, Ligor T, Rudnicka J, Buszewski B. Study of the art: canine olfaction used for cancer detection on the basis of breath odour. Perspectives and limitations. Journal of Breath Research, 2015, 9: 027001.

5. Zhou X, Li D, Wang X, Zhang B, Zhu H, et al. Galectin-1 is overexpressed in CD133+ human lung adenocarcinoma cells and promotes their growth and invasiveness. Oncotarget, 2015, 6: 3111–3122.

6. Lee KS, Pua BB. Alternative to surgery in early stage NSCLC-interventional radiologic approaches. Translational Lung Cancer Research, 2013, 2: 340–353.

7. Grogan EL, Deppen SA, Chen H, Ballman KV, Verdial FC, et al. Abstract LB-296: FDG-PET avidity negatively impacts survival in pStage I NSCLC in the ACOSOG Z4031 trial. Cancer Research, 2013, 73: 296–296.

8. Sanfiorenzo C, Ilie MI, Belaid A, Barlesi F, Mouroux J, et al. Two panels of plasma microRNAs as non-invasive biomarkers for prediction of recurrence in resectable NSCLC. PloS One, 2013, 8: e54596.

9. Westra SJ. The communication of the radiation risk from CT in relation to its clinical benefit in the era of personalized medicine: part 2: benefits versus risk of CT. Pediatric Radiology, 2014, 44 Suppl 3: 525–533.

10. Devan L, Santhosham R, Hariharan R. Non-invasive Method of Characterization of Fibrosis and Carcinoma Using Low-Dose Lung CT Images. Proceedings of the IEEE International Conference on Systems, Man, and Cybernetics, Manchester, UK, 2013, pp. 2168-2172.

11. Cabrera-Alarcon JL, Carrillo-Vico A, Santotoribio JD, Leon-Justel A, Sanchez-Gil R, et al. CYFRA 21-1 as a tool for distant metastasis detection in lung cancer. Clinical Laboratory, 2011, 57: 1011–1014.

12. Wang H, Wu S, Zhao L, Zhao J, Liu J, et al. Clinical use of microRNAs as potential non-invasive biomarkers for detecting non-small cell lung cancer: a meta-analysis. Respirology, 2015, 20: 56–65.

13. Petriella D, De Summa S, Lacalamita R, Galetta D, Catino A, et al. miRNA profiling in serum and tissue samples to assess noninvasive biomarkers for NSCLC clinical outcome. Tumour Biology: The Journal of the International Society for Oncodevelopmental Biology and Medicine, 2016, 37: 5503–5513.

14. Sun L, Zhao M, Zhang J, Lv M, Li Y, et al. MiR-29b Downregulation Induces Phenotypic Modulation of Vascular Smooth Muscle Cells: Implication for Intracranial Aneurysm Formation and Progression to Rupture. Cellular physiology and biochemistry: International Journal of Experimental Cellular Physiology, Biochemistry, and Pharmacology, 2017, 41: 510–518.

15. Latini A, Ciccacci C, Novelli G, Borgiani P. Polymorphisms in miRNA genes and their involvement in autoimmune diseases susceptibility. Immunologic Research, 2017, 65: 811–827.

16. Wang L, Zhang X, Jia LT, Hu SJ, Zhao J, et al. c-Myc-mediated epigenetic silencing of MicroRNA-101 contributes to dysregulation of multiple pathways in hepatocellular carcinoma. Hepatology, 2014, 59: 1850–1863.

17. Najafi Z, Sharifi M, Javadi G. Degradation of miR-21 induces apoptosis and inhibits cell proliferation in human hepatocellular carcinoma. Cancer Gene Therapy, 2015, 22: 530–535.

18. Syring I, Bartels J, Holdenrieder S, Kristiansen G, Muller SC, et al. Circulating serum miRNA (miR-367-3p, miR-371a-3p, miR-372-3p and miR-373-3p) as biomarkers in patients with testicular germ cell cancer. The Journal of Urology, 2015, 193: 331–337.

19. Mar-Aguilar F, Mendoza-Ramirez JA, Malagon-Santiago I, Espino-Silva PK, Santuario-Facio SK, et al. Serum circulating microRNA profiling for identification of potential breast cancer biomarkers. Disease Markers, 2013, 34: 163–169.

20. Cuk K, Zucknick M, Heil J, Madhavan D, Schott S, et al. Circulating microRNAs in plasma as early detection markers for breast cancer. International Journal of Cancer, 2013, 132: 1602–1612.

21. Mitchell PS, Parkin RK, Kroh EM, Fritz BR, Wyman SK, et al. Circulating microRNAs as stable blood-based markers for cancer detection. Proceedings of the National Academy of Sciences of the United States of America, 2008, 105: 10513–10518.

22. Patnaik SK, Kannisto ED, Mallick R, Vachani A, Yendamuri S. Whole blood microRNA expression may not be useful for screening non-small cell lung cancer. PloS One, 2017, 12: e0181926.

23. Geng Q, Fan T, Zhang B, Wang W, Xu Y, et al. Five microRNAs in plasma as novel biomarkers for screening of early-stage non-small cell lung cancer. Respiratory Research, 2014, 15: 149.

24. Huang Y, Hu Q, Deng Z, Hang Y, Wang J, et al. MicroRNAs in body fluids as biomarkers for non-small cell lung cancer: a systematic review. Technology in Cancer Research & Treatment, 2014, 13: 277–287.

25. Ollier M, Chamoux A, Naughton G, Pereira B, Dutheil F. Chest CT scan screening for lung cancer in asbestos occupational exposure: a systematic review and meta-analysis. Chest, 2014, 145: 1339–1346.

26. Arcangeli S, Agolli L, Portalone L, Migliorino MR, Lopergolo MG, et al. Patterns of CT lung injury and toxicity after stereotactic radiotherapy delivered with helical tomotherapy in early stage medically inoperable NSCLC. The British Journal of Radiology, 2015, 88: 20140728.

27. Kozu Y, Maniwa T, Takahashi S, Isaka M, Ohde Y. Results of surgical treatment for non-small cell lung cancer with positive sputum cytology: experience from a single institution. The Thoracic and Cardiovascular Surgeon, 2014, 62: 588–592.

28. Lo CK, Mertz D, Loeb M. Newcastle-Ottawa Scale: comparing reviewers' to authors' assessments. BMC Medical Research Methodology, 2014, 14: 45.

29. Edge SB, Compton CC. The American Joint Committee on Cancer: the 7th edition of the AJCC cancer staging manual and the future of TNM. Annals of Surgical Oncology, 2010, 17: 1471–1474.

30. Minami S, Kijima T, Shiroyama T, Okafuji K, Hirashima T, et al. Randomized Phase II trial of paclitaxel and carboplatin followed by gemcitabine switch-maintenance therapy versus gemcitabine and carboplatin followed by gemcitabine continuation-maintenance therapy in previously untreated advanced non-small cell lung cancer. BMC Research Notes, 2013, 6: 3.

31. Lu G, Ades AE. Combination of direct and indirect evidence in mixed treatment comparisons. Statistics in Medicine, 2004, 23: 3105–3124.

32. Dias S, Sutton AJ, Ades AE, Welton NJ. Evidence synthesis for decision making 2: a generalized linear modeling framework for pairwise and network meta-analysis of randomized controlled trials. Medical Decision Making: An International Journal of the Society for Medical Decision Making, 2013, 33: 607–617.

33. Dias S, Welton NJ, Caldwell DM, Ades AE. Checking consistency in mixed treatment comparison meta-analysis. Statistics in Medicine, 2010, 29: 932–944.

34. Chaimani A, Higgins JP, Mavridis D, Spyridonos P, Salanti G. Graphical tools for network meta-analysis in STATA. PloS One, 2013, 8: e76654.

35. Salanti G, Ades AE, Ioannidis JP. Graphical methods and numerical summaries for presenting results from multiple-treatment meta-analysis: an overview and tutorial. Journal of Clinical Epidemiology, 2011, 64: 163–171.

36. Yang C, Sun C, Liang X, Xie S, Huang J, et al. Integrative analysis of microRNA and mRNA expression profiles in non-small-cell lung cancer. Cancer Gene Therapy, 2016, 23: 90–97.

37. Tian F, Li R, Chen Z, Shen Y, Lu J, et al. Differentially Expressed miRNAs in Tumor, Adjacent, and Normal Tissues of Lung Adenocarcinoma. BioMed Research International, 2016, 2016: 1428271.

38. Tafsiri E, Darbouy M, Shadmehr MB, Zagryazhskaya A, Alizadeh J, et al. Expression of miRNAs in non-small-cell lung carcinomas and their association with clinicopathological features. Tumour Biology: The Journal of the International Society for Oncodevelopmental Biology and Medicine, 2015, 36: 1603–1612.

39. Nadal E, Truini A, Nakata A, Lin J, Reddy RM, et al. A Novel Serum 4-microRNA Signature for Lung Cancer Detection. Scientific Reports, 2015, 5: 12464.

40. Qi Z, Yang DY, Cao J. Increased micro-RNA 17, 21, and 192 gene expressions improve early diagnosis in non-small cell lung cancer. Medical Oncology, 2014, 31: 195.

41. Kim MK, Jung SB, Kim JS, Roh MS, Lee JH, et al. Expression of microRNA miR-126 and miR-200c is associated with prognosis in patients with non-small cell lung cancer. Virchows Archiv: An International Journal of Pathology, 2014, 465: 463–471.

42. Rani S, Gately K, Crown J, O'Byrne K, O'Driscoll L. Global analysis of serum microRNAs as potential biomarkers for lung adenocarcinoma. Cancer Biology & Therapy, 2013, 14: 1104–1112.

43. Capodanno A, Boldrini L, Proietti A, Ali G, Pelliccioni S, et al. Let-7g and miR-21 expression in non-small cell lung cancer: correlation with clinicopathological and molecular features. International Journal of Oncology, 2013, 43: 765–774.

44. 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. Medical Oncology, 2012, 29: 618–626.

45. Kosaka N, Iguchi H, Ochiya T. Circulating microRNA in body fluid: a new potential biomarker for cancer diagnosis and prognosis. Cancer Science, 2010, 101: 2087–2092.

46. Pfeffer SR, Yang CH, Pfeffer LM. The Role of miR-21 in Cancer. Drug Development Research, 2015, 76: 270–277.

47. Wang Y, Gao X, Wei F, Zhang X, Yu J, et al. Diagnostic and prognostic value of circulating miR-21 for cancer: a systematic review and meta-analysis. Gene, 2014, 533: 389–397.

48. Zhang JG, Wang JJ, Zhao F, Liu Q, Jiang K, et al. MicroRNA-21 (miR-21) represses tumor suppressor PTEN and promotes growth and invasion in non-small cell lung cancer (NSCLC). Clinica Chimica Acta; International Journal of Clinical Chemistry, 2010, 411: 846–852.

49. Hu S, Yuan Y, Song Z, Yan D, Kong X. Expression Profiles of microRNAs in Drug-Resistant Non-Small Cell Lung Cancer Cell Lines Using microRNA Sequencing. Cellular Physiology and Biochemistry, 2018, 51: 2509–2522.

50. Zhang Z, Wang J, Cheng J, Yu X. Effects of miR-126 on the STAT3 signaling pathway and the regulation of malignant behavior in lung cancer cells. Oncology Letters, 2018, 15: 8412–8416.

51. Grimolizzi F, Monaco F, Leoni F, Bracci M, Staffolani S, et al. Exosomal miR-126 as a circulating biomarker in non-small-cell lung cancer regulating cancer progression. Scientific Reports, 2017, 7: 15277.

52. Si L, Tian H, Yue W, Li L, Li S, et al. Potential use of microRNA-200c as a prognostic marker in non-small cell lung cancer. Oncology Letters, 2017, 14: 4325–4330.

53. Liu PL, Liu WL, Chang JM, Chen YH, Liu YP, et al. MicroRNA-200c inhibits epithelial-mesenchymal transition, invasion, and migration of lung cancer by targeting HMGB1. PLoS One, 2017, 12: e0180844.

54. Lu S, Kong H, Hou Y, Ge D, Huang W, et al. Two plasma microRNA panels for diagnosis and subtype discrimination of lung cancer. Lung Cancer, 2018, 123: 44–51.

55. He J. Knocking down MiR-15a expression promotes the occurrence and development and induces the EMT of NSCLC cells in vitro. Saudi Journal of Biological Sciences, 2017, 24: 1859–1865.

56. Wang H, Zhang Y, Wu Q, Wang YB, Wang W. miR-16 mimics inhibit TGF-beta1-induced epithelial-to-mesenchymal transition via activation of autophagy in non-small cell lung carcinoma cells. Oncology Reports, 2018, 39: 247–254.

57. Zhao K, Cheng J, Chen B, Liu Q, Xu D, et al. Circulating microRNA-34 family low expression correlates with poor prognosis in patients with non-small cell lung cancer. Journal of Thoracic Disease, 2017, 9: 3735–3746.

58. Hetta HF, Zahran AM, Shafik EA, El-Mahdy RI, Mohamed NA, et al. Circulating miRNA-21 and miRNA-23a Expression Signature as Potential Biomarkers for Early Detection of Non-Small-Cell Lung Cancer. Microrna, 2019, 8: 206–215.

59. Zhu K, Ding H, Wang W, Liao Z, Fu Z, et al. Tumor-suppressive miR-218-5p inhibits cancer cell proliferation and migration via EGFR in non-small cell lung cancer. Oncotarget, 2016, 7: 28075–28085.

60. Fassina A, Cappellesso R, Fassan M. Classification of non-small cell lung carcinoma in transthoracic needle specimens using microRNA expression profiling. Chest, 2011, 140: 1305–1311.