Cell-free microRNAs as non-invasive biomarkers in glioma: a diagnostic meta-analysis
Jinfeng Wang, Fengyuan Che, Jinling Zhang First Published April 10, 2019 Research Article
https://doi.org/10.1177/1724600819840033
Article information
Open Access Creative Commons Attribution, Non Commercial 4.0 License
Abstract
Objective:
Since the diagnostic value of microRNAs for detecting glioma is contentious, we aimed to carry out a meta-analysis to synthetically evaluate the diagnostic significance of cell-free microRNAs in cerebrospinal fluid and blood in the detection of glioma.
Methods:
A systematic document retrieval of public databases was performed to obtain eligible studies. Specificity was applied to draw the summary receiver operator characteristic (SROC) curve against sensitivity, and the pooled diagnostic efficiency was assessed by generating the area under the SROC curve. Meta-regression and subgroup analyses were utilized to explore the latent sources of heterogeneity. STATA 12.0, RevMan 5.3 and Meta-DiSc 1.4 were used to conduct all statistical analyses.
Results:
A total of 47 studies from 20 articles comprising 2262 glioma patients and 1986 controls were included in our meta-analysis. Cell-free microRNAs exhibited relatively good diagnostic efficiency in glioma detection, with a sensitivity of 0.83, a specificity of 0.87, and an area under the curve of 0.91. Cell-free miR-21 performed best with pooled area under the curve of 0.88, followed by miR-125 and miR-222. Subgroup analyses and meta-regression indicated that there was substantial heterogeneity existing among the studies, which was in part caused by sample size, World Health Organization grade, reference gene, microRNA origin (extracellular vesicles or non-extracellular vesicle-based-microRNA), microRNA profiling (single- or multiple-microRNA), specimen types, and ethnicity.
Conclusions:
Cell-free microRNAs in cerebrospinal fluid and blood may play an important role as promising non-invasive biomarkers in the early diagnosis of glioma. Further comprehensive forward-looking research is required to validate their clinical significance in glioma diagnosis.
Keywords Cell-free miRNAs, glioma, meta-analysis, non-invasive diagnosis
Introduction
Glioma, as the most common and highly invasive human primary brain tumor, accounts for roughly 30% of all malignancies in the central nervous system.1 Based on histopathologic evaluation, gliomas can be classified into four tumor grades (I–IV), of which grade IV (glioblastoma multiforme) is characterized by lethal aggressiveness as well as extremely poor clinical outcome.2 At present, computed tomography (CT) and magnetic resonance imaging (MRI) are the primary conventional approaches for the diagnosis and surveillance of glioma. Recently, the interest in liquid biopsies for the diagnosis of glioma has been rapidly expanding. Liquid biopsy offers a non-invasive approach by detecting circulating molecular biomarkers, including circulating tumor nucleic acids (DNA, different forms of RNA), circulating protein, circulating tumor cells, and extracellular vesicles (EVs).3
MicroRNAs (miRNAs), as short non-coding RNA molecules, can regulate protein expression post-transcriptionally and function as oncogenes and/or tumor suppressors.4 Recently, numerous studies have verified that miRNAs can stably exist in different types of body fluids (including serum, plasm, cerebrospinal fluid (CSF), urine, tears, bronchial lavage, etc.) and be exploited as non-invasive biomarkers for glioma detection.5–24 However, there have been differentials or inconformity among these studies concerning the reliability and application of miRNAs for the early non-invasive detection of glioma. Therefore, we carried out this meta-analysis to further comprehensively expound the diagnostic value of cell-free miRNAs in early glioma detection based on the former studies.
Materials and methods
Literature retrieval
This meta-analysis was conducted according to guidelines for diagnostic meta-analysis.25 The systematic document retrieval of PubMed, Cochrane Library and Embase for studies published in English up to 20 October 2018 was performed to obtain qualified original articles regarding focusing on the diagnostic value of cell-free miRNAs for glioma. The retrieval strategy was utilized and adopted as follows: (“glioma” OR “neuroglioma” OR “Cerebral glioma” OR “neurospongioma” OR “Malignant glioma” OR “glioblastoma” OR “glioblastoma multiforme”) AND (“microRNA” OR “miRNA” OR “miR”) AND (“Cerebrospinal fluid” OR “blood” OR “serum” OR “plasma” OR “cell-free” OR “circulating”) AND (“diagnosis” OR “sensitivity and specificity” OR “ROC curve”). Also, references lists of all correlative publications were manually searched to obtain additional articles.
Inclusion and exclusion criteria
Eligible publications were appraised independently by two reviewers. In the event of controversy, a third reviewer was consulted and a consensus was reached through multilateral discussion. Studies qualified to be included met the following criteria: (a) studies adopting the gold reference standard to make definite diagnosis of glioma patients; (b) studies regarding the diagnostic performance of cell-free miRNAs in blood and CSF for glioma; and (c) studies providing adequate data for reestablishing two-by-two tables, consisting of true positive (TP), false positive (FP), true negative (TN), and false negative (FN). Exclusion criteria were: (a) review, case reports, letters or seminar articles; (b) studies irrelevant to the diagnostic significance of cell-free miRNAs in blood and CSF for glioma; and (c) studies with replicated data published in other articles.
Data extraction
Two reviewers independently extracted the following data from all qualified publications using standardized forms: (a) basic characteristics of studies, comprising first author, publication year, country of publication, ethnicity, sample size, mean age, gender ratio, cancer type, specimen type, miRNA profiling, methods of miRNAs assay, and reference gene; and (b) diagnostic value, including sensitivity, specificity, TP, FP, FN, TN, and area under the curve (AUC).
Quality evaluation
Each included study was assessed systematically and scored independently by two reviewers in accordance with the revised Quality Assessment of Diagnostic Accuracy Studies 2 (QUADAS-2) criteria.26 This scoring system consists of four basic domains: patient selection, index test, reference standard, as well as flow and timing, utilizing seven questions required to be answered with “yes,” “no,” or “unclear” to evaluate the quality of included articles. An answer of “yes” demonstrates that the risk of bias can be judged low, while “no” or “unclear” was equivalent to high risk or unclear, respectively. Each domain was assessed in terms of the risk of bias, and the first three domains were also evaluated in terms of applicability. Any dispute between the two reviewers was settled by discussion with a third reviewer to arrive at a consensus.
Statistical analysis
The I2 statistic and Q test were undertaken to assess significant heterogeneity among included studies.27 The I2 value⩾ 50% or P value < 0.10 for the Q test implied substantial heterogeneity, and then the random-effects model were employed. Moreover, subgroup analyses and meta-regression were applied to further illuminate potential sources of heterogeneity. A bivariate meta-analysis model was adopted to ascertain pooled sensitivity, specificity, positive likelihood ratio (PLR), negative likelihood ratio (NLR), and diagnostic odds ratio (DOR) of all included publications. The sensitivity and specificity of each included study were used to plot the summary receiver operator characteristic (SROC) curve and to calculate the AUC representing a pooled evaluation indicator of diagnostic efficiency. The Fagan’s nomogram was applied to evaluate the post-diagnostic effect after pooled analysis. Deek’s funnel plot asymmetry test was performed to estimate the publication bias with P< 0.10 indicating any significant difference.28 All analyses were conducted using the STATA 12.0 (StataCorp LP, College Station, TX, US), RevMan 5.3 (Nordic Cochrane Centre, Copenhagen, Denmark) and Meta-DiSc 1.4 (Ramóny Cajal Hospital, Madrid, Spain).
Results
Literature search
The flow diagram of literature retrieval is shown in Supplementary Figure 1. A total of 132 potentially relevant articles were obtained by an initial database search, of which 16 duplicates were eliminated. There were 36 letters, reviews and meta-analysis; 10 articles that were not about humans; and 36 articles irrelevant to our research (about lncRNA, prognosis, therapy, or molecular mechanism, etc.) were further taken out through assessing the abstracts. Then, 34 articles remained for the full-text review, of which 14 were excluded (including 3 unrelated to diagnosis, 10 without sufficient data, and 1 about cell-miRNAs assays). Lastly, 20 qualified articles were involved in this meta-analysis, 3 of which concerned CSF-based miRNAs6,8,20 and 17 involving blood-based ones.5,7,9–19,21–24
Basic characteristics of included studies
The main characteristics of included articles are displayed in Table 1. A total of 47 studies from 20 articles were adopted in our meta-analysis. Six studies were involved in the diagnostic value of cell-free single miR-21 in glioma, 4 referred to single miR-125, and 4 concerned single miR-222. Eight studies assessed the diagnostic value of cell-free miRNAs in low-grade glioma (LGG, World Health Organization (WHO) grade I–II), 18 involved high-grade glioma (HGG, WHO grade III–IV), and 21 involved both LGG and HGG. All studies employed quantitative real-time reverse transcription polymerase chain reaction (qRT-PCR) to detect expression levels of miRNAs, of which 39 studies normalized the miRNA concentration to the reference gene, and 8 normalized the miRNA concentration to serum volume or standard curve. Seventeen studies exploited miRNAs originating from EVs, while 30 investigated miRNAs that were not derived from EVs. Only 11 of these 47 studies applied multiple-miRNA assay in glioma detection, while 36 involved single-miRNA assays. The specimens used by the included articles incorporated CSF, CSF EVs, cisternal CSF, lumbar CSF, serum exosome, serum, and plasma, which can be categorized into two general types: “CSF” and “blood.” Six studies explored the diagnostic value of CSF-based specimens and the other 41 focused on blood-based specimens. Twenty-five studies employed Asian participants and 22 selected Caucasian participants. The publication years of the included articles ranged from 2012 to 2018. The total numbers of glioma patients and controls were 2262 and 1986, respectively. Quality assessment of the included studies according to the QUADAS-2 tool is presented in Supplementary Figure 2. The majority of studies involved in this meta-analysis fulfilled at least four items in QUADAS-2, exhibiting good overall quality of the included studies.
Table
Table 1. Main characteristics of 47 studies included in the meta-analysis.
Table 1. Main characteristics of 47 studies included in the meta-analysis.
View larger version
Diagnostic efficacy of cell-free miRNAs for glioma
Considering that significant heterogeneity among studies exist in sensitivity and specificity data (I2= 81.96% and I2= 82.60%, respectively) (P< 0.01), the random-effects model was applied accordingly. As shown in Table 2, the pooled parameters calculated from all 47 studies were as follows: sensitivity, 0.83 (95% CI (confidence interval): 0.79, 0.87); specificity, 0.87 (95% CI 0.83, 0.91); PLR, 6.5 (95% CI 4.7, 9.1); NLR, 0.19 (95% CI 0.15, 0.25); DOR, 34 (95% CI 20, 56) and AUC, 0.91 (95 % CI 0.89, 0.94) (Figure 1(a)), manifesting that cell-free miRNAs in blood and CSF may be utilized as a good indicator of glioma diagnosis with a relatively high accuracy. As displayed in the Fagan’s plot (Figure 1(b)), the pre-test probability was 53%, the post-test probability of glioma for a positive test result was 88%, while a negative test result was 19%, signifying that both the likelihood ratios and post-test probabilities were moderate. The positive likelihood ratio of 7 indicated that a person with glioma is seven times more likely to have a positive test result than a healthy person. Also, the diagnostic odds ratio value was 34 (95% CI 20, 56), demonstrating that cell-free miRNAs in blood and CSF can be eligible for differentiating glioma patients from controls. Furthermore, cell-free single miR-21 demonstrated the best diagnostic performance with pooled AUC of 0.88, followed by miR-125 with AUC of 0.86, and miR-222 with AUC of 0.85 (Table 2 and Figure 2). Supplementary Figure 3 showed the weight and sensitivity of each study with the pooled sensitivity of 0.83 (95% CI 0.79, 0.87) (P< 0.001).
Table
Table 2. Summary estimates of diagnostic criteria for miRNAs profiling in glioma detection.
Table 2. Summary estimates of diagnostic criteria for miRNAs profiling in glioma detection.
View larger version
figure
Figure 1. Diagnostic accuracy of miRNAs for glioma. (a) SROC curve with pooled estimates of sensitivity, specificity and AUC of overall studies. (b) Fagan’s Nomogram for assessment of post-test probabilities based on pooled estimates of PLR and NLR of overall studies.
AUC: area under the curve; miRNA: microRNA; NLR: negative likelihood ratio; PLR: positive likelihood ratio; SROC: summary receiver operating characteristic.
figure
Figure 2. SROC curves based on diagnostic studies of (a) miR-21, (b) miR-125 and (c) miR-222.
SROC: summary receiver operating characteristic.
Subgroup analysis
Subgroup analyses based on WHO grade, reference gene, miRNA origin (EVs- or non-EVs based-miRNA), miRNA profiling (single- or multiple-miRNA), specimen types, and ethnicity were carried out respectively. The pooled results for diagnostic implication in different subgroups are presented in Table 2. A subgroup analysis based on WHO grade suggested that pooled AUC of cell-free miRNAs in LGG, HGG, and glioma across all grades was 0.81, 0.88, and 0.95, respectively. Cell-free miRNAs normalized to the reference gene gave rise to a pooled AUC of 0.90, while that normalized to serum volume or standard curve gave rise to a pooled AUC of 0.97. miRNAs originating from EVs and non-EVs manifested a pooled AUC of 0.84 and 0.94, respectively. Diagnostic properties of multiple-miRNA assays versus single ones were 0.87 versus 0.82 for sensitivity, 0.93 versus 0.85 for specificity, and 0.95 versus 0.90 for AUC. The diagnostic accuracy of CSF-based assays versus blood-based ones was 0.96 versus 0.91 for the pooled AUC. Moreover, Asian-based miRNA assays exhibited good diagnostic performance with a pooled AUC of 0.93, while Caucasian-based ones had a pooled AUC of 0.88.
Meta-regression and publication bias
Meta-regression analysis was conducted to further probe into potential sources of the heterogeneity and to validate the results of subgroup analyses. As depicted in Figure 3(a), sample size, WHO grade, reference gene, miRNA origin, miRNA profiling, types of specimen, and ethnicity may be the major sources of heterogeneity for miRNAs assay in glioma.
figure
Figure 3. Meta-regression and publication bias based on overall studies. (a) Forest plots of multivariable meta-regression analyses for sensitivity and specificity (vertical lines indicate pooled estimates of sensitivity and specificity respectively). (b) Deeks’ funnel plot asymmetry test.
Deeks’ funnel plot asymmetry test was performed to assess publication bias. As presented in Figure 3(b), the P-value of 0.10 suggested no significant publication bias was observed in our meta-analysis.
Discussion
CT and MRI are currently the most commonly used tools for diagnosing and surveilling glioma. Even with their current limitations, circulating biomarkers have shown tremendous clinical and research potential in glioma.29 It seems likely that these biomarkers may be complementary with each able to provide unique information about tumor burden, genetics, and biology.29 Recently, increasing studies have focused on exploring the potential of blood-based and CSF-based cell-free miRNAs as biomarkers of glioma.5–24 In view of inconsistencies and differentials existing among these studies, we conducted this meta-analysis to investigate the feasibility of miRNAs as non-invasive diagnostic biomarkers in glioma.
Thus far, our report is among the few evidence-based meta-analyses to verify the diagnostic significance of cell-free miRNAs in the early identification of glioma with a pooled AUC value of 0.91 in distinguishing glioma patients from controls (pooled sensitivity = 83%; pooled specificity = 87%), implying the potential diagnostic implication of cell-free miRNAs as non-invasive biomarkers. Furthermore, the DOR was 34 (95% CI 20, 56), suggesting that cell-free miRNA test-positive patients have a 34-times higher chance of glioma than controls.
As there were six studies involving diagnostic significance of cell-free single miR-21 in glioma—four referring to single miR-125 and four concerning single miR-222—we performed independent meta-analyses of these three miRNAs. Cell-free single miR-21 performed best with a pooled AUC of 0.88, sensitivity of 0.84, specificity of 0.87, PLR of 6.6, NLR of 0.18, and DOR of 36, manifesting relatively high diagnostic efficacy. Moreover, miR-21 even exhibited a relatively higher value of sensitivity, PLR, NLR, and DOR compared to overall miRNAs and a subgroup of a single-miRNA assay. Similarly, single miR-125 and single miR-222 also displayed satisfactory diagnostic accuracy with AUCs of 0.86 and 0.85, respectively. These results suggest that a single cell-free miR-21, miR-125, and miR-222 have the potential to be forceful non-invasive biomarkers for diagnosing glioma.
Next, we conducted subgroup analyses and meta-regression analyses to probe into potential sources of the heterogeneity. Our results indicated that pooled sensitivity was affected by WHO grade, reference gene, miRNA origin, miRNA profiling, specimen types, and ethnicity, whereas pooled specificity was influenced only by WHO grade, reference gene, miRNA origin, and ethnicity. This suggests that the above factors may be the major sources of heterogeneity for miRNA assays in glioma. First, the WHO grade of glioma seemed to have an impact on both the pooled sensitivity and specificity of cell-free miRNAs assays, which maybe because LGG and HGG are characterized by different somatic mutations, gene expression profiles, and survival rates of the patients.30 Second, the reference gene also contributed to heterogeneity for miRNA assays. There have been no acknowledged endogenous reference genes so far, which greatly limits the clinical application of accurate quantification of miRNAs in body fluids. Moreover, miRNAs originating from EVs manifested a pooled AUC of 0.84. EVs are small lipid membrane-bound satchels expelled by cells to mediate cell-to-cell communication and contain soluble materials, such as nucleic acids, lipids, and proteins, which are hence protected from degradation in body fluids.31 Exosomes are small EVs (40–150 nm) with a multivesicular endosomal origin secreted by both normal and neoplastic cells.32 Furthermore, the diagnostic property of multiple-miRNA assays was 0.95 for AUC, indicating the strength of exploiting panels of miRNAs to obtain the whole picture. As demonstrated in Table 1, miR-15b and miR-21 from CSF,6 7-miRNA panel from serum,7 miR-15b and miR-21 from plasma,10 9-miRNA panel from serum,15 9-miRNA panel from CSF,20 as well as miR-21, miR-222, and miR-124-3p from serum exosome24 performed well in the diagnosis of glioma, which may be because these panels of miRNAs have common mRNA targets or participate in the common pathways such as Hedgehog, Notch, Wnt, EGFR, TGFβ and HIF1α, etc.33 For example, ANKS1B, COX15, CREBRF, GABRB2, PDHA1, PLAG1, and UBR3 are common mRNA targets for the panel of miR-15b and miR-21 predicted by miRDB.34 The molecular mechanism underlying the limitation of a single-miRNA biomarker may be that aberrant levels of single miRNA might be associated with several different types of cancers.35 More importantly, the development of cancer can be perceived as the result of a complex multi-stage process of epigenetic and genomic abnormalities, and hence should be targeted by multiple miRNAs.36 Therefore, it is sensible to employ panels of miRNAs for the purpose of evading the limitations in applying miRNAs as non-invasive biomarkers in cancer identification, especially when it is hard to conduct a regular biopsy subjected to localized pathological conditions. In addition, diagnostic accuracy of CSF-based assays was 0.96 for the pooled AUC. However, as there were only six included studies involving in CSF-based miRNAs, it is best to treat the result cautiously. Last but not least, ethnicity seemed to be another potential source of the heterogeneity, with a pooled AUC of 0.93 for Asian-based miRNA assays, and a pooled AUC of 0.88 for Caucasian-based ones. This might be attributed to different hereditary factors and living backgrounds between the two populations. The origin of ethnicity-related discrepancy is still unclear and might be ascribed to unknown mechanisms. Therefore, large scale investigations should be carried out to further verify whether these ethnicity-related discrepancies truly exist.
In a meta-analysis from 2018, Zhou et al.37 systematically assessed the potential diagnostic value of miRNAs for glioma based on 28 articles (23 in blood, 3 in CSF, and 2 in frozen brain tissue) including 2528 glioma patients and 2563 controls. The results showed that pooled sensitivity, specificity, and AUC of overall miRNAs were 0.85, 0.90, and 0.93, respectively. However, our study aimed to evaluate the diagnostic significance of cell-free miRNAs from serum, plasma, and CSF, which differed from Zhou et al’s37 meta-analysis. Moreover, they did not conduct subgroup analyses of WHO grade, reference gene, and miRNA origin (EVs- vs. non-EVs based-miRNA), which was less comprehensive than our study.
Although we did our best to perform an accurate analysis, the following limitations existed and should be noted. Above all, there were no data published regarding African populations. Moreover, it is vital to identify and exploit miRNA profiles capable of distinguishing cancer from other diseases with similar symptoms. Nonetheless, the majority of studies included in this meta-analysis merely sought to differentiate glioma patients from healthy controls, and did not involve diseases with similar symptoms. In addition, there was a low amount of miRNAs in body fluids, and almost no acknowledged endogenous reference genes, which limits the clinical application of accurate quantification of miRNAs in body fluids. It is difficult for cross-comparison among studies performed by different laboratories due to lack of convention for accurate quantification of miRNAs in body fluids; thus, a normalized procedure should be set up and preferably be followed across all studies to minimize procedure-based bias. Lastly, statistical heterogeneity existed in our analysis due to sample size, WHO grade, reference gene, miRNA origin, miRNA profiling, specimen types, and ethnicity, which may have had an impact on the results.
Declaration of conflicting interest
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/ or publication of this article: This work was supported by The Key Research and Development Program of Linyi City (grant number 2016ZK001) and Medicine and health science technology development plan of Shandong Province (grant number 2017WS323).
Supplemental material
Supplemental material for this article is available online.
ORCID iD
Jinfeng Wang https://orcid.org/0000-0002-4454-8997
References
1. Clarke, J, Butowski, N, Chang, S. Recent advances in therapy for glioblastoma. Arch Neurol 2010; 67: 279–283.
Google Scholar | Medline
2. Brandes, AA, Tosoni, A, Franceschi, E. Recurrence pattern after temozolomide concomitant with and adjuvant to radiotherapy in newly diagnosed patients with glioblastoma: correlation With MGMT promoter methylation status. J Clin Oncol 2009; 27: 1275–1279.
Google Scholar | Medline | ISI
3. Westphal, M, Lamszus, K. Circulating biomarkers for gliomas. Nat Rev Neurol 2015; 11: 556–566.
Google Scholar | Medline
4. Kong, YW, Ferland-McCollough, D, Jackson, TJ. microRNAs in cancer management. Lancet Oncol 2012; 13: e249–e258.
Google Scholar | Medline | ISI
5. Wang, Q, Li, P, Li, A. Plasma specific miRNAs as predictive biomarkers for diagnosis and prognosis of glioma. J Exp Clin Cancer Res 2012; 31: 97.
Google Scholar | Medline
6. Baraniskin, A, Kuhnhenn, J, Schlegel, U. Identification of microRNAs in the cerebrospinal fluid as biomarker for the diagnosis of glioma. Neuro Oncol 2012; 14: 29–33.
Google Scholar | Medline
7. Yang, C, Wang, C, Chen, X. Identification of seven serum microRNAs from a genome-wide serum microRNA expression profile as potential noninvasive biomarkers for malignant astrocytomas. Int J Cancer 2013; 132: 116–127.
Google Scholar | Medline
8. Akers, JC, Ramakrishnan, V, Kim, R. MiR-21 in the extracellular vesicles (EVs) of cerebrospinal fluid (CSF): a platform for glioblastoma biomarker development. PLoS One 2013; 8: e78115.
Google Scholar | Medline
9. Manterola, L, Guruceaga, E, Gallego Perez-Larraya, J. A small noncoding RNA signature found in exosomes of GBM patient serum as a diagnostic tool. Neuro Oncol 2014; 16: 520–527.
Google Scholar | Medline
10. Ivo D’Urso, P, Fernando D’Urso, O, Damiano Gianfreda, C. miR-15b and miR-21 as Circulating Biomarkers for Diagnosis of Glioma. Curr Genomics 2015; 16: 304–311.
Google Scholar | Medline
11. Sun, J, Liao, K, Wu, X. Serum microRNA-128 as a biomarker for diagnosis of glioma. Int J Clin Exp Med 2015; 8: 456–463.
Google Scholar | Medline
12. Lai, NS, Wu, DG, Fang, XG. Serum microRNA-210 as a potential noninvasive biomarker for the diagnosis and prognosis of glioma. Br J Cancer 2015; 112: 1241–1246.
Google Scholar | Medline | ISI
13. Wu, J, Li, L, Jiang, C. Identification and evaluation of serum MicroRNA-29 family for glioma screening. Mol Neurobiol 2015; 52: 1540–1546.
Google Scholar | Medline
14. Shao, N, Wang, L, Xue, L. Plasma miR-454-3p as a potential prognostic indicator in human glioma. Neurol Sci 2015; 36: 309–313.
Google Scholar | Medline
15. Zhi, F, Shao, N, Wang, R. Identification of 9 serum microRNAs as potential noninvasive biomarkers of human astrocytoma. Neuro Oncol 2015; 17: 383–391.
Google Scholar | Medline
16. Xiao, Y, Zhang, L, Song, Z. Potential Diagnostic and Prognostic Value of Plasma Circulating MicroRNA-182 in Human Glioma. Med Sci Monit 2016; 22: 855–862.
Google Scholar | Medline
17. Yue, X, Lan, F, Hu, M. Downregulation of serum microRNA-205 as a potential diagnostic and prognostic biomarker for human glioma. J Neurosurg 2016; 124: 122–128.
Google Scholar | Medline
18. Zhang, R, Pang, B, Xin, T. Plasma miR-221/222 family as novel descriptive and prognostic biomarkers for glioma. Mol Neurobiol 2016; 53: 1452–1460.
Google Scholar | Medline
19. Wei, X, Chen, D, Lv, T. Serum MicroRNA-125b as a Potential Biomarker for Glioma Diagnosis. Mol Neurobiol 2016; 53: 163–170.
Google Scholar | Medline
20. Akers, JC, Hua, W, Li, H. A cerebrospinal fluid microRNA signature as biomarker for glioblastoma. Oncotarget 2017; 8: 68769–68779.
Google Scholar | Medline
21. Tang, Y, Zhao, S, Wang, J. Plasma miR-122 as a potential diagnostic and prognostic indicator in human glioma. Neurol Sci 2017; 38: 1087–1092.
Google Scholar | Medline
22. Huang, Q, Wang, C, Hou, Z. Serum microRNA-376 family as diagnostic and prognostic markers in human gliomas. Cancer Biomark 2017; 19: 137–144.
Google Scholar | Medline
23. Lan, F, Qing, Q, Pan, Q. Serum exosomal miR-301a as a potential diagnostic and prognostic biomarker for human glioma. Cell Oncol (Dordr) 2018; 41: 25–33.
Google Scholar | Medline
24. Santangelo, A, Imbruce, P, Gardenghi, B. A microRNA signature from serum exosomes of patients with glioma as complementary diagnostic biomarker. J Neurooncol 2018; 136: 51–62.
Google Scholar | Medline
25. Leeflang, MM, Deeks, JJ, Gatsonis, C. Systematic reviews of diagnostic test accuracy. Ann Intern Med 2008; 149: 889–897.
Google Scholar | Medline | ISI
26. Whiting, PF, Rutjes, AW, Westwood, ME. QUADAS-2: a revised tool for the quality assessment of diagnostic accuracy studies. Ann Intern Med 2011; 155: 529–536.
Google Scholar | Medline | ISI
27. Higgins, JP, Thompson, SG, Deeks, JJ. Measuring inconsistency in meta-analyses. Bmj 2003; 327: 557–560.
Google Scholar | Medline
28. Deeks, JJ, Macaskill, P, Irwig, L. The performance of tests of publication bias and other sample size effects in systematic reviews of diagnostic test accuracy was assessed. J Clin Epidemiol 2005; 58: 882–893.
Google Scholar | Medline | ISI
29. Wang, J, Bettegowda, C. Applications of DNA-based liquid biopsy for central nervous system neoplasms. J Mol Diagn 2017; 19: 24–34.
Google Scholar | Medline
30. Li, Y, Wang, D, Wang, L. Distinct genomic aberrations between low-grade and high-grade gliomas of Chinese patients. PLoS One 2013; 8: e57168.
Google Scholar | Medline
31. Kosaka, N, Yoshioka, Y, Fujita, Y. Versatile roles of extracellular vesicles in cancer. J Clin Invest 2016; 126: 1163–1172.
Google Scholar | Medline
32. Kalluri, R. The biology and function of exosomes in cancer. J Clin Invest 2016; 126: 1208–1215.
Google Scholar | Medline | ISI
33. Kit, OI, Vodolazhsky, DI, Rostorguev, EE. The role of microRNA in regulation of signaling pathways in gliomas. Biochem Moscow Suppl Ser B 2018; 12: 1–21.
Google Scholar
34. Wong, N, Wang, X. miRDB: an online resource for microRNA target prediction and functional annotations. Nucleic Acids Res 2015; 43: D146–152.
Google Scholar | Medline | ISI
35. Sita-Lumsden, A, Dart, DA, Waxman, J. Circulating microRNAs as potential new biomarkers for prostate cancer. Br J Cancer 2013; 108: 1925–1930.
Google Scholar | Medline | ISI
36. Zen, K, Zhang, CY. Circulating microRNAs: a novel class of biomarkers to diagnose and monitor human cancers. Med Res Rev 2012; 32: 326–348.
Google Scholar | Medline | ISI
37. Zhou, Q, Liu, J, Quan, J. MicroRNAs as potential biomarkers for the diagnosis of glioma: A systematic review and meta-analysis. Cancer Sci 2018; 109: 2651–2659.
Google Scholar | Medline
View Abstract
Jinfeng Wang, Fengyuan Che, Jinling Zhang First Published April 10, 2019 Research Article
https://doi.org/10.1177/1724600819840033
Article information
Open Access Creative Commons Attribution, Non Commercial 4.0 License
Abstract
Objective:
Since the diagnostic value of microRNAs for detecting glioma is contentious, we aimed to carry out a meta-analysis to synthetically evaluate the diagnostic significance of cell-free microRNAs in cerebrospinal fluid and blood in the detection of glioma.
Methods:
A systematic document retrieval of public databases was performed to obtain eligible studies. Specificity was applied to draw the summary receiver operator characteristic (SROC) curve against sensitivity, and the pooled diagnostic efficiency was assessed by generating the area under the SROC curve. Meta-regression and subgroup analyses were utilized to explore the latent sources of heterogeneity. STATA 12.0, RevMan 5.3 and Meta-DiSc 1.4 were used to conduct all statistical analyses.
Results:
A total of 47 studies from 20 articles comprising 2262 glioma patients and 1986 controls were included in our meta-analysis. Cell-free microRNAs exhibited relatively good diagnostic efficiency in glioma detection, with a sensitivity of 0.83, a specificity of 0.87, and an area under the curve of 0.91. Cell-free miR-21 performed best with pooled area under the curve of 0.88, followed by miR-125 and miR-222. Subgroup analyses and meta-regression indicated that there was substantial heterogeneity existing among the studies, which was in part caused by sample size, World Health Organization grade, reference gene, microRNA origin (extracellular vesicles or non-extracellular vesicle-based-microRNA), microRNA profiling (single- or multiple-microRNA), specimen types, and ethnicity.
Conclusions:
Cell-free microRNAs in cerebrospinal fluid and blood may play an important role as promising non-invasive biomarkers in the early diagnosis of glioma. Further comprehensive forward-looking research is required to validate their clinical significance in glioma diagnosis.
Keywords Cell-free miRNAs, glioma, meta-analysis, non-invasive diagnosis
Introduction
Glioma, as the most common and highly invasive human primary brain tumor, accounts for roughly 30% of all malignancies in the central nervous system.1 Based on histopathologic evaluation, gliomas can be classified into four tumor grades (I–IV), of which grade IV (glioblastoma multiforme) is characterized by lethal aggressiveness as well as extremely poor clinical outcome.2 At present, computed tomography (CT) and magnetic resonance imaging (MRI) are the primary conventional approaches for the diagnosis and surveillance of glioma. Recently, the interest in liquid biopsies for the diagnosis of glioma has been rapidly expanding. Liquid biopsy offers a non-invasive approach by detecting circulating molecular biomarkers, including circulating tumor nucleic acids (DNA, different forms of RNA), circulating protein, circulating tumor cells, and extracellular vesicles (EVs).3
MicroRNAs (miRNAs), as short non-coding RNA molecules, can regulate protein expression post-transcriptionally and function as oncogenes and/or tumor suppressors.4 Recently, numerous studies have verified that miRNAs can stably exist in different types of body fluids (including serum, plasm, cerebrospinal fluid (CSF), urine, tears, bronchial lavage, etc.) and be exploited as non-invasive biomarkers for glioma detection.5–24 However, there have been differentials or inconformity among these studies concerning the reliability and application of miRNAs for the early non-invasive detection of glioma. Therefore, we carried out this meta-analysis to further comprehensively expound the diagnostic value of cell-free miRNAs in early glioma detection based on the former studies.
Materials and methods
Literature retrieval
This meta-analysis was conducted according to guidelines for diagnostic meta-analysis.25 The systematic document retrieval of PubMed, Cochrane Library and Embase for studies published in English up to 20 October 2018 was performed to obtain qualified original articles regarding focusing on the diagnostic value of cell-free miRNAs for glioma. The retrieval strategy was utilized and adopted as follows: (“glioma” OR “neuroglioma” OR “Cerebral glioma” OR “neurospongioma” OR “Malignant glioma” OR “glioblastoma” OR “glioblastoma multiforme”) AND (“microRNA” OR “miRNA” OR “miR”) AND (“Cerebrospinal fluid” OR “blood” OR “serum” OR “plasma” OR “cell-free” OR “circulating”) AND (“diagnosis” OR “sensitivity and specificity” OR “ROC curve”). Also, references lists of all correlative publications were manually searched to obtain additional articles.
Inclusion and exclusion criteria
Eligible publications were appraised independently by two reviewers. In the event of controversy, a third reviewer was consulted and a consensus was reached through multilateral discussion. Studies qualified to be included met the following criteria: (a) studies adopting the gold reference standard to make definite diagnosis of glioma patients; (b) studies regarding the diagnostic performance of cell-free miRNAs in blood and CSF for glioma; and (c) studies providing adequate data for reestablishing two-by-two tables, consisting of true positive (TP), false positive (FP), true negative (TN), and false negative (FN). Exclusion criteria were: (a) review, case reports, letters or seminar articles; (b) studies irrelevant to the diagnostic significance of cell-free miRNAs in blood and CSF for glioma; and (c) studies with replicated data published in other articles.
Data extraction
Two reviewers independently extracted the following data from all qualified publications using standardized forms: (a) basic characteristics of studies, comprising first author, publication year, country of publication, ethnicity, sample size, mean age, gender ratio, cancer type, specimen type, miRNA profiling, methods of miRNAs assay, and reference gene; and (b) diagnostic value, including sensitivity, specificity, TP, FP, FN, TN, and area under the curve (AUC).
Quality evaluation
Each included study was assessed systematically and scored independently by two reviewers in accordance with the revised Quality Assessment of Diagnostic Accuracy Studies 2 (QUADAS-2) criteria.26 This scoring system consists of four basic domains: patient selection, index test, reference standard, as well as flow and timing, utilizing seven questions required to be answered with “yes,” “no,” or “unclear” to evaluate the quality of included articles. An answer of “yes” demonstrates that the risk of bias can be judged low, while “no” or “unclear” was equivalent to high risk or unclear, respectively. Each domain was assessed in terms of the risk of bias, and the first three domains were also evaluated in terms of applicability. Any dispute between the two reviewers was settled by discussion with a third reviewer to arrive at a consensus.
Statistical analysis
The I2 statistic and Q test were undertaken to assess significant heterogeneity among included studies.27 The I2 value⩾ 50% or P value < 0.10 for the Q test implied substantial heterogeneity, and then the random-effects model were employed. Moreover, subgroup analyses and meta-regression were applied to further illuminate potential sources of heterogeneity. A bivariate meta-analysis model was adopted to ascertain pooled sensitivity, specificity, positive likelihood ratio (PLR), negative likelihood ratio (NLR), and diagnostic odds ratio (DOR) of all included publications. The sensitivity and specificity of each included study were used to plot the summary receiver operator characteristic (SROC) curve and to calculate the AUC representing a pooled evaluation indicator of diagnostic efficiency. The Fagan’s nomogram was applied to evaluate the post-diagnostic effect after pooled analysis. Deek’s funnel plot asymmetry test was performed to estimate the publication bias with P< 0.10 indicating any significant difference.28 All analyses were conducted using the STATA 12.0 (StataCorp LP, College Station, TX, US), RevMan 5.3 (Nordic Cochrane Centre, Copenhagen, Denmark) and Meta-DiSc 1.4 (Ramóny Cajal Hospital, Madrid, Spain).
Results
Literature search
The flow diagram of literature retrieval is shown in Supplementary Figure 1. A total of 132 potentially relevant articles were obtained by an initial database search, of which 16 duplicates were eliminated. There were 36 letters, reviews and meta-analysis; 10 articles that were not about humans; and 36 articles irrelevant to our research (about lncRNA, prognosis, therapy, or molecular mechanism, etc.) were further taken out through assessing the abstracts. Then, 34 articles remained for the full-text review, of which 14 were excluded (including 3 unrelated to diagnosis, 10 without sufficient data, and 1 about cell-miRNAs assays). Lastly, 20 qualified articles were involved in this meta-analysis, 3 of which concerned CSF-based miRNAs6,8,20 and 17 involving blood-based ones.5,7,9–19,21–24
Basic characteristics of included studies
The main characteristics of included articles are displayed in Table 1. A total of 47 studies from 20 articles were adopted in our meta-analysis. Six studies were involved in the diagnostic value of cell-free single miR-21 in glioma, 4 referred to single miR-125, and 4 concerned single miR-222. Eight studies assessed the diagnostic value of cell-free miRNAs in low-grade glioma (LGG, World Health Organization (WHO) grade I–II), 18 involved high-grade glioma (HGG, WHO grade III–IV), and 21 involved both LGG and HGG. All studies employed quantitative real-time reverse transcription polymerase chain reaction (qRT-PCR) to detect expression levels of miRNAs, of which 39 studies normalized the miRNA concentration to the reference gene, and 8 normalized the miRNA concentration to serum volume or standard curve. Seventeen studies exploited miRNAs originating from EVs, while 30 investigated miRNAs that were not derived from EVs. Only 11 of these 47 studies applied multiple-miRNA assay in glioma detection, while 36 involved single-miRNA assays. The specimens used by the included articles incorporated CSF, CSF EVs, cisternal CSF, lumbar CSF, serum exosome, serum, and plasma, which can be categorized into two general types: “CSF” and “blood.” Six studies explored the diagnostic value of CSF-based specimens and the other 41 focused on blood-based specimens. Twenty-five studies employed Asian participants and 22 selected Caucasian participants. The publication years of the included articles ranged from 2012 to 2018. The total numbers of glioma patients and controls were 2262 and 1986, respectively. Quality assessment of the included studies according to the QUADAS-2 tool is presented in Supplementary Figure 2. The majority of studies involved in this meta-analysis fulfilled at least four items in QUADAS-2, exhibiting good overall quality of the included studies.
Table
Table 1. Main characteristics of 47 studies included in the meta-analysis.
Table 1. Main characteristics of 47 studies included in the meta-analysis.
View larger version
Diagnostic efficacy of cell-free miRNAs for glioma
Considering that significant heterogeneity among studies exist in sensitivity and specificity data (I2= 81.96% and I2= 82.60%, respectively) (P< 0.01), the random-effects model was applied accordingly. As shown in Table 2, the pooled parameters calculated from all 47 studies were as follows: sensitivity, 0.83 (95% CI (confidence interval): 0.79, 0.87); specificity, 0.87 (95% CI 0.83, 0.91); PLR, 6.5 (95% CI 4.7, 9.1); NLR, 0.19 (95% CI 0.15, 0.25); DOR, 34 (95% CI 20, 56) and AUC, 0.91 (95 % CI 0.89, 0.94) (Figure 1(a)), manifesting that cell-free miRNAs in blood and CSF may be utilized as a good indicator of glioma diagnosis with a relatively high accuracy. As displayed in the Fagan’s plot (Figure 1(b)), the pre-test probability was 53%, the post-test probability of glioma for a positive test result was 88%, while a negative test result was 19%, signifying that both the likelihood ratios and post-test probabilities were moderate. The positive likelihood ratio of 7 indicated that a person with glioma is seven times more likely to have a positive test result than a healthy person. Also, the diagnostic odds ratio value was 34 (95% CI 20, 56), demonstrating that cell-free miRNAs in blood and CSF can be eligible for differentiating glioma patients from controls. Furthermore, cell-free single miR-21 demonstrated the best diagnostic performance with pooled AUC of 0.88, followed by miR-125 with AUC of 0.86, and miR-222 with AUC of 0.85 (Table 2 and Figure 2). Supplementary Figure 3 showed the weight and sensitivity of each study with the pooled sensitivity of 0.83 (95% CI 0.79, 0.87) (P< 0.001).
Table
Table 2. Summary estimates of diagnostic criteria for miRNAs profiling in glioma detection.
Table 2. Summary estimates of diagnostic criteria for miRNAs profiling in glioma detection.
View larger version
figure
Figure 1. Diagnostic accuracy of miRNAs for glioma. (a) SROC curve with pooled estimates of sensitivity, specificity and AUC of overall studies. (b) Fagan’s Nomogram for assessment of post-test probabilities based on pooled estimates of PLR and NLR of overall studies.
AUC: area under the curve; miRNA: microRNA; NLR: negative likelihood ratio; PLR: positive likelihood ratio; SROC: summary receiver operating characteristic.
figure
Figure 2. SROC curves based on diagnostic studies of (a) miR-21, (b) miR-125 and (c) miR-222.
SROC: summary receiver operating characteristic.
Subgroup analysis
Subgroup analyses based on WHO grade, reference gene, miRNA origin (EVs- or non-EVs based-miRNA), miRNA profiling (single- or multiple-miRNA), specimen types, and ethnicity were carried out respectively. The pooled results for diagnostic implication in different subgroups are presented in Table 2. A subgroup analysis based on WHO grade suggested that pooled AUC of cell-free miRNAs in LGG, HGG, and glioma across all grades was 0.81, 0.88, and 0.95, respectively. Cell-free miRNAs normalized to the reference gene gave rise to a pooled AUC of 0.90, while that normalized to serum volume or standard curve gave rise to a pooled AUC of 0.97. miRNAs originating from EVs and non-EVs manifested a pooled AUC of 0.84 and 0.94, respectively. Diagnostic properties of multiple-miRNA assays versus single ones were 0.87 versus 0.82 for sensitivity, 0.93 versus 0.85 for specificity, and 0.95 versus 0.90 for AUC. The diagnostic accuracy of CSF-based assays versus blood-based ones was 0.96 versus 0.91 for the pooled AUC. Moreover, Asian-based miRNA assays exhibited good diagnostic performance with a pooled AUC of 0.93, while Caucasian-based ones had a pooled AUC of 0.88.
Meta-regression and publication bias
Meta-regression analysis was conducted to further probe into potential sources of the heterogeneity and to validate the results of subgroup analyses. As depicted in Figure 3(a), sample size, WHO grade, reference gene, miRNA origin, miRNA profiling, types of specimen, and ethnicity may be the major sources of heterogeneity for miRNAs assay in glioma.
figure
Figure 3. Meta-regression and publication bias based on overall studies. (a) Forest plots of multivariable meta-regression analyses for sensitivity and specificity (vertical lines indicate pooled estimates of sensitivity and specificity respectively). (b) Deeks’ funnel plot asymmetry test.
Deeks’ funnel plot asymmetry test was performed to assess publication bias. As presented in Figure 3(b), the P-value of 0.10 suggested no significant publication bias was observed in our meta-analysis.
Discussion
CT and MRI are currently the most commonly used tools for diagnosing and surveilling glioma. Even with their current limitations, circulating biomarkers have shown tremendous clinical and research potential in glioma.29 It seems likely that these biomarkers may be complementary with each able to provide unique information about tumor burden, genetics, and biology.29 Recently, increasing studies have focused on exploring the potential of blood-based and CSF-based cell-free miRNAs as biomarkers of glioma.5–24 In view of inconsistencies and differentials existing among these studies, we conducted this meta-analysis to investigate the feasibility of miRNAs as non-invasive diagnostic biomarkers in glioma.
Thus far, our report is among the few evidence-based meta-analyses to verify the diagnostic significance of cell-free miRNAs in the early identification of glioma with a pooled AUC value of 0.91 in distinguishing glioma patients from controls (pooled sensitivity = 83%; pooled specificity = 87%), implying the potential diagnostic implication of cell-free miRNAs as non-invasive biomarkers. Furthermore, the DOR was 34 (95% CI 20, 56), suggesting that cell-free miRNA test-positive patients have a 34-times higher chance of glioma than controls.
As there were six studies involving diagnostic significance of cell-free single miR-21 in glioma—four referring to single miR-125 and four concerning single miR-222—we performed independent meta-analyses of these three miRNAs. Cell-free single miR-21 performed best with a pooled AUC of 0.88, sensitivity of 0.84, specificity of 0.87, PLR of 6.6, NLR of 0.18, and DOR of 36, manifesting relatively high diagnostic efficacy. Moreover, miR-21 even exhibited a relatively higher value of sensitivity, PLR, NLR, and DOR compared to overall miRNAs and a subgroup of a single-miRNA assay. Similarly, single miR-125 and single miR-222 also displayed satisfactory diagnostic accuracy with AUCs of 0.86 and 0.85, respectively. These results suggest that a single cell-free miR-21, miR-125, and miR-222 have the potential to be forceful non-invasive biomarkers for diagnosing glioma.
Next, we conducted subgroup analyses and meta-regression analyses to probe into potential sources of the heterogeneity. Our results indicated that pooled sensitivity was affected by WHO grade, reference gene, miRNA origin, miRNA profiling, specimen types, and ethnicity, whereas pooled specificity was influenced only by WHO grade, reference gene, miRNA origin, and ethnicity. This suggests that the above factors may be the major sources of heterogeneity for miRNA assays in glioma. First, the WHO grade of glioma seemed to have an impact on both the pooled sensitivity and specificity of cell-free miRNAs assays, which maybe because LGG and HGG are characterized by different somatic mutations, gene expression profiles, and survival rates of the patients.30 Second, the reference gene also contributed to heterogeneity for miRNA assays. There have been no acknowledged endogenous reference genes so far, which greatly limits the clinical application of accurate quantification of miRNAs in body fluids. Moreover, miRNAs originating from EVs manifested a pooled AUC of 0.84. EVs are small lipid membrane-bound satchels expelled by cells to mediate cell-to-cell communication and contain soluble materials, such as nucleic acids, lipids, and proteins, which are hence protected from degradation in body fluids.31 Exosomes are small EVs (40–150 nm) with a multivesicular endosomal origin secreted by both normal and neoplastic cells.32 Furthermore, the diagnostic property of multiple-miRNA assays was 0.95 for AUC, indicating the strength of exploiting panels of miRNAs to obtain the whole picture. As demonstrated in Table 1, miR-15b and miR-21 from CSF,6 7-miRNA panel from serum,7 miR-15b and miR-21 from plasma,10 9-miRNA panel from serum,15 9-miRNA panel from CSF,20 as well as miR-21, miR-222, and miR-124-3p from serum exosome24 performed well in the diagnosis of glioma, which may be because these panels of miRNAs have common mRNA targets or participate in the common pathways such as Hedgehog, Notch, Wnt, EGFR, TGFβ and HIF1α, etc.33 For example, ANKS1B, COX15, CREBRF, GABRB2, PDHA1, PLAG1, and UBR3 are common mRNA targets for the panel of miR-15b and miR-21 predicted by miRDB.34 The molecular mechanism underlying the limitation of a single-miRNA biomarker may be that aberrant levels of single miRNA might be associated with several different types of cancers.35 More importantly, the development of cancer can be perceived as the result of a complex multi-stage process of epigenetic and genomic abnormalities, and hence should be targeted by multiple miRNAs.36 Therefore, it is sensible to employ panels of miRNAs for the purpose of evading the limitations in applying miRNAs as non-invasive biomarkers in cancer identification, especially when it is hard to conduct a regular biopsy subjected to localized pathological conditions. In addition, diagnostic accuracy of CSF-based assays was 0.96 for the pooled AUC. However, as there were only six included studies involving in CSF-based miRNAs, it is best to treat the result cautiously. Last but not least, ethnicity seemed to be another potential source of the heterogeneity, with a pooled AUC of 0.93 for Asian-based miRNA assays, and a pooled AUC of 0.88 for Caucasian-based ones. This might be attributed to different hereditary factors and living backgrounds between the two populations. The origin of ethnicity-related discrepancy is still unclear and might be ascribed to unknown mechanisms. Therefore, large scale investigations should be carried out to further verify whether these ethnicity-related discrepancies truly exist.
In a meta-analysis from 2018, Zhou et al.37 systematically assessed the potential diagnostic value of miRNAs for glioma based on 28 articles (23 in blood, 3 in CSF, and 2 in frozen brain tissue) including 2528 glioma patients and 2563 controls. The results showed that pooled sensitivity, specificity, and AUC of overall miRNAs were 0.85, 0.90, and 0.93, respectively. However, our study aimed to evaluate the diagnostic significance of cell-free miRNAs from serum, plasma, and CSF, which differed from Zhou et al’s37 meta-analysis. Moreover, they did not conduct subgroup analyses of WHO grade, reference gene, and miRNA origin (EVs- vs. non-EVs based-miRNA), which was less comprehensive than our study.
Although we did our best to perform an accurate analysis, the following limitations existed and should be noted. Above all, there were no data published regarding African populations. Moreover, it is vital to identify and exploit miRNA profiles capable of distinguishing cancer from other diseases with similar symptoms. Nonetheless, the majority of studies included in this meta-analysis merely sought to differentiate glioma patients from healthy controls, and did not involve diseases with similar symptoms. In addition, there was a low amount of miRNAs in body fluids, and almost no acknowledged endogenous reference genes, which limits the clinical application of accurate quantification of miRNAs in body fluids. It is difficult for cross-comparison among studies performed by different laboratories due to lack of convention for accurate quantification of miRNAs in body fluids; thus, a normalized procedure should be set up and preferably be followed across all studies to minimize procedure-based bias. Lastly, statistical heterogeneity existed in our analysis due to sample size, WHO grade, reference gene, miRNA origin, miRNA profiling, specimen types, and ethnicity, which may have had an impact on the results.
Declaration of conflicting interest
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/ or publication of this article: This work was supported by The Key Research and Development Program of Linyi City (grant number 2016ZK001) and Medicine and health science technology development plan of Shandong Province (grant number 2017WS323).
Supplemental material
Supplemental material for this article is available online.
ORCID iD
Jinfeng Wang https://orcid.org/0000-0002-4454-8997
References
1. Clarke, J, Butowski, N, Chang, S. Recent advances in therapy for glioblastoma. Arch Neurol 2010; 67: 279–283.
Google Scholar | Medline
2. Brandes, AA, Tosoni, A, Franceschi, E. Recurrence pattern after temozolomide concomitant with and adjuvant to radiotherapy in newly diagnosed patients with glioblastoma: correlation With MGMT promoter methylation status. J Clin Oncol 2009; 27: 1275–1279.
Google Scholar | Medline | ISI
3. Westphal, M, Lamszus, K. Circulating biomarkers for gliomas. Nat Rev Neurol 2015; 11: 556–566.
Google Scholar | Medline
4. Kong, YW, Ferland-McCollough, D, Jackson, TJ. microRNAs in cancer management. Lancet Oncol 2012; 13: e249–e258.
Google Scholar | Medline | ISI
5. Wang, Q, Li, P, Li, A. Plasma specific miRNAs as predictive biomarkers for diagnosis and prognosis of glioma. J Exp Clin Cancer Res 2012; 31: 97.
Google Scholar | Medline
6. Baraniskin, A, Kuhnhenn, J, Schlegel, U. Identification of microRNAs in the cerebrospinal fluid as biomarker for the diagnosis of glioma. Neuro Oncol 2012; 14: 29–33.
Google Scholar | Medline
7. Yang, C, Wang, C, Chen, X. Identification of seven serum microRNAs from a genome-wide serum microRNA expression profile as potential noninvasive biomarkers for malignant astrocytomas. Int J Cancer 2013; 132: 116–127.
Google Scholar | Medline
8. Akers, JC, Ramakrishnan, V, Kim, R. MiR-21 in the extracellular vesicles (EVs) of cerebrospinal fluid (CSF): a platform for glioblastoma biomarker development. PLoS One 2013; 8: e78115.
Google Scholar | Medline
9. Manterola, L, Guruceaga, E, Gallego Perez-Larraya, J. A small noncoding RNA signature found in exosomes of GBM patient serum as a diagnostic tool. Neuro Oncol 2014; 16: 520–527.
Google Scholar | Medline
10. Ivo D’Urso, P, Fernando D’Urso, O, Damiano Gianfreda, C. miR-15b and miR-21 as Circulating Biomarkers for Diagnosis of Glioma. Curr Genomics 2015; 16: 304–311.
Google Scholar | Medline
11. Sun, J, Liao, K, Wu, X. Serum microRNA-128 as a biomarker for diagnosis of glioma. Int J Clin Exp Med 2015; 8: 456–463.
Google Scholar | Medline
12. Lai, NS, Wu, DG, Fang, XG. Serum microRNA-210 as a potential noninvasive biomarker for the diagnosis and prognosis of glioma. Br J Cancer 2015; 112: 1241–1246.
Google Scholar | Medline | ISI
13. Wu, J, Li, L, Jiang, C. Identification and evaluation of serum MicroRNA-29 family for glioma screening. Mol Neurobiol 2015; 52: 1540–1546.
Google Scholar | Medline
14. Shao, N, Wang, L, Xue, L. Plasma miR-454-3p as a potential prognostic indicator in human glioma. Neurol Sci 2015; 36: 309–313.
Google Scholar | Medline
15. Zhi, F, Shao, N, Wang, R. Identification of 9 serum microRNAs as potential noninvasive biomarkers of human astrocytoma. Neuro Oncol 2015; 17: 383–391.
Google Scholar | Medline
16. Xiao, Y, Zhang, L, Song, Z. Potential Diagnostic and Prognostic Value of Plasma Circulating MicroRNA-182 in Human Glioma. Med Sci Monit 2016; 22: 855–862.
Google Scholar | Medline
17. Yue, X, Lan, F, Hu, M. Downregulation of serum microRNA-205 as a potential diagnostic and prognostic biomarker for human glioma. J Neurosurg 2016; 124: 122–128.
Google Scholar | Medline
18. Zhang, R, Pang, B, Xin, T. Plasma miR-221/222 family as novel descriptive and prognostic biomarkers for glioma. Mol Neurobiol 2016; 53: 1452–1460.
Google Scholar | Medline
19. Wei, X, Chen, D, Lv, T. Serum MicroRNA-125b as a Potential Biomarker for Glioma Diagnosis. Mol Neurobiol 2016; 53: 163–170.
Google Scholar | Medline
20. Akers, JC, Hua, W, Li, H. A cerebrospinal fluid microRNA signature as biomarker for glioblastoma. Oncotarget 2017; 8: 68769–68779.
Google Scholar | Medline
21. Tang, Y, Zhao, S, Wang, J. Plasma miR-122 as a potential diagnostic and prognostic indicator in human glioma. Neurol Sci 2017; 38: 1087–1092.
Google Scholar | Medline
22. Huang, Q, Wang, C, Hou, Z. Serum microRNA-376 family as diagnostic and prognostic markers in human gliomas. Cancer Biomark 2017; 19: 137–144.
Google Scholar | Medline
23. Lan, F, Qing, Q, Pan, Q. Serum exosomal miR-301a as a potential diagnostic and prognostic biomarker for human glioma. Cell Oncol (Dordr) 2018; 41: 25–33.
Google Scholar | Medline
24. Santangelo, A, Imbruce, P, Gardenghi, B. A microRNA signature from serum exosomes of patients with glioma as complementary diagnostic biomarker. J Neurooncol 2018; 136: 51–62.
Google Scholar | Medline
25. Leeflang, MM, Deeks, JJ, Gatsonis, C. Systematic reviews of diagnostic test accuracy. Ann Intern Med 2008; 149: 889–897.
Google Scholar | Medline | ISI
26. Whiting, PF, Rutjes, AW, Westwood, ME. QUADAS-2: a revised tool for the quality assessment of diagnostic accuracy studies. Ann Intern Med 2011; 155: 529–536.
Google Scholar | Medline | ISI
27. Higgins, JP, Thompson, SG, Deeks, JJ. Measuring inconsistency in meta-analyses. Bmj 2003; 327: 557–560.
Google Scholar | Medline
28. Deeks, JJ, Macaskill, P, Irwig, L. The performance of tests of publication bias and other sample size effects in systematic reviews of diagnostic test accuracy was assessed. J Clin Epidemiol 2005; 58: 882–893.
Google Scholar | Medline | ISI
29. Wang, J, Bettegowda, C. Applications of DNA-based liquid biopsy for central nervous system neoplasms. J Mol Diagn 2017; 19: 24–34.
Google Scholar | Medline
30. Li, Y, Wang, D, Wang, L. Distinct genomic aberrations between low-grade and high-grade gliomas of Chinese patients. PLoS One 2013; 8: e57168.
Google Scholar | Medline
31. Kosaka, N, Yoshioka, Y, Fujita, Y. Versatile roles of extracellular vesicles in cancer. J Clin Invest 2016; 126: 1163–1172.
Google Scholar | Medline
32. Kalluri, R. The biology and function of exosomes in cancer. J Clin Invest 2016; 126: 1208–1215.
Google Scholar | Medline | ISI
33. Kit, OI, Vodolazhsky, DI, Rostorguev, EE. The role of microRNA in regulation of signaling pathways in gliomas. Biochem Moscow Suppl Ser B 2018; 12: 1–21.
Google Scholar
34. Wong, N, Wang, X. miRDB: an online resource for microRNA target prediction and functional annotations. Nucleic Acids Res 2015; 43: D146–152.
Google Scholar | Medline | ISI
35. Sita-Lumsden, A, Dart, DA, Waxman, J. Circulating microRNAs as potential new biomarkers for prostate cancer. Br J Cancer 2013; 108: 1925–1930.
Google Scholar | Medline | ISI
36. Zen, K, Zhang, CY. Circulating microRNAs: a novel class of biomarkers to diagnose and monitor human cancers. Med Res Rev 2012; 32: 326–348.
Google Scholar | Medline | ISI
37. Zhou, Q, Liu, J, Quan, J. MicroRNAs as potential biomarkers for the diagnosis of glioma: A systematic review and meta-analysis. Cancer Sci 2018; 109: 2651–2659.
Google Scholar | Medline
View Abstract
Δεν υπάρχουν σχόλια:
Δημοσίευση σχολίου