Accepted Article
Article Type: 15 Original Article - Asia (rest of)
MicroRNAs as a potential tool in the differential diagnosis of thyroid cancer: a systematic review and meta-analysis
Wei-Jun Wei*; Chen-Tian Shen*; Hong-Jun Song; Zhong-Ling Qiu# and Quan-Yong Luo# Department of Nuclear Medicine, Shanghai Jiao Tong University Affiliated Sixth People’s Hospital, 600 Yishan Road, Shanghai 200233, People’s Republic of China *
Wei-Jun Wei and Chen-Tian Shen contributed equally to this work.
Short title: MicroRNAs in DTC
#
Correspondence and reprint requests to:
Quan-Yong Luo, Department of Nuclear Medicine, Shanghai Jiao Tong University Affiliated Sixth People’s Hospital, 600 Yishan Road, Shanghai, 200233, P.R. China. E-mail: [email protected]. Tel: 86-21-64369181, Fax: 86-21-64701361 Zhong-Ling Qiu, Department of Nuclear Medicine, Shanghai Jiao Tong University Affiliated Sixth This article has been accepted for publication and undergone full peer review but has not been through the copyediting, typesetting, pagination and proofreading process, which may lead to differences between this version and the Version of Record. Please cite this article as doi: 10.1111/cen.12696 This article is protected by copyright. All rights reserved.
Accepted Article
People’s Hospital, 600 Yishan Road, Shanghai, 200233, P.R. China. E-mail:[email protected]. Tel: 86-21-64369181, Fax: 86-21-64701361 Keywords MicroRNAs, Thyroid cancer, FNAB, Diagnostic value, Meta-analysis
Declaration of interest The authors declare that there is no conflict of interest that could be perceived as prejudicing the impartiality of the research reported.
Funding This study was sponsored by the National Natural Science Foundation of Chin a (No: 81271611& 81201115).
Summary Objective Thyroid cancer is the most common endocrine malignancy and its incidence has been increasing over the last thirty years. Several studies have suggested that miRNAs may play a significant role in the differential diagnosis of indeterminate thyroid nodules. To systematically evaluate the utility of miRNAs in discriminating malignant thyroid nodules from benign ones on fine-needle aspiration biopsy (FNAB) samples, a systematic review and meta-analysis of the published literatures were carried out.
This article is protected by copyright. All rights reserved.
Accepted Article
Patients and Design 361 samples, obtained from 341 patients, were included in the research and summary sensitivity (SEN), specificity (SPE), positive likelihood ratios (PLR), negative likelihood ratios (NLR), diagnostic odds ratio (DOR) were calculated. Then summary receiver operating characteristic curves (SROCs) and areas under the SROC curves (AUCs) were calculated to further estimate the overall diagnostic value of miRNAs in thyroid cancer.
Results The overall pooled SEN, SPE and AUC are 0.75, 0.81, 0.89, respectively. For multiple miRNAs assays, the pooled SEN, SPE and AUC are 0.87, 0.75, 0.68, respectively. For single miRNA assays, the corresponding results are 0.71, 0.84 ,0.87, respectively. The corresponding statistical results for differentiating indeterminate FNAB samples are 0.92, 0.68, 0.86, respectively.
Conclusion Our current meta-analysis suggests that miRNAs may serve as a novel diagnostic tool in distinguishing malignant thyroid nodules from benign ones on FNAB specimens. In addition, subgroup analysis suggests that a panel of miRNAs may have a higher sensitivity but a relatively lower specificity than that of single miRNA in distinguishing thyroid nodules.
This article is protected by copyright. All rights reserved.
Accepted Article
Introduction Thyroid cancer is the most common endocrine neoplasms, which accounts for approximately 1.7% of all cancer diagnoses.1The incidence of thyroid cancer has increased continuously over the last three decades.2Differentiated thyroid cancer (DTC) comprises the vast majority of thyroid cancers, nearly 85% of cases are papillary thyroid cancer (PTC) compared to about 10% that have follicular histology.3 Currently, the most important diagnostic method in the
detection
of
thyroid
cancer
is
fine-needle
aspiration
biopsy
(FNAB).Nevertheless, up to 16.6-22.5 % of the detections can not be diagnosed definitely , owing to wrong diagnosis or sampling errors.4 In order to avoid a 20-30% risk of malignance in indeterminate samples, repeated aspiration or a diagnostic thyroidectomy is always performed for definite diagnosis.5 However, given the vast majority of final histopathology diagnoses are benign and unnecessary thyroidectomy harbor risks like recurrent laryngeal nerve injury, hypoparathyroidism and lifelong thyroid hormone replacement, developing noninvasive and accurate method to use along with FNAB becomes quite imperative. MicroRNAs are small non-protein-coding RNAs of approximately 21-23 nucleotides and have a role in many cellular processes, including differentiation, proliferation, apoptosis and carcinogenesis.6-7Previous investigations have detected the dysregulation of miRNAs in malignant thyroid nodules, implying the potential role of miRNAs in differentiating benign thyroid nodules from
This article is protected by copyright. All rights reserved.
Accepted Article
malignant ones. And there is mounting evidence supporting the roles of miRNAs in thyroid carcinogenesis.8-10 MicroRNAs are remarkably stable in vitro and can be extracted from fresh FNAB samples or from residual cells left within the needle cup, which set the foundation for miRNAs as a clinical application in distinguishing indeterminate thyroid nodules. 11 Previous studies have identified and validated the feasibility of applying miRNAs in FNAB samples to increase the diagnostic accuracy for indeterminate thyroid nodules. However, the results of these studies are inconsistent. In Agretti’s study, the decision model comprising three miRNAs (miR-146b, -155 and -221) correctly predicted 59% of the 53 indeterminate lesions.12By measuring the expression levels of four miRNAs (miR-146b, miR-221, miR-187, and miR-30d), Shen et al. constructed a model that identified 73% of 30 atypia thyroid samples in the validation group.13 Considering the limits of single study, we conduct this systematic review and meta-analysis to evaluate the comprehensive diagnostic efficacy of miRNAs based on the data mainly extracted from validation groups.
Materials and Methods Keywords including ‘‘thyroid cancer’’ or “thyroid carcinoma,” “MicroRNA” or “miRNA” and “diagnosis” with limitations in ‘‘humans’’ and ‘‘English’’ were searched on PubMed and EMBASE up to July 1, 2014. Publications were considered eligible if they met the following criteria:(1) articles on the
This article is protected by copyright. All rights reserved.
Accepted Article
diagnostic value of miRNA for thyroid cancer;(2) Enough data were provided in the study for further analysis;(3) control groups were confirmed to be free of malignancy by final histopathology;(4) miRNAs were extracted from FNAB samples or Snap-frozen tissue;(5) diagnoses of thyroid cancer were based on standard histopathology. Articles were excluded based on the following criteria: (1)reviews, letters or meta-analyses;(2)non-English articles;(3)unpublished articles;(4)studies without key data;(5) articles focused on the prognostic value of miRNAs.
Quality assessment The qualities of the included studies were assessed independently by two reviewers (W-J Wei and C-T Shen) using the revised Quality Assessment of Diagnostic Accuracy Studies (QUADAS-2) (15). The QUADAS-2 comprises 4 domains: patient selection, index test, reference standard, flow and timing. Questions related to the specific domain were set to assess the applicability and risk of bias. As the data used for meta-analysis were mainly collected from validation groups, we primarily emphasize and evaluate the quality of studies chosen for meta-analysis.
Data extraction Two reviewers independently read the titles and abstracts of the included articles and judged their eligibility. After excluding articles that did not met our
This article is protected by copyright. All rights reserved.
Accepted Article
inclusion criteria, the full texts were read and relevant data were extracted from included studies. Any disagreements were resolved by consensus. The extracted data were recorded on a standardized form including information as follows: first author, year, country, sample size, patients/lesions, sensitivity, specificity, materials, miRNA profiles as well as the status of miRNAs.
Statistical methods Diagnostic meta-analysis was conducted on Meta-disc 1.4 using true-positive (TP), false-positive (FP), false-negative (FN), and true negative (TN) extracted directly or through recalculation based on reported accuracy in combination with the sensitivity, specificity and sample size of a study. The overall statistical estimates of miRNA for differentiating thyroid nodules were calculated. Considering the fact that we have collected data from both multiple miRNAs assays and single miRNA assays, we have drawn not only the pooled sensitivity, specificity and SROC of multiple miRNAs assays but also the corresponding results of single miRNA assays. Subgroup analysis of studies, which used indeterminate FNAB samples, was also performed. The AUCs and the maximum point of intersection between sensitivity and specificity (Q value), which reflects the closeness of SROC to the upper left corner and the accuracy of diagnostic tests, were calculated in further analysis. In addition, we drew a funnel plot to investigate publication bias using stata12.
This article is protected by copyright. All rights reserved.
Accepted Article
Results Literature Search In total, 140 articles were identified from a primary literature search in Pubmed and EMBASE. Following the inclusion and exclusion criteria illustrated above, two researchers (W-J Wei and C-T Shen) independently screened the titles, abstracts and key words, 70 articles were excluded as they were duplicates. Then we systematically assessed the remaining 70 articles by reading full-text, as a result, 60 articles were omitted during this process because they were reviews, meta-analysis, non-English articles, articles using circulating miRNAs or focusing on other cancer types. The flow diagram of the article selection process is presented in Fig. 1.Finally, 10 selected articles were included in this review.11-13, 15-21Of the 10 candidate articles, 3 articles were excluded in further analysis because they lacked relevant data needed to carry out our meta-analysis.16,
20-21
Therefore, only data extracted from the
remaining 7 publications were used in our meta-analysis and main characteristics of the 7 publications are listed in Table 1.
Study Characteristics and Quality Assessments Finally, 16 groups from 7 articles were included in our meta-analysis, containing 341 patients with 163 malignant samples and 198 benign samples. Among the 16 diagnostic studies, 7 studies were conducted in Israel, 5 in USA, 3 in China and 1 in Italy. All the 16 studies used the quantitative real-time
This article is protected by copyright. All rights reserved.
Accepted Article
reverse transcription-PCR (RT-PCR) method to measure the expression of miRNAs based on cells obtained mainly from fine-needle aspiration samples. 11 studies evaluated the diagnostic value of single miRNA while 5 studies assessed the diagnostic value of multiple miRNAs. In the seven articles included into our meta-analysis, three miRNAs (miR-30a-3p, -30d, -7) were reported significantly downregulated in PTC. You Peng et al. found down-regulation of miR-30a-3p in PTC tissue compared to the adjacent normal tissue.15Shen et al. demonstrated that miRNA amplification from FNA samples is feasible and the level of miR-30d was significantly lower in the malignant tissue.13 Kitano et al. found that downregulated miRNA-7 performed best when distinguishing benign from malignant thyroid FNAB samples.17
Five studies used indeterminate FNAB samples to validate the diagnostic value of miRNA.11-13,17,18 Agretti et al. obtained an established prediction model (miRNA-146b, -155, -221) to classify a nodule as benign or malignant and validated the diagnostic value of the model using 53 undetermined thyroid nodules.12 Shen et al. identified a set of four miRNAs (miR-146b, -221, -187, and -30d) that could differentiate non-diagnostic FNAB specimens.13 Kitano and his colleagues found that miRNA-7, which was significantly downregulated in malignant samples, was the best predictor to distinguish indeterminate thyroid nodules. When the model was applied to the 59 indeterminate lesions it had a
This article is protected by copyright. All rights reserved.
Accepted Article
sensitivity of 100% with a specificity of 20%.17 Keutgen X.M. used a model derived from 29 indeterminate thyroid lesions to validate on an independent set of 72 indeterminate FNA samples.18 Mazeh et al. extracted miRNAs (miR-21, -31,-146b, -187, -221, and -222) from residual cells left within FNAB cup and analyzed their expressions on 11 indeterminate FNA samples.11 The qualities of the 16 studies were methodologically assessed using QUADAS-2.14 We firmly believe that patient selection plays such a role in conducting experiments that data used in this meta-analysis are mainly from validation groups. As a whole, the qualities of included studies are satisfying and eligible. Quality assessments are shown in a bar chart of QUADAS-2, as is shown in Fig. 2.
Diagnostic accuracy of miRNAs for thyroid cancer According to Berlin, the I2 index not only assessed heterogeneity in a meta-analysis but also the extent of that heterogeneity.22 It is considered a more appropriate procedure than Dixon’s Q test for assessing whether or not there is true heterogeneity among the studies in a meta-analysis.23 By calculating the I2, which was interpreted as low (25%), moderate (50%), and high (75%),we found heterogeneity between studies in sensitivity and specificity data (I2=78.6 % and I2=90.5 %，respectively).24 Therefore, the random-effect model was used to calculate the pool estimates in this study. The overall pooled SEN, SPE, PLR, NLR, DOR and AUC are 0.75, 0.81, 5.88, 0.33,
This article is protected by copyright. All rights reserved.
Accepted Article
25.5 and 0.89, respectively. The overall forest plots of sensitivity and specificity are shown in Fig.3. And the overall SROC curve is presented in Fig.4 (a). For multiple miRNAs assays, the corresponding results are 0.87, 0.75, 3.50, 0.20, 20.3 and 0.68, respectively. For single miRNA assays, the corresponding pooled results are 0.71, 0.84, 9.47, 0.37, 28.9 and 0.87. The SROC curves of multiple miRNAs and single miRNA are shown in Fig.5. In subgroup analysis of five studies using indeterminate FNAB samples, the corresponding pooled results are 0.92, 0.68, 2.9, 0.12, 25.0 and 0.86, respectively. The AUC values of four analyses are 0.89, 0.68, 0.87 and 0.86, indicating that miRNAs have high potential utility for the differential diagnosis of thyroid cancer. From the statistical results, we find that multiple miRNAs assays appear to be more sensitive than single miRNA assays. However, multiple miRNAs assays showed a relatively lower specificity and AUC value when compared with single miRNA assays. The potential diagnostic utility of miRNA is mainly for the indeterminate FNAB specimens, and subgroup analysis using indeterminate FNAB samples showed high sensitivity of 0.92 and satisfying AUC value of 0.86. The overall statistical results of this meta-analysis are presented in Table 2.
Discussion As the incidence and the morbidity of thyroid cancer have increased over the recent years, development of suitable biomarkers is critical for accurately diagnosing thyroid cancer.2,25 MicroRNAs were first described in Cell in 1993,26
This article is protected by copyright. All rights reserved.
Accepted Article
and were subsequently found to be of vital importance in many biological and pathological
processes,
impacting
cell
differentiation,
proliferation,
survival ,apoptosis and tumorigenesis.27 By using microarray analysis and RT-PCR technology, the aberrant expressions of miRNAs have been detected and analyzed in various human cancers, suggesting that miRNA dysregulation is crucial in tumor development and progression.28-31 In thyroid tumors, most studies have focused on miRNA analysis of papillary thyroid cancer and have identified many upregulated as well as several downregulated miRNAs. Chen et al. found that miRNA-146b was most consistently overexpressed in both PTC and follicular variants of papillary thyroid cancer (FVPTC) and miRNAs from formalin-fixed paraffin-embedded (FFPE) tissues were analyzable.32 Pallante, P. analyzed miRNA expression profile in human thyroid papillary carcinoma and found that a significant increase in miRNAs (miR-221, -222 and -181b) was detected in PTC when compared to normal thyroid tissue.33 Nikiforova found that a set of seven miRNAs (miR-187, -221, -222, -146b, -155, -224, and -197) were significantly overexpressed in thyroid tumors.34 Under this circumstances, we designed this meta-analysis to estimate the potential diagnostic value of miRNAs for thyroid specimens. Making full use of the data extracted from 16 previously published independent studies, we found that miRNAs may serve as a novel diagnostic tool in distinguishing thyroid nodules. When it comes to differentiating benign thyroid nodules from malignant ones, a set of miRNAs seems to be more
This article is protected by copyright. All rights reserved.
Accepted Article
sensitive than single miRNA with sensitivity=0.87 compared to that of 0.71. However, our meta-analysis has several limitations. First, even if we tried to incorporate all related articles, we may omit some relevant publications. Second, among 16 studies enrolled in our meta-analysis, only 5 studies used indeterminate FNAB samples to validate the potential diagnostic value of miRNAs. The small cohort of patients limits our work’s eligibility and may cause heterogeneity. Third, even though the funnel plot showed no obvious publication bias with pvalue= 0.30, as is depicted in Fig.4 (b), several factors may contribute to heterogeneity: patient selection, characteristics of subjects, study design, different pathological interpretations. Unfortunately, we failed to find the potential sources for heterogeneity because of limited information.
In summary, we showed that miRNAs are promising molecular biomarkers that may significantly improve the diagnostic accuracy of thyroid nodules when using along with FNAB. However, future researches should construct a panel of stably detectable miRNAs and validate the diagnostic value of the panel using a larger cohort of indeterminate FNAB specimens, which could provide highly accurate clinical information for the differential diagnosis of thyroid nodules. Fortunately, miRNAs are comparatively stable in FFPE tissues and can be extracted from FNAB samples, implicating a great opportunity to design large retrospective or prospective studies.
This article is protected by copyright. All rights reserved.
Accepted Article
References 1
Ferlay, J., Shin, H.R., Bray, F. et al. (2010) Estimates of worldwide burden of
cancer in 2008: GLOBOCAN 2008. Int J Cancer, 127, 2893-2917. 2
Pellegriti, G., Frasca, F., Regalbuto, C. et al. (2013) Worldwide increasing
incidence of thyroid cancer: update on epidemiology and risk factors. J Cancer
Epidemiol, 2013, 965212. 3
Hundahl, S.A., Fleming, I.D., Fremgen, A.M. et al. (1998) A National Cancer
Data Base report on 53,856 cases of thyroid carcinoma treated in the U.S., 1985-1995 [see commetns]. Cancer, 83, 2638-2648. 4
Papini, E., Guglielmi, R., Bianchini, A. et al. (2002) Risk of malignancy in
nonpalpable thyroid nodules: predictive value of ultrasound and color-Doppler features. J Clin Endocrinol Metab, 87, 1941-1946. 5
Mehanna, H.M., Jain, A., Morton, R.P. et al. (2009) Investigating the thyroid
nodule. BMJ, 338, b733. 6
Wilmott, J.S., Zhang, X.D., Hersey, P. et al. (2011) The emerging important
role of microRNAs in the pathogenesis, diagnosis and treatment of human cancers. Pathology, 43, 657-671. 7
Martinez, I. & Dimaio, D. (2011) B-Myb, cancer, senescence, and
microRNAs. Cancer Res, 71, 5370-5373. 8
Marini, F., Luzi, E. & Brandi, M.L. (2011) MicroRNA Role in Thyroid Cancer
Development. J Thyroid Res, 2011, 407123. 9
de la Chapelle, A. & Jazdzewski, K. (2011) MicroRNAs in thyroid cancer. J
This article is protected by copyright. All rights reserved.
Accepted Article
Clin Endocrinol Metab, 96, 3326-3336. 10 Schmid, K.W. (2010) [Molecular pathology of thyroid tumors]. Pathologe, 31 Suppl 2, 229-233. 11 Mazeh, H., Levy, Y., Mizrahi, I. et al. (2013) Differentiating benign from malignant thyroid nodules using micro ribonucleic acid amplification in residual cells obtained by fine needle aspiration biopsy. J Surg Res, 180, 216-221. 12 Agretti, P., Ferrarini, E., Rago, T. et al. (2012) MicroRNA expression profile helps to distinguish benign nodules from papillary thyroid carcinomas starting from cells of fine-needle aspiration. Eur J Endocrinol, 167, 393-400. 13 Shen, R., Liyanarachchi, S., Li, W. et al. (2012) MicroRNA signature in thyroid fine needle aspiration cytology applied to "atypia of undetermined significance" cases. Thyroid, 22, 9-16. 14 Whiting, P.F., Rutjes, A.W., Westwood, M.E. et al. (2011) QUADAS-2: a revised tool for the quality assessment of diagnostic accuracy studies. Ann
Intern Med, 155, 529-536. 15 Peng, Y., Li, C., Luo, D.C. et al. (2014) Expression profile and clinical significance of microRNAs in papillary thyroid carcinoma. Molecules, 19, 11586-11599. 16 Vriens, M.R., Weng, J., Suh, I. et al. (2012) MicroRNA expression profiling is a potential diagnostic tool for thyroid cancer. Cancer, 118, 3426-3432. 17 Kitano, M., Rahbari, R., Patterson, E.E. et al. (2012) Evaluation of candidate diagnostic microRNAs in thyroid fine-needle aspiration biopsy samples. Thyroid,
This article is protected by copyright. All rights reserved.
Accepted Article
22, 285-291. 18 Keutgen, X.M., Filicori, F., Crowley, M.J. et al. (2012) A panel of four miRNAs accurately differentiates malignant from benign indeterminate thyroid lesions on fine needle aspiration. Clin Cancer Res, 18, 2032-2038. 19 Mazeh, H., Mizrahi, I., Halle, D. et al. (2011) Development of a microRNA-based molecular assay for the detection of papillary thyroid carcinoma in aspiration biopsy samples. Thyroid, 21, 111-118. 20 Kitano, M., Rahbari, R., Patterson, E.E. et al. (2011) Expression profiling of difficult-to-diagnose thyroid histologic subtypes shows distinct expression profiles and identify candidate diagnostic microRNAs. Ann Surg Oncol, 18, 3443-3452. 21 He, H., Jazdzewski, K., Li, W. et al. (2005) The role of microRNA genes in papillary thyroid carcinoma. Proc Natl Acad Sci U S A, 102, 19075-19080. 22 Berlin, J.A. (1995) Invited commentary: benefits of heterogeneity in meta-analysis of data from epidemiologic studies. Am J Epidemiol, 142, 383-387. 23 Huedo-Medina, T.B., Sanchez-Meca, J., Marin-Martinez, F. et al. (2006) Assessing heterogeneity in meta-analysis: Q statistic or I2 index? Psychol
Methods, 11, 193-206. 24 Moher, D., Liberati, A., Tetzlaff, J. et al. (2009) Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement. J Clin Epidemiol, 62, 1006-1012.
This article is protected by copyright. All rights reserved.
Accepted Article
25 Sipos, J.A. & Mazzaferri, E.L. (2010) Thyroid cancer epidemiology and prognostic variables. Clin Oncol (R Coll Radiol), 22, 395-404. 26 Lee, R.C., Feinbaum, R.L. & Ambros, V. (1993) The C. elegans heterochronic gene lin-4 encodes small RNAs with antisense complementarity to lin-14. Cell, 75, 843-854. 27 Hwang, H.W. & Mendell, J.T. (2007) MicroRNAs in cell proliferation, cell death, and tumorigenesis. Br J Cancer, 96 Suppl, R40-44. 28 Calin, G.A., Liu, C.G., Sevignani, C. et al. (2004) MicroRNA profiling reveals distinct signatures in B cell chronic lymphocytic leukemias. Proc Natl Acad Sci U
S A, 101, 11755-11760. 29 Tong, A.W. (2006) Small RNAs and non-small cell lung cancer. Curr Mol Med, 6, 339-349. 30 Iorio, M.V., Ferracin, M., Liu, C.G. et al. (2005) MicroRNA gene expression deregulation in human breast cancer. Cancer Res, 65, 7065-7070. 31 Akao, Y., Nakagawa, Y. & Naoe, T. (2007) MicroRNA-143 and -145 in colon cancer. DNA Cell Biol, 26, 311-320. 32 Chen, Y.T., Kitabayashi, N., Zhou, X.K. et al. (2008) MicroRNA analysis as a potential diagnostic tool for papillary thyroid carcinoma. Mod Pathol, 21, 1139-1146. 33 Pallante, P., Visone, R., Ferracin, M. et al. (2006) MicroRNA deregulation in human thyroid papillary carcinomas. Endocr Relat Cancer, 13, 497-508. 34 Nikiforova, M.N., Tseng, G.C., Steward, D. et al. (2008) MicroRNA
This article is protected by copyright. All rights reserved.
Accepted Article
expression profiling of thyroid tumors: biological significance and diagnostic utility. J Clin Endocrinol Metab, 93, 1600-1608.
Figure legends: FIGURE1:Flow chart showing the process of study selection.
FIGURE2: (a) Proportion of studies with low, high, or unclear risk of bias; (b) Proportion of studies with low or unclear concerns regarding applicability.
FIGURE3 :The pooled sensitivity (a) ,specificity (b) for miRNAs in discriminating thyroid nodules.
FIGURE4: The overall SROC (a) and Deeks’ Funnel Plot(b).
FIGURE5: The SROC curves of multiple miRNAs(a) and single miRNA(b).
This article is protected by copyright. All rights reserved.
Accepted Article
Table 1 The main characteristics of seven articles included in the meta-analysis First Ye Auth ar or
Cou ntry
Kita no Kita no
20 12 20 12
US A US A
Agre 20 tti 12
Ita ly
Maz eh#1
20 11
Isr ael
47(20/ Lesi 27) ons
You Pen g#2
20 14
Chi na
51(36/ Patie Tissue 15) nts After Resecti on
Keut 20 gen 12
US A
72(22/ Patie Indeter 50) nts minate FNAB
She
20
US
Validatio n Group (cancer/ normal) 59(17/ 42) 24(5/1 9)
Patie nts/ Lesi ons Patie nts Patie nts
Material MicroR NAs
Sta tus
Sensi tivity
Speci ficity
FNAB
miR-7
↓
29%
miR-7
↓
100 % 100 %
miR-14 6 miR-15 5 miR-22 1 miR-21 miR-31 miR-14 6b miR-18 7 miR-22 1 miR-22 2
↑ ↑ ↑
60%
58%
↑ ↑ ↑ ↑ ↑ ↑
50% 45% 80% 50% 95% 90%
miR-30 a-3p miR-14 6b-5p miR-19 9b-5p miR-22 2 miR-32 8 miR-19 7, miR-21 miR-14
↓ ↑ ↑
68% 91% 73%
100 % 100 % 100 % 100 % 100 % 100 % 73% 89% 82%
↑ ↑ ↑ ↑
100 %
Indeter minate FNAB 53(15/ Patie Indeter 38) nts minate FNAB
FNAB after resectio n
68(44/ Patie FNAB
This article is protected by copyright. All rights reserved.
↑
20%
86%
Accepted Article
n
12
A
24)
nts
Slides
↑ 89% 78% 6b ↑ miR-22 ↓ 1 miR-18 7 miR-30 d ↑ She 20 US 30(11/ Patie Indeter miR-14 ↑ 64% 79% n 12 A 19) nts minate 6b ↑ FNAB miR-22 ↓ 1 miR-18 7 miR-30 d ↑ Isr 11(9/2 Patie Indeter miR-21 Maz 20 #2 ↑ 13 ael ) nts minate miR-31 eh ↑ 88% 100 FNAB miR-14 ↑ 6b % ↑ miR-18 ↑ 7 miR-22 1 miR-22 2 F=female; M=male; TG=training group; VG=validation group; NR=not reported; #1 47 samples were obtained from 27 subjects in this study; #2 These two studies were not validation groups.
This article is protected by copyright. All rights reserved.
Accepted Arti
Table 2: The overall diagnostic value of miRNA in thyroid cancer with 95 % confidence intervals
#1:
Results
Sensitivity
Specificity
(95 %CI)
(95 %CI)
AUC
DOR
PLR
NLR
(95 %CI)
(95 %CI)
(95 %CI)
Overall
0.75(0.70-0.80)
Multiple#1
0.87 (0.78-0.93 ) 0.75 (0.67-0.83 ) 0.68 20.3 (2.43-170.16 ) 3.50( 1.40-8.77 ) 0.20(0.05-0.75 )
Single#2
0.71(0.65-0.77)
0.84(0.78-0.88)
0.87 28.9(9.48-88.03)
9.47(2.00-44.80) 0.37(0.27-0.53)
Subgroup#3 0.92(0.58-0.99)
0.68(0.43-0.86)
0.86 25.0(3.00-246)
2.90(1.40-6.10)
0.81(0.77-0.85)
0.89 25.5(9.73-66.67)
The diagnostic value of multiple miRNAs; #2: The diagnostic value of single miRNA. #3: The diagnostic efficacy of miRNA for indeterminate FNAB samples;
This article is protected by copyright. All rights reserved.
5.88(2.48-13.90) 0.33(0.23-0.47)
0.12(0.02-0.83)
Accepted Article This article is protected by copyright. All rights reserved.
Accepted Article This article is protected by copyright. All rights reserved.
Accepted Article This article is protected by copyright. All rights reserved.
Article Type: 15 Original Article - Asia (rest of)
MicroRNAs as a potential tool in the differential diagnosis of thyroid cancer: a systematic review and meta-analysis
Wei-Jun Wei*; Chen-Tian Shen*; Hong-Jun Song; Zhong-Ling Qiu# and Quan-Yong Luo# Department of Nuclear Medicine, Shanghai Jiao Tong University Affiliated Sixth People’s Hospital, 600 Yishan Road, Shanghai 200233, People’s Republic of China *
Wei-Jun Wei and Chen-Tian Shen contributed equally to this work.
Short title: MicroRNAs in DTC
#
Correspondence and reprint requests to:
Quan-Yong Luo, Department of Nuclear Medicine, Shanghai Jiao Tong University Affiliated Sixth People’s Hospital, 600 Yishan Road, Shanghai, 200233, P.R. China. E-mail: [email protected]. Tel: 86-21-64369181, Fax: 86-21-64701361 Zhong-Ling Qiu, Department of Nuclear Medicine, Shanghai Jiao Tong University Affiliated Sixth This article has been accepted for publication and undergone full peer review but has not been through the copyediting, typesetting, pagination and proofreading process, which may lead to differences between this version and the Version of Record. Please cite this article as doi: 10.1111/cen.12696 This article is protected by copyright. All rights reserved.
Accepted Article
People’s Hospital, 600 Yishan Road, Shanghai, 200233, P.R. China. E-mail:[email protected]. Tel: 86-21-64369181, Fax: 86-21-64701361 Keywords MicroRNAs, Thyroid cancer, FNAB, Diagnostic value, Meta-analysis
Declaration of interest The authors declare that there is no conflict of interest that could be perceived as prejudicing the impartiality of the research reported.
Funding This study was sponsored by the National Natural Science Foundation of Chin a (No: 81271611& 81201115).
Summary Objective Thyroid cancer is the most common endocrine malignancy and its incidence has been increasing over the last thirty years. Several studies have suggested that miRNAs may play a significant role in the differential diagnosis of indeterminate thyroid nodules. To systematically evaluate the utility of miRNAs in discriminating malignant thyroid nodules from benign ones on fine-needle aspiration biopsy (FNAB) samples, a systematic review and meta-analysis of the published literatures were carried out.
This article is protected by copyright. All rights reserved.
Accepted Article
Patients and Design 361 samples, obtained from 341 patients, were included in the research and summary sensitivity (SEN), specificity (SPE), positive likelihood ratios (PLR), negative likelihood ratios (NLR), diagnostic odds ratio (DOR) were calculated. Then summary receiver operating characteristic curves (SROCs) and areas under the SROC curves (AUCs) were calculated to further estimate the overall diagnostic value of miRNAs in thyroid cancer.
Results The overall pooled SEN, SPE and AUC are 0.75, 0.81, 0.89, respectively. For multiple miRNAs assays, the pooled SEN, SPE and AUC are 0.87, 0.75, 0.68, respectively. For single miRNA assays, the corresponding results are 0.71, 0.84 ,0.87, respectively. The corresponding statistical results for differentiating indeterminate FNAB samples are 0.92, 0.68, 0.86, respectively.
Conclusion Our current meta-analysis suggests that miRNAs may serve as a novel diagnostic tool in distinguishing malignant thyroid nodules from benign ones on FNAB specimens. In addition, subgroup analysis suggests that a panel of miRNAs may have a higher sensitivity but a relatively lower specificity than that of single miRNA in distinguishing thyroid nodules.
This article is protected by copyright. All rights reserved.
Accepted Article
Introduction Thyroid cancer is the most common endocrine neoplasms, which accounts for approximately 1.7% of all cancer diagnoses.1The incidence of thyroid cancer has increased continuously over the last three decades.2Differentiated thyroid cancer (DTC) comprises the vast majority of thyroid cancers, nearly 85% of cases are papillary thyroid cancer (PTC) compared to about 10% that have follicular histology.3 Currently, the most important diagnostic method in the
detection
of
thyroid
cancer
is
fine-needle
aspiration
biopsy
(FNAB).Nevertheless, up to 16.6-22.5 % of the detections can not be diagnosed definitely , owing to wrong diagnosis or sampling errors.4 In order to avoid a 20-30% risk of malignance in indeterminate samples, repeated aspiration or a diagnostic thyroidectomy is always performed for definite diagnosis.5 However, given the vast majority of final histopathology diagnoses are benign and unnecessary thyroidectomy harbor risks like recurrent laryngeal nerve injury, hypoparathyroidism and lifelong thyroid hormone replacement, developing noninvasive and accurate method to use along with FNAB becomes quite imperative. MicroRNAs are small non-protein-coding RNAs of approximately 21-23 nucleotides and have a role in many cellular processes, including differentiation, proliferation, apoptosis and carcinogenesis.6-7Previous investigations have detected the dysregulation of miRNAs in malignant thyroid nodules, implying the potential role of miRNAs in differentiating benign thyroid nodules from
This article is protected by copyright. All rights reserved.
Accepted Article
malignant ones. And there is mounting evidence supporting the roles of miRNAs in thyroid carcinogenesis.8-10 MicroRNAs are remarkably stable in vitro and can be extracted from fresh FNAB samples or from residual cells left within the needle cup, which set the foundation for miRNAs as a clinical application in distinguishing indeterminate thyroid nodules. 11 Previous studies have identified and validated the feasibility of applying miRNAs in FNAB samples to increase the diagnostic accuracy for indeterminate thyroid nodules. However, the results of these studies are inconsistent. In Agretti’s study, the decision model comprising three miRNAs (miR-146b, -155 and -221) correctly predicted 59% of the 53 indeterminate lesions.12By measuring the expression levels of four miRNAs (miR-146b, miR-221, miR-187, and miR-30d), Shen et al. constructed a model that identified 73% of 30 atypia thyroid samples in the validation group.13 Considering the limits of single study, we conduct this systematic review and meta-analysis to evaluate the comprehensive diagnostic efficacy of miRNAs based on the data mainly extracted from validation groups.
Materials and Methods Keywords including ‘‘thyroid cancer’’ or “thyroid carcinoma,” “MicroRNA” or “miRNA” and “diagnosis” with limitations in ‘‘humans’’ and ‘‘English’’ were searched on PubMed and EMBASE up to July 1, 2014. Publications were considered eligible if they met the following criteria:(1) articles on the
This article is protected by copyright. All rights reserved.
Accepted Article
diagnostic value of miRNA for thyroid cancer;(2) Enough data were provided in the study for further analysis;(3) control groups were confirmed to be free of malignancy by final histopathology;(4) miRNAs were extracted from FNAB samples or Snap-frozen tissue;(5) diagnoses of thyroid cancer were based on standard histopathology. Articles were excluded based on the following criteria: (1)reviews, letters or meta-analyses;(2)non-English articles;(3)unpublished articles;(4)studies without key data;(5) articles focused on the prognostic value of miRNAs.
Quality assessment The qualities of the included studies were assessed independently by two reviewers (W-J Wei and C-T Shen) using the revised Quality Assessment of Diagnostic Accuracy Studies (QUADAS-2) (15). The QUADAS-2 comprises 4 domains: patient selection, index test, reference standard, flow and timing. Questions related to the specific domain were set to assess the applicability and risk of bias. As the data used for meta-analysis were mainly collected from validation groups, we primarily emphasize and evaluate the quality of studies chosen for meta-analysis.
Data extraction Two reviewers independently read the titles and abstracts of the included articles and judged their eligibility. After excluding articles that did not met our
This article is protected by copyright. All rights reserved.
Accepted Article
inclusion criteria, the full texts were read and relevant data were extracted from included studies. Any disagreements were resolved by consensus. The extracted data were recorded on a standardized form including information as follows: first author, year, country, sample size, patients/lesions, sensitivity, specificity, materials, miRNA profiles as well as the status of miRNAs.
Statistical methods Diagnostic meta-analysis was conducted on Meta-disc 1.4 using true-positive (TP), false-positive (FP), false-negative (FN), and true negative (TN) extracted directly or through recalculation based on reported accuracy in combination with the sensitivity, specificity and sample size of a study. The overall statistical estimates of miRNA for differentiating thyroid nodules were calculated. Considering the fact that we have collected data from both multiple miRNAs assays and single miRNA assays, we have drawn not only the pooled sensitivity, specificity and SROC of multiple miRNAs assays but also the corresponding results of single miRNA assays. Subgroup analysis of studies, which used indeterminate FNAB samples, was also performed. The AUCs and the maximum point of intersection between sensitivity and specificity (Q value), which reflects the closeness of SROC to the upper left corner and the accuracy of diagnostic tests, were calculated in further analysis. In addition, we drew a funnel plot to investigate publication bias using stata12.
This article is protected by copyright. All rights reserved.
Accepted Article
Results Literature Search In total, 140 articles were identified from a primary literature search in Pubmed and EMBASE. Following the inclusion and exclusion criteria illustrated above, two researchers (W-J Wei and C-T Shen) independently screened the titles, abstracts and key words, 70 articles were excluded as they were duplicates. Then we systematically assessed the remaining 70 articles by reading full-text, as a result, 60 articles were omitted during this process because they were reviews, meta-analysis, non-English articles, articles using circulating miRNAs or focusing on other cancer types. The flow diagram of the article selection process is presented in Fig. 1.Finally, 10 selected articles were included in this review.11-13, 15-21Of the 10 candidate articles, 3 articles were excluded in further analysis because they lacked relevant data needed to carry out our meta-analysis.16,
20-21
Therefore, only data extracted from the
remaining 7 publications were used in our meta-analysis and main characteristics of the 7 publications are listed in Table 1.
Study Characteristics and Quality Assessments Finally, 16 groups from 7 articles were included in our meta-analysis, containing 341 patients with 163 malignant samples and 198 benign samples. Among the 16 diagnostic studies, 7 studies were conducted in Israel, 5 in USA, 3 in China and 1 in Italy. All the 16 studies used the quantitative real-time
This article is protected by copyright. All rights reserved.
Accepted Article
reverse transcription-PCR (RT-PCR) method to measure the expression of miRNAs based on cells obtained mainly from fine-needle aspiration samples. 11 studies evaluated the diagnostic value of single miRNA while 5 studies assessed the diagnostic value of multiple miRNAs. In the seven articles included into our meta-analysis, three miRNAs (miR-30a-3p, -30d, -7) were reported significantly downregulated in PTC. You Peng et al. found down-regulation of miR-30a-3p in PTC tissue compared to the adjacent normal tissue.15Shen et al. demonstrated that miRNA amplification from FNA samples is feasible and the level of miR-30d was significantly lower in the malignant tissue.13 Kitano et al. found that downregulated miRNA-7 performed best when distinguishing benign from malignant thyroid FNAB samples.17
Five studies used indeterminate FNAB samples to validate the diagnostic value of miRNA.11-13,17,18 Agretti et al. obtained an established prediction model (miRNA-146b, -155, -221) to classify a nodule as benign or malignant and validated the diagnostic value of the model using 53 undetermined thyroid nodules.12 Shen et al. identified a set of four miRNAs (miR-146b, -221, -187, and -30d) that could differentiate non-diagnostic FNAB specimens.13 Kitano and his colleagues found that miRNA-7, which was significantly downregulated in malignant samples, was the best predictor to distinguish indeterminate thyroid nodules. When the model was applied to the 59 indeterminate lesions it had a
This article is protected by copyright. All rights reserved.
Accepted Article
sensitivity of 100% with a specificity of 20%.17 Keutgen X.M. used a model derived from 29 indeterminate thyroid lesions to validate on an independent set of 72 indeterminate FNA samples.18 Mazeh et al. extracted miRNAs (miR-21, -31,-146b, -187, -221, and -222) from residual cells left within FNAB cup and analyzed their expressions on 11 indeterminate FNA samples.11 The qualities of the 16 studies were methodologically assessed using QUADAS-2.14 We firmly believe that patient selection plays such a role in conducting experiments that data used in this meta-analysis are mainly from validation groups. As a whole, the qualities of included studies are satisfying and eligible. Quality assessments are shown in a bar chart of QUADAS-2, as is shown in Fig. 2.
Diagnostic accuracy of miRNAs for thyroid cancer According to Berlin, the I2 index not only assessed heterogeneity in a meta-analysis but also the extent of that heterogeneity.22 It is considered a more appropriate procedure than Dixon’s Q test for assessing whether or not there is true heterogeneity among the studies in a meta-analysis.23 By calculating the I2, which was interpreted as low (25%), moderate (50%), and high (75%),we found heterogeneity between studies in sensitivity and specificity data (I2=78.6 % and I2=90.5 %，respectively).24 Therefore, the random-effect model was used to calculate the pool estimates in this study. The overall pooled SEN, SPE, PLR, NLR, DOR and AUC are 0.75, 0.81, 5.88, 0.33,
This article is protected by copyright. All rights reserved.
Accepted Article
25.5 and 0.89, respectively. The overall forest plots of sensitivity and specificity are shown in Fig.3. And the overall SROC curve is presented in Fig.4 (a). For multiple miRNAs assays, the corresponding results are 0.87, 0.75, 3.50, 0.20, 20.3 and 0.68, respectively. For single miRNA assays, the corresponding pooled results are 0.71, 0.84, 9.47, 0.37, 28.9 and 0.87. The SROC curves of multiple miRNAs and single miRNA are shown in Fig.5. In subgroup analysis of five studies using indeterminate FNAB samples, the corresponding pooled results are 0.92, 0.68, 2.9, 0.12, 25.0 and 0.86, respectively. The AUC values of four analyses are 0.89, 0.68, 0.87 and 0.86, indicating that miRNAs have high potential utility for the differential diagnosis of thyroid cancer. From the statistical results, we find that multiple miRNAs assays appear to be more sensitive than single miRNA assays. However, multiple miRNAs assays showed a relatively lower specificity and AUC value when compared with single miRNA assays. The potential diagnostic utility of miRNA is mainly for the indeterminate FNAB specimens, and subgroup analysis using indeterminate FNAB samples showed high sensitivity of 0.92 and satisfying AUC value of 0.86. The overall statistical results of this meta-analysis are presented in Table 2.
Discussion As the incidence and the morbidity of thyroid cancer have increased over the recent years, development of suitable biomarkers is critical for accurately diagnosing thyroid cancer.2,25 MicroRNAs were first described in Cell in 1993,26
This article is protected by copyright. All rights reserved.
Accepted Article
and were subsequently found to be of vital importance in many biological and pathological
processes,
impacting
cell
differentiation,
proliferation,
survival ,apoptosis and tumorigenesis.27 By using microarray analysis and RT-PCR technology, the aberrant expressions of miRNAs have been detected and analyzed in various human cancers, suggesting that miRNA dysregulation is crucial in tumor development and progression.28-31 In thyroid tumors, most studies have focused on miRNA analysis of papillary thyroid cancer and have identified many upregulated as well as several downregulated miRNAs. Chen et al. found that miRNA-146b was most consistently overexpressed in both PTC and follicular variants of papillary thyroid cancer (FVPTC) and miRNAs from formalin-fixed paraffin-embedded (FFPE) tissues were analyzable.32 Pallante, P. analyzed miRNA expression profile in human thyroid papillary carcinoma and found that a significant increase in miRNAs (miR-221, -222 and -181b) was detected in PTC when compared to normal thyroid tissue.33 Nikiforova found that a set of seven miRNAs (miR-187, -221, -222, -146b, -155, -224, and -197) were significantly overexpressed in thyroid tumors.34 Under this circumstances, we designed this meta-analysis to estimate the potential diagnostic value of miRNAs for thyroid specimens. Making full use of the data extracted from 16 previously published independent studies, we found that miRNAs may serve as a novel diagnostic tool in distinguishing thyroid nodules. When it comes to differentiating benign thyroid nodules from malignant ones, a set of miRNAs seems to be more
This article is protected by copyright. All rights reserved.
Accepted Article
sensitive than single miRNA with sensitivity=0.87 compared to that of 0.71. However, our meta-analysis has several limitations. First, even if we tried to incorporate all related articles, we may omit some relevant publications. Second, among 16 studies enrolled in our meta-analysis, only 5 studies used indeterminate FNAB samples to validate the potential diagnostic value of miRNAs. The small cohort of patients limits our work’s eligibility and may cause heterogeneity. Third, even though the funnel plot showed no obvious publication bias with pvalue= 0.30, as is depicted in Fig.4 (b), several factors may contribute to heterogeneity: patient selection, characteristics of subjects, study design, different pathological interpretations. Unfortunately, we failed to find the potential sources for heterogeneity because of limited information.
In summary, we showed that miRNAs are promising molecular biomarkers that may significantly improve the diagnostic accuracy of thyroid nodules when using along with FNAB. However, future researches should construct a panel of stably detectable miRNAs and validate the diagnostic value of the panel using a larger cohort of indeterminate FNAB specimens, which could provide highly accurate clinical information for the differential diagnosis of thyroid nodules. Fortunately, miRNAs are comparatively stable in FFPE tissues and can be extracted from FNAB samples, implicating a great opportunity to design large retrospective or prospective studies.
This article is protected by copyright. All rights reserved.
Accepted Article
References 1
Ferlay, J., Shin, H.R., Bray, F. et al. (2010) Estimates of worldwide burden of
cancer in 2008: GLOBOCAN 2008. Int J Cancer, 127, 2893-2917. 2
Pellegriti, G., Frasca, F., Regalbuto, C. et al. (2013) Worldwide increasing
incidence of thyroid cancer: update on epidemiology and risk factors. J Cancer
Epidemiol, 2013, 965212. 3
Hundahl, S.A., Fleming, I.D., Fremgen, A.M. et al. (1998) A National Cancer
Data Base report on 53,856 cases of thyroid carcinoma treated in the U.S., 1985-1995 [see commetns]. Cancer, 83, 2638-2648. 4
Papini, E., Guglielmi, R., Bianchini, A. et al. (2002) Risk of malignancy in
nonpalpable thyroid nodules: predictive value of ultrasound and color-Doppler features. J Clin Endocrinol Metab, 87, 1941-1946. 5
Mehanna, H.M., Jain, A., Morton, R.P. et al. (2009) Investigating the thyroid
nodule. BMJ, 338, b733. 6
Wilmott, J.S., Zhang, X.D., Hersey, P. et al. (2011) The emerging important
role of microRNAs in the pathogenesis, diagnosis and treatment of human cancers. Pathology, 43, 657-671. 7
Martinez, I. & Dimaio, D. (2011) B-Myb, cancer, senescence, and
microRNAs. Cancer Res, 71, 5370-5373. 8
Marini, F., Luzi, E. & Brandi, M.L. (2011) MicroRNA Role in Thyroid Cancer
Development. J Thyroid Res, 2011, 407123. 9
de la Chapelle, A. & Jazdzewski, K. (2011) MicroRNAs in thyroid cancer. J
This article is protected by copyright. All rights reserved.
Accepted Article
Clin Endocrinol Metab, 96, 3326-3336. 10 Schmid, K.W. (2010) [Molecular pathology of thyroid tumors]. Pathologe, 31 Suppl 2, 229-233. 11 Mazeh, H., Levy, Y., Mizrahi, I. et al. (2013) Differentiating benign from malignant thyroid nodules using micro ribonucleic acid amplification in residual cells obtained by fine needle aspiration biopsy. J Surg Res, 180, 216-221. 12 Agretti, P., Ferrarini, E., Rago, T. et al. (2012) MicroRNA expression profile helps to distinguish benign nodules from papillary thyroid carcinomas starting from cells of fine-needle aspiration. Eur J Endocrinol, 167, 393-400. 13 Shen, R., Liyanarachchi, S., Li, W. et al. (2012) MicroRNA signature in thyroid fine needle aspiration cytology applied to "atypia of undetermined significance" cases. Thyroid, 22, 9-16. 14 Whiting, P.F., Rutjes, A.W., Westwood, M.E. et al. (2011) QUADAS-2: a revised tool for the quality assessment of diagnostic accuracy studies. Ann
Intern Med, 155, 529-536. 15 Peng, Y., Li, C., Luo, D.C. et al. (2014) Expression profile and clinical significance of microRNAs in papillary thyroid carcinoma. Molecules, 19, 11586-11599. 16 Vriens, M.R., Weng, J., Suh, I. et al. (2012) MicroRNA expression profiling is a potential diagnostic tool for thyroid cancer. Cancer, 118, 3426-3432. 17 Kitano, M., Rahbari, R., Patterson, E.E. et al. (2012) Evaluation of candidate diagnostic microRNAs in thyroid fine-needle aspiration biopsy samples. Thyroid,
This article is protected by copyright. All rights reserved.
Accepted Article
22, 285-291. 18 Keutgen, X.M., Filicori, F., Crowley, M.J. et al. (2012) A panel of four miRNAs accurately differentiates malignant from benign indeterminate thyroid lesions on fine needle aspiration. Clin Cancer Res, 18, 2032-2038. 19 Mazeh, H., Mizrahi, I., Halle, D. et al. (2011) Development of a microRNA-based molecular assay for the detection of papillary thyroid carcinoma in aspiration biopsy samples. Thyroid, 21, 111-118. 20 Kitano, M., Rahbari, R., Patterson, E.E. et al. (2011) Expression profiling of difficult-to-diagnose thyroid histologic subtypes shows distinct expression profiles and identify candidate diagnostic microRNAs. Ann Surg Oncol, 18, 3443-3452. 21 He, H., Jazdzewski, K., Li, W. et al. (2005) The role of microRNA genes in papillary thyroid carcinoma. Proc Natl Acad Sci U S A, 102, 19075-19080. 22 Berlin, J.A. (1995) Invited commentary: benefits of heterogeneity in meta-analysis of data from epidemiologic studies. Am J Epidemiol, 142, 383-387. 23 Huedo-Medina, T.B., Sanchez-Meca, J., Marin-Martinez, F. et al. (2006) Assessing heterogeneity in meta-analysis: Q statistic or I2 index? Psychol
Methods, 11, 193-206. 24 Moher, D., Liberati, A., Tetzlaff, J. et al. (2009) Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement. J Clin Epidemiol, 62, 1006-1012.
This article is protected by copyright. All rights reserved.
Accepted Article
25 Sipos, J.A. & Mazzaferri, E.L. (2010) Thyroid cancer epidemiology and prognostic variables. Clin Oncol (R Coll Radiol), 22, 395-404. 26 Lee, R.C., Feinbaum, R.L. & Ambros, V. (1993) The C. elegans heterochronic gene lin-4 encodes small RNAs with antisense complementarity to lin-14. Cell, 75, 843-854. 27 Hwang, H.W. & Mendell, J.T. (2007) MicroRNAs in cell proliferation, cell death, and tumorigenesis. Br J Cancer, 96 Suppl, R40-44. 28 Calin, G.A., Liu, C.G., Sevignani, C. et al. (2004) MicroRNA profiling reveals distinct signatures in B cell chronic lymphocytic leukemias. Proc Natl Acad Sci U
S A, 101, 11755-11760. 29 Tong, A.W. (2006) Small RNAs and non-small cell lung cancer. Curr Mol Med, 6, 339-349. 30 Iorio, M.V., Ferracin, M., Liu, C.G. et al. (2005) MicroRNA gene expression deregulation in human breast cancer. Cancer Res, 65, 7065-7070. 31 Akao, Y., Nakagawa, Y. & Naoe, T. (2007) MicroRNA-143 and -145 in colon cancer. DNA Cell Biol, 26, 311-320. 32 Chen, Y.T., Kitabayashi, N., Zhou, X.K. et al. (2008) MicroRNA analysis as a potential diagnostic tool for papillary thyroid carcinoma. Mod Pathol, 21, 1139-1146. 33 Pallante, P., Visone, R., Ferracin, M. et al. (2006) MicroRNA deregulation in human thyroid papillary carcinomas. Endocr Relat Cancer, 13, 497-508. 34 Nikiforova, M.N., Tseng, G.C., Steward, D. et al. (2008) MicroRNA
This article is protected by copyright. All rights reserved.
Accepted Article
expression profiling of thyroid tumors: biological significance and diagnostic utility. J Clin Endocrinol Metab, 93, 1600-1608.
Figure legends: FIGURE1:Flow chart showing the process of study selection.
FIGURE2: (a) Proportion of studies with low, high, or unclear risk of bias; (b) Proportion of studies with low or unclear concerns regarding applicability.
FIGURE3 :The pooled sensitivity (a) ,specificity (b) for miRNAs in discriminating thyroid nodules.
FIGURE4: The overall SROC (a) and Deeks’ Funnel Plot(b).
FIGURE5: The SROC curves of multiple miRNAs(a) and single miRNA(b).
This article is protected by copyright. All rights reserved.
Accepted Article
Table 1 The main characteristics of seven articles included in the meta-analysis First Ye Auth ar or
Cou ntry
Kita no Kita no
20 12 20 12
US A US A
Agre 20 tti 12
Ita ly
Maz eh#1
20 11
Isr ael
47(20/ Lesi 27) ons
You Pen g#2
20 14
Chi na
51(36/ Patie Tissue 15) nts After Resecti on
Keut 20 gen 12
US A
72(22/ Patie Indeter 50) nts minate FNAB
She
20
US
Validatio n Group (cancer/ normal) 59(17/ 42) 24(5/1 9)
Patie nts/ Lesi ons Patie nts Patie nts
Material MicroR NAs
Sta tus
Sensi tivity
Speci ficity
FNAB
miR-7
↓
29%
miR-7
↓
100 % 100 %
miR-14 6 miR-15 5 miR-22 1 miR-21 miR-31 miR-14 6b miR-18 7 miR-22 1 miR-22 2
↑ ↑ ↑
60%
58%
↑ ↑ ↑ ↑ ↑ ↑
50% 45% 80% 50% 95% 90%
miR-30 a-3p miR-14 6b-5p miR-19 9b-5p miR-22 2 miR-32 8 miR-19 7, miR-21 miR-14
↓ ↑ ↑
68% 91% 73%
100 % 100 % 100 % 100 % 100 % 100 % 73% 89% 82%
↑ ↑ ↑ ↑
100 %
Indeter minate FNAB 53(15/ Patie Indeter 38) nts minate FNAB
FNAB after resectio n
68(44/ Patie FNAB
This article is protected by copyright. All rights reserved.
↑
20%
86%
Accepted Article
n
12
A
24)
nts
Slides
↑ 89% 78% 6b ↑ miR-22 ↓ 1 miR-18 7 miR-30 d ↑ She 20 US 30(11/ Patie Indeter miR-14 ↑ 64% 79% n 12 A 19) nts minate 6b ↑ FNAB miR-22 ↓ 1 miR-18 7 miR-30 d ↑ Isr 11(9/2 Patie Indeter miR-21 Maz 20 #2 ↑ 13 ael ) nts minate miR-31 eh ↑ 88% 100 FNAB miR-14 ↑ 6b % ↑ miR-18 ↑ 7 miR-22 1 miR-22 2 F=female; M=male; TG=training group; VG=validation group; NR=not reported; #1 47 samples were obtained from 27 subjects in this study; #2 These two studies were not validation groups.
This article is protected by copyright. All rights reserved.
Accepted Arti
Table 2: The overall diagnostic value of miRNA in thyroid cancer with 95 % confidence intervals
#1:
Results
Sensitivity
Specificity
(95 %CI)
(95 %CI)
AUC
DOR
PLR
NLR
(95 %CI)
(95 %CI)
(95 %CI)
Overall
0.75(0.70-0.80)
Multiple#1
0.87 (0.78-0.93 ) 0.75 (0.67-0.83 ) 0.68 20.3 (2.43-170.16 ) 3.50( 1.40-8.77 ) 0.20(0.05-0.75 )
Single#2
0.71(0.65-0.77)
0.84(0.78-0.88)
0.87 28.9(9.48-88.03)
9.47(2.00-44.80) 0.37(0.27-0.53)
Subgroup#3 0.92(0.58-0.99)
0.68(0.43-0.86)
0.86 25.0(3.00-246)
2.90(1.40-6.10)
0.81(0.77-0.85)
0.89 25.5(9.73-66.67)
The diagnostic value of multiple miRNAs; #2: The diagnostic value of single miRNA. #3: The diagnostic efficacy of miRNA for indeterminate FNAB samples;
This article is protected by copyright. All rights reserved.
5.88(2.48-13.90) 0.33(0.23-0.47)
0.12(0.02-0.83)
Accepted Article This article is protected by copyright. All rights reserved.
Accepted Article This article is protected by copyright. All rights reserved.
Accepted Article This article is protected by copyright. All rights reserved.
