Tumor Biol. DOI 10.1007/s13277-014-2544-2

RESEARCH ARTICLE

MicroRNAs as a novel class of diagnostic biomarkers in detection of hepatocellular carcinoma: a meta-analysis Hongmin Yin & Xinguo Peng & Peng Ren & Bo Cheng & Shumin Li & Chengyong Qin

Received: 17 July 2014 / Accepted: 25 August 2014 # International Society of Oncology and BioMarkers (ISOBM) 2014

Abstract MicroRNAs (miRNAs) have been proposed as promising diagnostic biomarkers for many diseases, particularly in the field of cancer research. Numerous studies have explored the use of miRNAs in the detection of hepatocellular carcinoma (HCC), with some reporting inconsistent results. Thus, we conducted this meta-analysis to evaluate the potential diagnostic value of miRNAs in HCC. All relevant literature was collected from the PubMed and other databases before June 3, 2014. The summary receiver operator characteristic (SROC) curve and other parameters were used to estimate overall predictive performance. Fourteen studies involving 1,848 cases with HCC and 1,187 controls (576 healthy controls and 611 individuals with chronic liver diseases) were included in this meta-analysis. SROC analyses for the diagnostic power of miRNAs yielded an area under the curve (AUC) of 0.93 with 86 % sensitivity and 86 % specificity in discriminating patients with HCC from healthy subjects and an AUC of 0.88 with 79 % sensitivity and 83 % specificity in differentiating patients with HCC from those with chronic liver diseases (CLDs). Furthermore, subgroup analyses showed that miRNA panels yielded excellent diagnostic characteristics, with an AUC of 0.99 (96 % sensitivity Electronic supplementary material The online version of this article (doi:10.1007/s13277-014-2544-2) contains supplementary material, which is available to authorized users. H. Yin : C. Qin (*) Department of Gastroenterology, Provincial Hospital Affiliated to Shandong University, 324 Jingwu Weiqi Road, Jinan 250021, China e-mail: [email protected] H. Yin : P. Ren : B. Cheng : S. Li Department of Gastroenterology, Binzhou Central Hospital Shandong Province, Binzhou 251700, China X. Peng Clinical laboratory, Affiliated Hospital of Binzhou Medical College, Binzhou 256603, China

and 96 % specificity) for detection of HCC from healthy controls and an AUC of 0.93 (85 % sensitivity and 88 % specificity) for HCC from those with CLDs. MiRNAs might be novel potential biomarkers for the diagnosis of HCC, and a combination of multiple miRNAs could significantly improve the diagnostic accuracy. Keywords microRNAs . Hepatocellular carcinoma . Diagnosis . Meta-analysis

Introduction Liver cancer is the fifth most frequently diagnosed cancer worldwide in males and the seventh most commonly diagnosed cancer in females, according to the global cancer statistics estimation in 2011 [1]. Among primary liver cancers, hepatocellular carcinoma (HCC) represents the predominant histological subtype and likely accounts for 70–85 % of the total liver cancer burden worldwide [2]. The global risk of HCC has been largely driven by chronic infections of hepatitis B (HBV) and/or C (HCV) viruses over the past century, along with aflatoxin B1, primary biliary cirrhosis, excessive alcohol consumption, non-alcoholic fatty liver disease, carcinogens exposure, and obesity/diabetes [3]. HCC is also highly malignant and lethal, with a poor overall five-year survival rate of 5–9 % from the time of clinical diagnosis [4]. This is primarily due to the fact that a large fraction of HCC cases are detected in advanced stages when curative therapy is no longer feasible and prognosis is thus very poor [5]. The lack of good diagnostic markers for early diagnosis has rendered the disease a major challenge. Therefore, discovering sensitive and specific tools for the early detection of HCC is vital to improving the prognosis of patients with HCC. Currently, clinical diagnostic approaches to HCC are usually based on serologic markers such as aberrantly alpha-

Tumor Biol.

fetoprotein (AFP) and PIVKA-II levels, as well as imaging techniques, including abdominal ultrasound, magnetic resonance imaging (MRI), contrast-enhanced computed tomography (CT) scan, and angiograms [6]. These diagnostic tools have greatly improved the diagnostic accuracy of HCC. However, cross-sectional imaging typically only detects large lesions (greater than 1 cm) and fails to detect small tumors [7]. The accuracy of diagnoses based on AFP (sensitivity 39– 65 %, specificity 76–91 %) and PIVKA-II (sensitivity 41– 77 %, specificity 72–98 %) levels are not satisfying since moderately raised levels are also found in benign liver diseases, such as hepatitis and cirrhosis [8]. Accordingly, there is an urgent need to identify and develop novel non-invasive biomarkers with high sensitivities and specificities that can be used to screen for early stage HCC in at-risk populations, especially in patients with chronic liver diseases (CLDs). Recently, accumulating evidence from both human and animal studies suggests a link between alterations in expression of microRNAs and many types of disease, particularly cancers [9–14]. MicroRNAs (miRNAs) are a class of small (~22 nucleotides in length), endogenous, non-coding RNAs with a fundamental role in the regulation of gene expression [15], and hence control various biological processes, including cellular development, apoptosis, proliferation, differentiation, and tumorigenesis [16]. Since initial observation, over 1,400 human miRNAs have been identified and approximately half of the known miRNAs, localized inside or close to fragile sites, have been found to be associated with the development of human cancers [17]. In addition, the miRNAs are remarkably abundant and stable in human serum/plasma, which laid the foundation for miRNAs as clinical biomarkers in the diagnosis and prognosis of human cancers. Numerous studies have shown that aberrantly expressed miRNAs have been linked to the development and progression of various types of human cancers [18–21]. Moreover, several recent studies have reported the potential application of miRNA expression in liver tissues or serum/plasma samples as predictive and diagnostic biomarkers of HCC [22–24]. However, those studies have produced inconsistent results. The conflicting conclusions of individual studies might be partly due to small sample sizes, inconsonant numbers of valid miRNAs, and different selection criteria for control groups. Some studies investigated the accuracy of one particular miRNA for HCC [7, 25–29], whereas others focused on a combination of multiple miRNAs [30–32]. Some of them recruited controls from healthy population [26, 28, 33, 34], whereas others from individuals with CLDs, such as chronic hepatitis and cirrhosis [27, 31, 35, 36]. Therefore, we conducted this meta-analysis of diagnostic studies from these previous publications to systematically explore the clinical applicability of miRNAs as non-invasive biomarkers in the diagnosis of HCC.

Materials and methods Search strategy and study selection We conducted a literature search in the PubMed, Embase, Chinese National Knowledge Infrastructure (CNKI), and Chinese Biomedical Literature Database (CBM) databases to identify original articles analyzing the diagnostic value of miRNAs in HCC. The last retrieval was performed on June 3, 2014. The medical subject headings (MeSH) and their entry terms employed in the literature search included (1) “liver neoplasms” or “liver cancer” or “carcinoma, hepatocellular” or “hepatocellular carcinomas”, (2) “microRNAs” or “miRNAs” or “primary microRNA” or “circulating microRNAs” or “circulating miRNA”, and (3) “diagnosis” or “sensitivity and specificity” or “receiver operating characteristics” or “ROC curve” or “ROC analyses” or “predictive value”. There were no restrictions in language or publication date. The reference lists of review articles and selected papers were also scanned to identify any additional eligible studies. Selection criteria This meta-analysis was conducted according to the guidelines of the preferred reporting items for systematic reviews and meta-analyses (PRISMA) statement. All included candidate studies had to comply with the following inclusion criteria: (1) studies evaluated the diagnostic value of miRNAs in HCC, (2) all the patients with HCC were confirmed by the standard test (such as histological examinations), and (3) studies reported sample size, sensitivity, specificity, or enough information to reconstruct the diagnostic four-fold contingency table. Accordingly, studies were excluded if they met any of the following criteria: (1) reviews, meeting abstracts, editorials, commentaries, or studies without complete data, (2) unrelated to HCC, (3) focused on survival or prognosis of HCC, (4) studies without comparison groups, and (5) duplicate publications. For studies with overlapping populations or data, only the most complete report was included in this meta-analysis. Retrieved articles were independently screened and carefully evaluated by two investigators. The selection process and inclusion and exclusion outcomes are presented in Fig. 1. Data extraction and quality assessment Two investigators examined the full-texts of eligible publications and extracted relevant data using a standardized table. For each included study, the following information was extracted: (1) study features, including the first author’s surname, publication year, and country, (2) participants’ general features, including ethnicity, type of case and control groups, sample size, gender, and age, (3) data needed for diagnostic meta-analysis, including studied miRNAs, specimen,

Tumor Biol.

Statistical analysis All statistical analyses were performed using Stata 12.0 software. We extracted the numbers of participants with TP, FP, FN, and TN from each study to obtain pooled sensitivity [TP/ (TP+FN)], specificity [TN/(TN+FP)], positive likelihood ratio (PLR), negative likelihood ratio (NLR), diagnostic odds ratio (DOR), and their corresponding 95 % confidence intervals (CIs). Simultaneously, the sensitivity and specificity of each included study were used to plot the bivariate summary receiver operator characteristic (SROC) curve and calculate the area under the SROC curve (AUC) and the maximum point of intersection between sensitivity and specificity (Q value) [38]. AUC is recommended for evaluating the accuracy of a diagnostic test and ranges in scores from 0.5 for a chance discrimination to 1.0 for perfect discrimination [39]. In addition, we explored potential publication bias using Deeks’ funnel plots [40]. All statistical tests were two-sided, and a P

MicroRNAs as a novel class of diagnostic biomarkers in detection of hepatocellular carcinoma: a meta-analysis.

MicroRNAs (miRNAs) have been proposed as promising diagnostic biomarkers for many diseases, particularly in the field of cancer research. Numerous stu...
2MB Sizes 1 Downloads 10 Views