| Objective:(1)To explore the current status of the methodological and reporting quality of diagnostic articles about miRNAs for pancreatic cancer;(2)To explore the diagnostic value of miRNAs for pancreatic cancer.Method: Pub Med,Scopus,Web of Science,CBM,CNKI,Wanfang Data and VIP data were retrieved from inception to December 31 th 2015,to collect diagnostic studies about miRNAs for pancreatic cancer.The quality of the articles were independently reviewed by two reviewers and analyzed by Rev Man 5.3.Then Meta-analysis was performed using Meta Di Sc 1.4 and Stata 12.0 software.Results: A total of 40 articles,109 studies,2878 patients with pancreatic cancer and 2269 controls were included.The pooled SEN(sensitivity)was 0.81(95%CI: 0.81,0.82);the SPE(specificity)was 0.78(95%CI: 0.77,0.79);the +LR(positive likelihood ratio)was 3.32(95%CI: 2.90,3.80);the –LR(nagtive likelihood ratio)was 0.27(95%CI: 0.24,0.31;the DOR(diagnostic odds ratio)was 14.56(95%CI: 11.55,18.34);the AUC(area under curve of ROC)was 0.86(95%CI: 0.84,0.88).Subgroups analysis showed that there were significant differences between some groups: caucasian-based studies,miRNA-21-based studies,tissue-based studies and multiple-miRNAs profiling-based assays showed higher diagnostic accuracies than that of their counterparts(AUC=0.89 VS 0.84;0.91 VS 0.84;0.92 VS 0.85;0.92 VS 0.82).Meta-regression showed that whether use multiple miRNAs to diagnose pancreatic cancer may be the source of heterogeneous(RDOR=3.07,95% CI: 1.73-5.44,P<0.001,RDOR: relative DOR).Sensitivity analysis showed the results were reliable.Because when those studies whose sample size <50 and QUADAS scored <7 points were removed from the overall analysis,the pooled estimates were not obviously influenced.Deek’s funnel plot test didn’t exhibit publication bias(Bias=-3.17,P=0.53).The quality of evaluated articles was generally medium,and those English papers were better than the Chinese.The STARD(standards for reporting of diagnostic accuracy)scored 10 to 20,English literature is superior to the Chinese literature(15.07±2.34 VS 12.67±1.82,t=-3.163,P=0.003);The QUADAS(Quality Assessment of Diagnostic Accuracy Studies)scored 5 to 10,English literature is superior to the Chinese literature(7.36±1.10 VS 6.42±0.90,t=-2.613,P=0.013).QUADAS-2 showed mainly defects were as follow: impertinent composition of disease(most papers chose healthy people as control),didn’t use blind method,didn’t determined the threshold in advance,didn’t report the time interval between the diagnostic test and gold standard.The corresponding clinical adaptability of the composition of disease,gold standard,and diagnostic test were high.Conclusion:The meta-analysis suggests that miRNAs have high diagnostic value for pancreatic cancer,particularly using multiple miRNAs.Due to limited quality and quantity of the included studies,more higher quality studies are needed to be down to verify the above conclusion.Researchers should cooperate with statisticians and epidemiologists closer to obtain higher quality studies. |
PDF Full Text Request