MicroRNAs level as an initial screening method for early-stage lung cancer: a bivariate diagnostic random-effects meta-analysis.

Accumulating studies suggested that microRNAs (miRNAs) can have high diagnostic value as a non-invasive and cost-effective procedure with high sensitivity and specificity in the detection of early-stage lung cancer. However, there is inconsistency observed in the results of relevant studies. Therefore, we performed this meta-analysis to evaluate diagnostic value of miRNAs based on all related studies. A total of 38 studies from 13 included articles were used for the analysis, consisting of 510 patients and 465 healthy controls. All analyses were performed on the R 3.2.0 software. The bivariate random-effects meta-analysis model was applied to obtain the following pooled parameters: sensitivity, 0.797 (95% CI: 0.756-0.832); false positive rate, 0.296 (95% CI: 0.250-0.346); and AUC, 0.818. In addition, subgroup analyses were conducted, showing not only that a combination of multiple miRNAs as biomarkers have greater diagnostic value for early-stage lung cancer (sensitivity, false positive rate and AUC of 83%, 25.2% and 0.858, respectively) had a higher diagnostic accuracy than single miRNA (sensitivity, false positive rate and AUC of 78.3%, 31.6% and 0.799, respectively), but also that specimen from circulating system (sensitivity, false positive rate and AUC of 82.5%, 30.5% and 0.836, respectively) provide better biomarkers than specimen from non-circulating system (sensitivity, false positive rate and AUC of 73.8%, 26.5% and 0.796, respectively). In summary, the current meta-analysis suggests that miRNAs as biomarkers, particularly a combination of multiple tumor-specific miRNAs from circulating system, have moderately high clinical diagnostic value in the detection of early-stage lung cancer. However, the clinical diagnostic utilization and additional improvements of miRNAs as biomarkers for early-stage lung cancer detection still remain to be further validated by more future studies.