Diagnostic rather than prognostic markers-relationship between EpCAM overexpression and lung cancer: a meta-analysis.

Background: Epithelial cell adhesion molecule (EpCAM) is one of the most commonly used markers of cancer stem cells (CSCs). However, the diagnostic and prognostic significance of EpCAM in lung cancer remains largely undetermined. In the present study, we systematically summarized and elucidated the correlation between EpCAM overexpression and lung cancer through a meta-analysis.

Methods: Six databases (PubMed, Web of Science, Cochrane Library, and Embase, CnKI and Wanfang Database) were systematically searched. Quality Assessment of Diagnostic Accuracy Studies-2 (QUADAS-2) criteria were adopted to assess the qualities of the included studies. Relevant data were extracted for meta-analysis using the Stata12.0 software. Unadjusted mixed odds ratios (ORs) or hazard ratios (HRs) with 95% confidence interval (95% CI) were estimated to evaluate the correlation between EpCAM overexpression and lung cancer. The sensitivity and specificity of the included studies were used to construct the summary receiver operator characteristic (SROC) curve and calculate the area under the SROC curve (AUC).

Results: A total of 14 studies consisting of 2,658 lung cancer patients were included following the PICOS principle. We found that the EpCAM expression was significantly higher in lung cancer patients compared with normal controls, including patients with benign pulmonary diseases (OR =63.71, 95% CI, 14.59-278.21, P=0.003) and healthy individuals (OR =520.08, 95% CI, 16.38-16,510.80, P=0.002), and its overexpression was negatively associated with the TNM stage (III + IV) (OR =0.41, 95% CI, 0.21-0.82, P=0.073. The combined sensitivity and specificity of EpCAM overexpression in the diagnosis of lung cancer were 0.79 (95% CI, 0.59-0.90) and 0.98 (95% CI, 0.95-0.99), respectively, and the SROC-AUC was 0.98 (95% CI, 0.97-0.99). Multivariate analysis of 322 lung cancer patients showed that there was no significant correlation between the EpCAM overexpression and prognosis of lung cancer (HR =2.28, 95% CI, 0.80-6.51, P=0.002). Deeks' funnel plot analysis showed the existence of publication bias (P=0.000).

Conclusions: Our present findings suggested that EpCAM overexpression was not sensitive enough to predict the prognosis of lung cancer. Moreover, it was also a potential diagnostic indicator for lung cancer and correlated with TNM staging of lung cancer.