Recent evidence indicates that miR-17-92 family might be an essential prognostic biomarker for human cancers. However, results are still inconsistent. We therefore performed a meta-analysis to evaluate the predictive role of miR-17-92 family in human cancer prognosis. We searched literatures published before March 31th, 2017 inPubMed, Cochrane and Embase databases. Twenty six studies were included in our analyses. The overall hazard ratios (HRs) showed that high expression level of miR-17-92 family was a predictor of poor overall survival (OS): adjusted HRs = 1.71, 95% confidence intervals (CIs): 1.39-2.11, p < 0.00001, and poor disease-free survival (DFS): adjusted HRs = 2.29, 95% CIs: 1.41-3.72, p = 0.0008. However, no association between miR-17-92 family expression and cancer progress-free survival (PFS) was found (p > 0.05). Subgroup analyses showed that high expression of miR-17-92 family was associated with poor OS (adjusted HRs = 1.89, 95% CIs: 1.43-2.49, p < 0.00001) and DFS (adjusted HRs = 2.83, 95% CIs: 1.59-5.04, p = 0.0003) among the Asian, and no association was found for the Caucasian (p > 0.05). Besides, the HRs of miR-17-92 family high expression in tissue and serum samples was 1.68 (1.35-2.09) and 2.20 (1.08-4.46) for OS, and 1.73 (0.80-3.74) and 3.37 (2.25-5.02) for DFS. It also found that high expression of miR-17-92 family predicted a poor OS in breast cancer, esophageal squamous cell carcinoma, lymphoma and other cancers. Findings suggest that miR-17-92 family can be an effective predictor for prognosis prediction in cancer patients.
Keywords: cancer; meta-analysis; miR-17-92 family; prognosis.