Objective To evaluate the sensitivity， specificity, and accuracy of magnetic resonance imaging （MRI） in characterizing adnexal masses. Methods The databases such as the Cochrane Library, PubMed， EMbase， CNKI， and WanFang Data were searched on computer from 1991 to 2011. The reviewers screened the trials according to inclusion and exclusion criteria strictly, extracted the data， and assessed the methodology quality. Meta-analysis were performed using the Metadisc 1.40 software. The acquired pooled sensitivity, specificity， and summary receiver operating characteristic curve （SROC） were used to describe the diagnostic value. The pooled likelihood ratios were calculated based on the pooled sensitivity and specificity. Results Ten case-control studies involving 649 women who were suspected to have pelvic masses were included and 729 masses were confirmed by the postoperative histopathology. The pooled statistical results of meta-analysis showed that：the sensitivity and specificity of MRI were 〔89%（84%-92%）， P=0.046 6〕 and 〔87% （83%-90%）， P=0.000 2〕 respectively, the positive and negative likelihood ratios of MRI were 6.25（P=0.008 5） and 0.14（P=0.029 1） respectively, and the area under the SROC curve （AUC） was 0.941. The sensitivity and specificity of ultrasound were 〔87%（82%-91%）， P=0.000 0〕 and 〔73%（69%-77%）, P=0.000 0〕 respectively, the positive and negative likelihood ratios of MRI were 3.07（P=0.000 0） and 0.18（P=0.000 1） respectively， and the AUC was 0.897. The speci?city and accuracy of MRI in characterizing female pelvic masses were higher than ultrasound obviously. Conclusion According these evidences， the MRI should be recommended to the women who are suspected to have pelvic masses as a preferred.
