ObjectiveTo evaluate the sensitivity, specificity, and accuracy of magnetic resonance imaging (MRI) in characterizing cirrhosis-related nodules. MethodsThe databases such as the Cochrane Library, PubMed, EMbase were searched on computer from 1998 to 2012.The reviewers screened the trials according to strict inclusion and exclusion criteria, extracted the data, and assessed the methodology quality.Meta-analysis was performed using the metadisc 1.40 software.The acquired pooled sensitivity, specificity, and summary receiver operating characteristic (SROC) curve were used to describe the diagnostic value.The pooled likelihood ratio was calculated based on the pooled sensitivity and specificity. ResultsSix case-control studies involving 917 patients diagnosed with cirrhosis who were suspected to have hepatic nodules were included and 776 masses were confirmed by the biopsy or postoperative histopathology.The pooled statistical results of meta-analysis showed that the sensitivity and specificity of MRI were 87%(83%-89%) and 79%(73%-84%) respectively, the positive and negative likelihood ratios of MRI were 3.95 and 0.18 respectively, and the area under the SROC curve (AUC) was 0.895 6.The sensitivity and specificity of CT were 69%(65%-73%) and 83%(77%-88%) respectively, the positive and negative likelihood ratios of CT were 3.29 and 0.42 respectively, and the AUC was 0.728 7.The sensitivity, positive likehood ratio, and accuracy of MRI in characterizing cirrhosis-related nodules were higher than those of CT. ConclusionAccording these evidences, the MRI should be the first imaging examination for qualitative diagnosis of cirrhosis-related nodules.