ObjectiveTo systematically review the accuracy of serum ferritin (SF) for detecting breast cancer. MethodsWe electronically and comprehensively searched databases including CNKI, WanFang Data, CBM, VIP, PubMed, EMbase and The Cochrane Library (Issue 5, 2013) up to April 2013, for diagnostic tests about using SF for detecting breast cancer. Four reviewers independently screened literature according to the exclusion and inclusion criteria, extracted data, and assessed methodological quality of included studies. Then, meta-analysis was performed using Meta-Disc 1.4 software and funnel plots were drawn using Stata 12.0 software. ResultsA total of 19 studies were finally included involving 2 977 patients. The results of meta-analysis showed that:Sen, Spe, +LR, -LR, DOR were 0.51 (95%CI 0.48 to 0.53), 0.91 (95%CI 0.89 to 0.92), 5.32 (95%CI 3.72 to 7.60), 0.45 (95%CI 0.35 to 0.57), and 13.22 (95%CI 7.22 to 24.18); SROC area under the curve (AUC) was 0.920 5 and Q* was 0.853 9. Besides, when cut off value was 101-150 μg/L, Sen, Spe, AUC and Q* were the largest, and the best cut-off value was probably 150 μg/L. ConclusionSF has relatively high Sen and Spe in the diagnosis of breast cancer which could not be used as specific index. Heterogeneities exist among the research results, which are possibly associated with researchers, severity of disease, instrument types, sources of reagents, and cut off values. Due to limited quality and quantity of the included studies, the above conclusion should be treated with caution, and in clinic, combing other tumour markers with SF is recommended in detecting breast cancer to augment accuracy.