Abstract Objective: We evaluated whether preoperative serum levels of matrix metalloproteinase-2 (MMP-2) work as a prognostic biomarker in breast cancer prognosis. Methods: Three hundred and three women with histologically confirmed breast cancer were recruited. The follow-up time for all patients was 4.24 years. The MMP-2 levels were quantitatively measured by enzyme-linked immunosorbent assay (ELISA) using the preoperative serum. The relationship of MMP-2 to survival was investigated using Cox's proportional hazard model adjusted for the TNM stage and estrogen receptor (ER) status. Results: In the multivariate analysis, disease-free survival (DFS) was worse among patients with the third tertile of MMP-2 compared to the first tertile of MMP-2 (hazard ratio (HR)=1.80, 95% confidence interval (CI)=1.04-3.11, P=0.04). Furthermore, when the patients were stratified by histological grade and nuclear grade, the worse DFS was predicted by high levels of MMP-2 (HR=2.90 and 95% CI=1.42-5.92 in histological grade III vs. I-II and HR=2.61 and 95% CI=1.26-5.39 in nuclear grade III vs. I-II). In ER negative patients, high levels of MMP-2 also tended to have a worse prognosis (HR=2.75 and 95% CI=1.32-5.73). Conclusions: Our results suggest that the preoperative serum levels of MMP-2 were associated with the survival of breast cancer as a potential prognostic biomarker. Citation Format: {Authors}. {Abstract title} [abstract]. In: Proceedings of the 103rd Annual Meeting of the American Association for Cancer Research; 2012 Mar 31-Apr 4; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2012;72(8 Suppl):Abstract nr 4494. doi:1538-7445.AM2012-4494