e11546 Background: As many Asian patients want breast conserving therapy (BCT), use of magnetic resonance imaging (MRI) increase in preoperative diagnosis for breast cancer. But the impact of MRI on these patients has not been unclear. Methods: From January 2008 to July 2008, 423 patients underwent breast cancer surgery in National Cancer Center, Korea. We enrolled 357 patients consecutively in this retrospective study; 290 patients (non-MRI group) with preoperative mammography (MMG) and ultrasonography (US) vs. 66 patients (MRI group) with additional MRI to MMG, US and excluded 67 patients (42 patients with preoperative chemotherapy, 8 patients with ipsilateral recurrence, 17 patients whose MRI showed no residual lesion after excisional biopsy). We examined MRI effect on mastectomy rate, intraoperative conversion from BCT to mastectomy, positive margin rate in frozen specimen in both group. In MRI group, we evaluated the correlation between tumor size on US, MRI and pathologic tumor size. Results: Mean age of this study was 48.89 years (Non-MRI group: 50.70 years vs. MRI group: 46.33 years, p=0.001). The rate of mastectomy wasn’t different in both groups (Non-MRI group: 13.7% vs. MRI group: 19.4%, p=0.252). Intraoperative conversion to mastectomy was performed frequently in MRI group. (Non-MRI group: 1.7% vs. MRI group: 7.5%, p=0.023). But positive margin rate in frozen specimen was similar in both groups (Non-MRI group: 23.2% vs. MRI group: 34.0%, p=0.111). In MRI group, mean tumor size on MRI, US was 3.07cm, 1.98cm respectively. Mean pathologic tumor size was 2.67cm. The tumor size on MRI correlated strongly with the pathologic tumor size. The correlation coefficient was 0.732 (p=0.0001). But the tumor size on US didn’t correlate with the pathologic tumor size (p=0.066). In twenty nine patients whose MMG showed suspicious microcalcification, tumor size on MRI also correlated strongly with pathologic tumor size. The correlation coefficient was 0.693 (p=0.0001). But US didn’t show the correlation with the pathologic tumor size in these patients. Conclusions: Preoperative breast MRI didn’t give the impact on breast cancer surgery in Asian patients and could overestimate the size of tumor. But it could strongly correlate with the pathologic tumor size in Asian patients. No significant financial relationships to disclose.