21171 Background: Background: Axillary lymph node metastasis is the most significant marker for the pathologic staging of breast cancer. However a proportion of lymph node-negative breast cancer will develop metastatic disease. Therefore, molecular markers of invasion in these patients are needed Methods: We selected 10 primary breast cancer cases, 5 lymph node-negative (T1N0) and 4 lymph node-positive (T1N1). In these cases we searched for gene expression of 8 genes (bbc3, cegp1, fgf18, flt1, cffm4, gstm3, hec, tgfb3) selected from previous studies as a good candidates for metastasis prediction. A quantitative Real-Time PCR was performed using beta-2- microglobulin gene expression to normalized gene expression of each gene. The expression average of beta-2-microglobulin was 303,291 among T1N0 and 342,533 among T1N1 cases (ratio 0.88). The expression average of these 8 genes was 2.11 and 0.002 at T1N0 and T1N1 respectively (ratio 1,055). At least 3 genes were significantly down-regulated in T1N1 (bbc3, flt1, gstm3) in comparison with T1N1 breast carcinomas Results: Computational analysis reveals that these 3 genes (bbc3, flt1, gstm3) contain CpG islands in their promoter regions. Although preliminary data, we found group of genes that are down-regulated in T1N1 tumor Conclusions: The finding that these genes are epigenetically regulated, suggest that could be a good candidates for metastasis prediction, by methylation approach (MethyLight System). Since this down-regulation occurs in primary tumor, the analysis of these genes might be useful to predict metastasis in lymph node-negative breast cancer. No significant financial relationships to disclose.