There are few data about the reliability of sentinel node biopsy in patients with multi-focal breast cancer. The aim of this study was to determine the factors affecting the identification and accuracy of the sentinel node, comparing multifocality with other variables, using peritumoral isosulfan blue dye injection technique alone. Between 1998 and 2001, 122 patients with clinically negative nodes from a single institute were eligible for sentinel lymph node biopsies (SLNBs). All patients underwent conventional axillary lymph node dissection (ALND). SLNs were identified in 111 of 122 (91%) cases, and analyzed by hematoxylin and eosin. Twenty-one patients with multifocal breast cancer were determined by clinical or pathologic examination (gross or microscopic). Success in locating the sentinel node was unrelated to patient's age, tumor size, type, location, histological or nuclear grade, multifocality, or a previous surgical biopsy. SLNBs accurately predicted the status of the axilla in 104 of the 111 patients (93.7%), while 18 of the 21 patients with multi-focal breast cancer (85.7%) had successful lymphatic mapping. The false negative (FN) rate was 11.3% among patients with successful SLNBs. Multifocality and tumor size (>2 cm) were associated significantly with decreased accuracy and increased FN rates (for multifocality, p = 0.007 and p = 0.006, and for tumor size >2 cm, p = 0.04 and p = 0.05, respectively) in binary logistic regression analysis, whereas excisional biopsy, tumor location in the upper outer quadrant and patient's age did not significantly affect the accuracy and FN rates in univariate analysis. These results suggest sentinel lymph node biopsy using peritumoral isosulfan blue injection method alone can accurately predict axillary status in patients with clinically negative nodes, but patients with multi-focal disease and large tumor size may not be ideal candidates.