The value of frozen sections in the intraoperative examination of sentinel nodes (SN) remains controversial. Accurate frozen sections will spare those patients with node metastasis from a second procedure to complete the axillary dissection. We examined our own experience with intraoperative examination of SN. Between January 1, 2006, and December 31, 2006, we performed 236 sentinel lymph node biopsy procedures that were read as "frozen-section-negative." An additional 47 sentinel lymph node biopsy patients were frozen-section-positive for metastatic disease and underwent immediate completion axillary dissection. At least 1 SN was found in all 283 women (100%). The number of patients with false-negative frozen sections was tallied; patient data were reviewed for a number of variables to see which factors might be associated with a false-negative result. Eleven patients had positive nodes on subsequent examination of the formalin-fixed, hematoxylin and eosin-stained slides; the false-negative rate of intraoperative frozen section was 4.7%. The sensitivity of the negative frozen section was > 95%. The following variables were compared for significance: pathologist, nuclear grade, histologic grade, margins, lymphovascular invasion, tumor type (ductal versus lobular), and estrogen receptor and progesterone receptor values. The only significant variables were lymphovascular invasion (p = 0.019) and presence of in situ ductal carcinoma (p = 0.001). Our data confirm the value of intraoperative examination of SN: > 95% sensitivity. Presence of in situ ductal carcinoma or lymphovascular invasion makes these tumors more likely than others to have micrometastases to SN overlooked.