Sentinel lymph node biopsy (SLNB) is the standard of care for axillary staging in clinically node-negative breast cancer patients. Ultrasound (US) has shown promise when used to assess axillary lymph nodes preoperatively, thus aiding surgical decision making. We examined the correlation between preoperative US and SLNB results to further clarify the role of US in clinicopathologic staging of breast cancer when the axilla is clinically negative on physical examination. Our institutional cancer registry was used to identify clinically node-negative patients diagnosed with breast cancer from January 1, 2009 to December 31, 2012. Variables including age, body mass index, date of surgery, date of diagnosis, US results, US-directed biopsy results, SLNB results, and final pathology were recorded. Incomplete charts were excluded. In all, 249 patients were included. Sensitivity/specificity of US in the clinically negative axilla were 7.4 per cent and 91.8 per cent, respectively. The false-positive rate was 80 per cent, whereas the negative predictive value was 78 per cent. The effect of time from diagnosis/US to SLNB, interpreting radiologist, year in which US was performed, and body mass index were not statistically significant. US in the clinically node-negative patient, although useful when it leads to a positive needle biopsy result, is unlikely to replace SLNB owing to its low sensitivity and a high false-positive rate. Further prospective study into the role of US in the evaluation of the clinically negative axilla is warranted.