The aim of the study was to describe our experience with sentinel lymph node biopsy in patients with breast cancer. 326 consecutive patients with breast cancer operated on from December 1998 to December 2002 were studied. All patients gave their informed consent. Patients were mapped with an intradermal injection of Tc-99 (median dose: 0.5 mCi) and/or Patent Blue. Sentinel lymph nodes were analyzed with serial sections. 333 procedures were performed in the 326 patients. A median of two sentinel lymph nodes were identified in 322 cases (97%). 3165 additional non-sentinel lymph nodes were removed and analyzed to assess the accuracy of the technique. The correlation between sentinel lymph nodes and final pathological status was 97% (314/322). In 66/133 cases with axillary metastases (50%) the sentinel lymph node was the only site of metastasis. Micrometastases were diagnosed in 35/66 cases (26%), while isolated tumour cells were found in 15 cases (11%). At a median follow-up of 21 months one patient presented an axillary relapse (0.3%). Our experience confirms that sentinel lymph node biopsy is accurate and reproducible. Routine axillary dissection is no longer the gold standard in patients with early breast cancer. Prospective studies are under way in an attempt to provide, definitive answers.