Sentinel lymph-node biopsy for breast cancer has been rapidly adopted in clinical practice. At the present time few prospective randomised studies exist, the false negative rate is variable and its role with regard to prognosis is not well known. The aim of this study was to evaluate the elements of sentinel lymph-node biopsy that have yet to be clearly defined, by reference to the literature and our own experience. From September 1999 to July 2005, we considered 235 consecutive patients undergoing sentinel lymph-node biopsy. We used the radioactive tracer in 143 cases, and the radiotracer combined with vital blue dye in 89 cases. Vital blue dye was used alone in only 3 cases. In 224 cases the sentinel lymph-node biopsy was performed in a single session, using frozen sections to evaluate the sentinel node. The identification rate obtained was 97.9% with the radiotracer, 100% with the combined procedure and 66.7% with vital blue alone. The sentinel node proved positive in 52 cases. The frozen sections correctly predicted the positive result of the definitive histopathological analysis in 26 cases and correctly predicted a negative result in 172. We discuss the indications and methods of sentinel lymph-node biopsy, analysing our own data and those reported in the literature.