Although axillary dissection was at one time an important staging procedure for breast cancer, patients were increasingly presenting with small carcinomas, and all too often axillary dissection revealed only healthy lymph nodes. Often too, axillary dissection was associated with postoperative sequelae such as chronic arm lymphedema, pain, and brachial plexopathy. We therefore turned our attention to the axilla, looking for ways to reduce axillary treatment. We studied the ability of sentinel node biopsy to predict axillary status and hence whether it was safe to forgo axillary dissection when the sentinel node was negative. After preliminary studies on radioactive dose and the best tracer substance to use to ensure migration within the lymphatic ducts but efficient retention by the lymph nodes, we recruited 376 consecutive breast cancer patients to a study in which we injected a small quantity of 99Tc-labeled human albumin peritumorally. The next day, the sentinel node was identified and removed surgically (via a small incision) guided by the acoustic signals emitted by a handheld gamma ray–detecting probe. Total axillary dissection followed. The pathologic status of the sentinel node was compared with that of the whole axilla. A sentinel node was identified in 371 patients (98.7%) and correctly predicted the condition of the axilla in 359 (96.8%). Twelve false-negative cases were found among 203 negative sentinel nodes (6.7%).18 Subsequently, the first major study in this area was conducted at the IEO. It randomized 516 patients and compared sentinel node biopsy plus immediate axillary dissection, with sentinel node biopsy plus axillary dissection only if the sentinel node was positive. After more than 5 years of follow-up, no differences between the arms that did and did not receive complete axillary dissection were found, either in terms of axillary recurrences or distant metastases.19 Sentinel node biopsy using a radioactive tracer subsequently became part of the routine treatment of breast cancer patients at the IEO and has since been performed on more than 15,000 patients. The pathologic examination developed to examine the removed sentinel nodes is more exhaustive than normally performed on lymph nodes. This led to the more frequent discovery of metastases, with improved staging accuracy as a consequence. Another result of extensive pathologic examination of sentinel nodes is that micrometastases (<2 mm) and isolated tumor cells are found with greater frequency.20 Current policy at the IEO is to perform complete axillary dissection whenever the sentinel node is macrometastatic. There are 2 main reasons for this. First, metastatic nodes in the axilla may grow and may be inoperable when discovered. Second, the prognosis depends on the number of involved axillary nodes and level of invasion. However, if the sentinel node contains micrometastasis only, it is not clear that complete axillary dissection is necessary. The ongoing trial of the International Breast Cancer Study Group (23-01) is designed to determine the prognostic significance of minimal (<2 mm) metastatic involvement of sentinel nodes in breast cancer; it randomizes patients with minimal involvement to total axillary dissection or no further axillary treatment.