Clinical assessment is an important part of the breast cancer patients' work-up, but it has low sensitivity and specificity. In a retrospective study, histological slides of axillary clearance specimens were used to model palpability of the axillary lymph nodes. Obvious nodes (enlarged and involving considerable amount of lymphatic and/or metastatic tissue) and nodes equal to or larger than 1 cm or 1.5 cm were counted and the slides were subsequently reviewed. The false positive and negative rates expected on the basis of the model ranged from 24 to 72 % and from 10 to 38 %, respectively. This model (also valid for intraoperative assessment of nodal status by palpation) documents the lack of specificity of clinical staging of the axilla. These results question the practice of excluding patients with palpable axillary lymph node enlargement from less radical staging procedures such as axillary sampling or sentinel node biopsy.