RESUMO
BACKGROUND: Intraoperative frozen sections (FS) of sentinel lymph nodes (SLN) were evaluated to avoid the need for deferred axillary lymph node dissection (ALND) in patients with early breast cancer (EBC). However, FS has low sensitivity for detecting micro-metastases (<2 mm), resulting in patients who later undergo deferred ALND. The aim of the study was to determine the best clinical approach for selecting patients who would derive real benefit from ALND, as well as to minimize the functional and psychological damage caused by delayed surgery, and the risk of undertreating EBC patients. METHODS: This study evaluated 1453 patients with early breast cancer (EBC) who underwent SLN biopsy, FS and definitive evaluation. Causes of discrepancies between SLN biopsy and FS results and the need for further surgery were evaluated. RESULTS: A total of 1226 (86%) patients underwent FS; of these patients, 146 (11.9%) were false negatives. The global sensitivity of FS in detecting both macro and micrometastases was 53.7%. Although ACOSOG Z0011 criteria found that ALND could be avoided in 236 patients, 40 (17%) of these had >3 positive axillary lymph nodes. In contrast, application of the IBCSG 23-10 trial criteria, found that only three patients (3.1%) had >3 positive axillary lymph nodes. CONCLUSIONS: FS has a low sensitivity in detecting micrometastases (19%), but a reasonable sensitivity for macrometastases (75%). Most false negatives were smaller metastases (mean 2.1 mm) and more likely in patients with infiltrating lobular carcinoma. Retrospective modelling of the IBCSG 23-10 criteria reduced the percentage of patients requiring deferred surgery from 12% to 4%. Guidelines recommend irradiation of lymph node drainage stations in patients with ≥4 axillary metastatic lymph nodes. Omission of ALND from 40% of patients who met Z0011 criteria would have resulted in their undertreatment. This risk decreases to 3% by omitting axillary clearing only in patients with micrometastases.