Histopathologic validation of the sentinel lymph node hypothesis for breast carcinoma

Abstract

Background and objective: The sentinel node hypothesis assumes that a primary tumor drains to a specific lymph node in the regional lymphatic basin. To determine whether the sentinel node is indeed the node most likely to harbor an axillary metastasis from breast carcinoma, the authors used cytokeratin immunohistochemical staining (IHC) to examine both sentinel and nonsentinel lymph nodes.

Methods: From February 1994 through October 1995, patients with breast cancer were staged with sentinel lymphadenectomy followed by completion level I and II axillary dissection. If the sentinel node was free of metastasis by hematoxylin and eosin staining (H&E), then sentinel and nonsentinel nodes were examined with IHC.

Results: The 103 patients had a median age of 55 years and a median tumor size of 1.8 cm (58.3% T1, 39.8% T2, and 1.9% T3). A mean of 2 sentinel (range, 1-8) and 18.9 nonsentinel (range, 7-37) nodes were excised per patient. The H&E identified 33 patients (32%) with a sentinel lymph node metastasis and 70 patients (68%) with tumor-free sentinel nodes. Applying IHC to the 157 tumor-free sentinel nodes in these 70 patients showed an additional 10 tumor-involved nodes, each in a different patient. Thus, 10 (14.3%) of 70 patients who were tumor-free by H&E actually were sentinel node-positive, and the IHC lymph node conversion rate from sentinel node-negative to sentinel node-positive was 6.4% (10/157). Overall, sentinel node metastases were detected in 43 (41.8%) of 103 patients. In the 60 patients whose sentinel nodes were metastasis-free by H&E and IHC, 1087 nonsentinel nodes were examined at 2 levels by IHC and only 1 additional tumor-positive lymph node was identified. Therefore, one H&E sentinel node-negative patient (1.7%) was actually node-positive (p < 0.0001), and the nonsentinel IHC lymph node conversion rate was 0.09% (1/1087; p < 0.0001).

Conclusions: If the sentinel node is tumor-free by both H&E and IHC, then the probability of nonsentinel node involvement is <0.1%. The true false-negative rate of this technique using multiple sections and IHC to examine all nonsentinel nodes for metastasis is 0.97% (1/103) in the authors' hands. The sentinel lymph node is indeed the most likely axillary node to harbor metastatic breast carcinoma.