Context.—The pathologic protocol for sentinel lymph node evaluation has yet to be standardized. Results from previous studies are troublesome to compare because they have been conducted on different sets of subjects with cancer.

Objective.—To compare the detection of sentinel lymph node metastases by using step-sectioning methods at 20-μm and 150-μm intervals for the same patient with primary breast cancer.

Design.—A total of 186, initially tumor-negative sentinel lymph nodes from a group of 80 patients with breast cancer were included. For all nodes, each paraffin block was cut serially to produce a total of 10 levels: 5 consecutive levels of sections for each of the 20-μm and 150-μm intervals. The nodal findings obtained at these intervals on hematoxylin-eosin and cytokeratin slides were compared by using the McNemar test.

Results.—The overall detection rate for sentinel lymph node metastasis at intervals of 20 μm and 150 μm was 27.5% (22/80) and 20% (16/80), respectively. The overall agreement between the 20-μm and 150-μm sections was 82.5%. No macrometastasis was missed by either method. At the 20-μm interval, 2 cases of micrometastasis were missed, while 10 cases of isolated tumor cells were missed at the 150-μm interval. However, no statistical difference was observed for the final sentinel lymph node results with either method. (McNemar test, P = .18 for case-based results and P = .052 for nodal-based results).

Conclusions.—The 20-μm and 150-μm interval step-sectioning methods produce comparable results for detection of metastatic deposits in sentinel lymph nodes.