Background: The purpose of the present study was to evaluate the usefulness of intraoperative imprint cytology and frozen sectioning of sentinel lymph nodes in patients with clinically node-negative breast cancer.
Methods: Sentinel node biopsy was performed in 101 patients with stage I or II breast cancer with clinically negative nodes using a dye-guided method. Intraoperative evaluation of sentinel node involvement by imprint cytology and frozen sectioning was compared with the final histopathological results of permanent sections. Tumour-negative nodes in paraffin sections stained by haematoxylin and eosin were further studied using an anticytokeratin antibody.
Results: The results of imprint cytology and frozen-section analysis were compared with those of haematoxylin and eosin-stained sections. The sensitivity, specificity and overall accuracy of imprint cytology were 96.0, 90.8 and 92.1 per cent respectively, and those of frozen-section examination were 52.0, 100 and 88.1 per cent. Ten sentinel nodes were tumour positive on imprint cytology and tumour negative on stained paraffin sections. Micrometastasis was found in eight of these nodes on immunohistochemistry. Taking these immunohistological results into consideration, the final sensitivity, specificity and overall accuracy of imprint cytology were 90.9, 98.5 and 96.0 per cent respectively.
Conclusion: Intraoperative imprint cytology is a useful method for evaluating the status of sentinel nodes and is more accurate than frozen-section analysis. In addition, imprint cytology can detect micrometastasis more accurately than conventional haematoxylin and eosin-stained sectioning.