Intraoperative Histopathologic Assessment of Sentinel Lymph Nodes in Breast Cancer: An Evaluation of Institutional Protocol

Abstract

Frozen section (FS) for evaluation of sentinel lymph nodes (SLNs) can be used to detect metastasis. However, there is no standard protocol concerning histopathologic analysis of SLNs on FS. Diverse methods have been proposed. Nonetheless, several issues such as cost, potential loss of micrometastasis, and sampling error, among other factors arose. This study highlights our institutional protocol in evaluating SLNs on FS. Statistical analyses of collected data are enumerated. Retrospective analysis of 519 SLNs (219 primary breast cancer patients), who underwent breast surgery with FS of SLNs were examined. The study was conducted at Saint Barnabas Medical Center, Livingston, New Jersey during a 1-year period. Per protocol, SLNs greater than 1 cm are serially sectioned perpendicular to long axis; alternate pieces, including tips, are submitted on FS; and the remainder submitted for permanent microscopic analysis. If nodes are 1 cm or less, they are bisected exposing the greatest surface; half is submitted for FS and the remaining half is for permanent section. Suspicious area noted on gross examination is submitted for FS. Two step sections of hematoxylin and eosin-stained tissue are analyzed on FS. The entire lymph node is not submitted for FS to avoid potential loss of micrometastasis. Residual lymph nodes after FS submitted entirely for definitive microscopic examination employing one step section of H&E-stained slide. Lymph nodes that are negative on FS are reflexively stained with pan-cytokeratin to detect false negatives. Tissues are processed according to defined standards, slide-mounted tissues are examined by experienced pathologists. Overall, intraoperative FS analysis of 519 SLNs revealed sensitivity of 72.73% and specificity of 100%. Separate statistical analysis revealed sensitivity of 23.53% for micrometastasis compared to 95% sensitivity for macrometastasis. The specificity for identifying macrometastasis is 100%. Our protocol revealed very good sensitivity and an excellent specificity.