A Prospective, Blinded Trial of Touch Prep Analysis versus Frozen Section for Intraoperative Evaluation of Sentinel Lymph Nodes in Breast Cancer

Background  Sentinel lymph node biopsy (SLNB) has largely replaced axillary dissection (ALND) for axillary staging in early breast cancer. However, intense pathologic evaluation is not routinely available intraoperatively; therefore, patients with SLN metastasis may require a second surgery for completion ALND. We hypothesized that a single-section approach (by either frozen section [FS] or touch preparation analysis [TPA]) could be accurate for intraoperative SLN evaluation. Methods  We performed a prospective, blinded study of patients undergoing SLNB for breast cancer from September 2004 to July 2006. SLNs were bivalved along the long axis, underwent FS and TPA of the facing halves, followed by routine sentinel node processing (serial sectioning with hematoxylin/eosin staining). A single pathologist reviewed all study slides and was blinded to the permanent section interpretation. Results  We analyzed 233 nodes from 118 patients. Overall, 21% of patients (N = 25) had SLN metastasis by serial-section histopathology. Single-section FS and TPA had similar sensitivities (0.67 and 0.66, P = .82) and specificities (0.995 and 0.995, P = 1.0) for detection of SLN metastasis, yielding equivalent accuracies (95%). All micrometastases (<2 mm; N = 4) were missed by both techniques. False positives were rare—only one in each group (2% overall). Conclusion  Single-section TPA and FS have similar accuracies and can be safely used to identify the majority of patients with SLN metastasis, sparing these patients a delayed ALND. False-negative results from TPA or FS occur in patients with micrometastatic disease, for which the role of completion ALND remains controversial.