RESUMO
PURPOSE: This study was designed to evaluate the reliability of the sentinel node concept in colonic cancer. METHODS: Patent blue was used as tracer. The four blue nodes closest to the tumor were defined as the sentinel node(s) by the pathologist. All nodes were examined by routine microscopy (hematoxylin-eosin staining). If no metastases were detected, all lymph nodes were examined immunohistochemically with antibody to cytokeratin. RESULTS: Two hundred colon specimens were examined. Sentinel node(s) were identified in 93 percent. Sixty contained metastases in hematoxylin-eosin sections. In 32 these were found in sentinel nodes (sensitivity 53 percent). Twenty-eight patients had metastases in nonsentinel nodes only, giving a false-negative rate of 47 percent. Immunostaining revealed 39 (30 percent) micrometastases or submicrometastases in 131 TNM Stages I and II patients, and in 17 of these patients metastases were found in nonsentinel nodes only (false-negative rate 44 percent). CONCLUSIONS: Sentinel lymph node mapping shows low sensitivity for detection of ordinary metastases, micrometastases, and submicrometastases. If only the sentinel nodes had been examined, approximately half of the metastases would have been lost after routine staining, as well as half of the micrometastases and submicrometastases when immunohistochemical examination was added.