Sentinel lymph node evaluation does not improve staging accuracy in colon cancer

Background: Lymph node involvement is an important prognostic factor in colorectal cancer. Sentinel lymph node (SLN) evaluation for assessing lymph node status in colorectal cancer remains controversial. Here we evaluated the sensitivity, predictive value, and accuracy of SLN evaluation for determining lymph node status in resectable colon cancer. Methods: A prospective phase 2 cohort study of SLN evaluation in colon cancer was conducted from September 1998 to April 2006. Patients underwent resection and SLN mapping with 1% isosulfan blue and ^m99Tc sulfur colloid injection. SLNs were evaluated by hematoxylin and eosin (HE) staining and, if findings were negative, by additional thin HE sections and immunohistochemical (IHC) staining for pancytokeratin and MOC31. Overall survival for patients with IHC-positive disease was evaluated by Kaplan-Meier analysis and the log rank test. Results: SLNs were identified in 119 (99%) of the 120 patients eligible for the study. Median number of SLNs identified was 4 (range, 0-13). Forty-nine patients (40%) had nodal metastases on HE. The SLN accurately identified nodal metastases in 29 (59%) of these 49 patients and was negative for metastases in 22 patients (41%). SLNs in eight patients (7%) were negative by HE but positive by IHC staining. Positive IHC status did not affect survival after a median follow-up of 33 months (P = .41). Conclusions: The low sensitivity and high false-negative rate of SLN evaluation does not support this technique for improving the accuracy of nodal staging for patients with colon cancer. The significance of IHC-positive SLNs remains uncertain.