Sentinel lymph node (SLN) mapping has evolved into the standard of care for melanoma and may replace routine node dissection in the treatment of breast cancer. There are few data evaluating sentinel node mapping in patients with cancer of the colon. This trial represents our initial experience with SLN mapping for carcinoma of the colon. SLN mapping was performed in 22 patients most of whom had biopsy-proven adenocarcinoma of the colon. One milliliter of isosulfan blue was injected with a 25-gauge needle into the subserosa at four sites around the edge of the palpable tumor. The SLN was identified visually and excised. A standard lymphadenectomy was then performed. The SLN was analyzed with standard hematoxylin and eosin evaluation. Immunohistochemical techniques for carcinoembryonic antigen and cytokeratin (Imm) were performed if the H&E was negative. The mapping added approximately 5 minutes to the total operative time and no adverse reactions to the dye occurred. A SLN was identified in 20 of 22 cases. In cases with negative lymph nodes the SLN was predictive of all the regional nodes by both H&E and Imm (14 of 14). In patients with positive lymph nodes the SLN was predictive in all cases (six of six). In one case the only node with disease was the SLN, and in this case the diease was identified by only Imm; thus this patient was upstaged. SLN mapping is feasible and safe and can readily be performed in patients with colonic cancer. In conjunction with SLN mapping, Imm techniques may upstage a subset of patients likely to be at increased risk for metastatic disease. Consequently SLN mapping of colon cancer should be evaluated in large prospective trials.