Sentinel lymph node (SLN) biopsy is a widely accepted method for staging breast cancer and melanoma, and it has recently been proposed as a means of improving staging in colorectal cancer. However, lymphatic mapping in colorectal cancer has been plagued by studies demonstrating high false-negative rates. The purpose of this study was to evaluate possible mechanisms for high false-negative rates after SLN biopsy in colorectal cancer. We hypothesized that poor accuracy may be due to bulky tumor or complete replacement of lymph nodes by tumor. Patients with colorectal adenocarcinoma underwent standard colorectal resection with lymphatic mapping. At operation, 1 mL of isosulfan blue dye was injected at the tumor site, using either an in vivo or an ex vivo technique. Routine pathological evaluation was performed. The sentinel node was examined by hematoxylin and eosin stains, and if these results were negative, by cytokeratin immunohistochemistry. The patient's age, operation type, tumor stage, tumor diameter, method of SLN detection, presence of palpable nodes, and pathological description of nodes completely replaced by tumor were recorded. Fifty patients (mean age, 62.8, 50% men) undergoing colorectal cancer resection underwent 51 lymphatic mapping procedures. Right- and left-sided colorectal resections were almost equally distributed (48% vs 42%). SLNs were successfully identified in 47 of 51 specimens (92%). The mean number of SLNs obtained from each specimen was 1.5 (range, 1-5). Routine pathological evaluation demonstrated lymph node metastasis in 20 of the 47 patients (43%) who had an SLN identified. The SLN was positive for metastasis in 10 of these 20 patients (50%). Ten of 20 patients with metastatic disease had a negative SLN, resulting in a false-negative rate of 50%. The false-negative rate was significantly higher in patients undergoing left-sided procedures versus right-sided procedures. Differences among gender, tumor stage, tumor diameter, method of SLN detection, presence of palpable nodes, and pathological description of nodes completely replaced by tumor were not associated with a higher false-negative rate. Identification of the SLN in colorectal cancer is technically possible in more than 90% of patients. However, SLN status correlates poorly with the true nodal status of the colorectal cancer, and the false-negative rate is 50%. This high false-negative rate is not clearly explained by extensive tumor burden, and it was also independent of gender, tumor stage, and type of lymphatic mapping technique. However, staging accuracy was lower in patients who underwent left-sided colorectal resection. Further studies are needed to clarify the limitations of lymphatic mapping in colorectal cancer.