PURPOSE: Metastasis to the regional lymph nodes is the most important prognostic indicator in patients with colorectal cancer (CRC). The number of lymph nodes examined for adequate staging is still controversial. The aim of this study was to determine if the number of lymph nodes after curative surgery is associated with long-term outcome in patents with Dukes B CRC. METHODS: A retrospective analysis was performed in 174 consecutive patients with Dukes B CRC who underwent curative resection from 1990 to 1999. Patients were stratified according to the number of nodes examined as group A (less than 12 nodes) and group B (12 or more nodes). End-points were local and systemic recurrence and relapse-free survival. Comparisons between the groups were performed by Kaplan-Meier methods and chi-square test as appropriate. RESULTS: There were 115 men (66%). The mean number of nodes examined was 13.4 with the median of 11. No significant difference was found in the number of nodes examined between colon and rectum (16+/-10.6 vs. 13+/-10.0, P=0.675). However, the number of lymph nodes examined tends to be more in recent period of study and if the specimens were examined in the fresh status. With the median follow-up of 44 months, there were 5 local recurrences (2.9%), 22 systemic recurrences (12.6%), and 2 combined local and systemic recurrences (1.1%). Most of the recurrences were observed in group A (79%). The difference of 5-year relapse-free survival rates between the groups was also statistically significant (group A: 73.5%, group B: 91.7%, log-rank test, P=0.0114). The pT stage and number of lymph nodes examined were the independent variables associated with relapse-free survival in multivariate analysis. CONCLUSIONS: The number of lymph nodes examined has prognostic value in patients undergoing curative resection for CRC. Based on our analysis, we recommended at least 12 lymph nodes should be analyzed for accurate staging of CRC.