BACKGROUND: Recent literature has shown that lymph node ratio (LNR) is superior to the number of positive lymph nodes (pLNs) in predicting the prognosis in several malignances other than colon cancer. We hypothesize that LNR may play a similar role in stage III colon cancer. METHODS: We included 24,477 stage III colon cancer cases from the Surveillance, Epidemiology, and End Results cancer registry. Patients were categorized into four groups, LNR1 to 4, according to cutoff points 1/14, 0.25, and 0.50. Kaplan-Meier and Cox proportional hazard model were used to evaluate the prognostic effect and estimate the relative risk (RR) and 95% confidence interval (CI) of LNR. RESULTS: The 5-year survival for patients with stage IIIA, IIIB, and IIIC was 71.3%, 51.7%, and 34.0%, respectively (P < .0001). There was no survival difference among LNR1 to LNR4 for stage IIIA patients. In stage IIIB patients, the 5-year survival for those with LNR1 to LNR4 was 63.5%, 54.7%, 44.4%, and 34.2%, respectively (P < .0001). In stage IIIC patients, the 5-year survival for those with LNR2 to LNR4 was 49.6%, 41.7%, and 25.2%, respectively (P < .0001). LNR is an independent predictor of survival after adjusting patient's age, tumor size, tumor grade, race, number of pLNs, and total number of LNs harvested. (RR 2.30, 95% CI 2.08-2.55). CONCLUSION: Patients with stage IIIB and IIIC colon cancer represent a heterogeneous group of patients with the majority either overstaged or understaged. LNR is a more accurate prognostic method for stage III colon cancer patients. We propose an algorithm to incorporate LNR into current AJCC staging system.