Background: Long-chain omega-3 fatty acids (LCN-3 FA) may prevent colon cancer (CC) progression, but data in humans are lacking. We hypothesized that LCN-3 FA and dark meat fish intake after CC diagnosis would be associated with longer disease-free survival (DFS), particularly among patients (pts) whose tumors expressed COX2. Methods: We conducted a prospective study among 1011 stage III CC pts enrolled in an adjuvant chemotherapy trial in 1999-2001. Intakes of dark meat fish; canned tuna; breaded fish; shrimp, lobster, scallops, clams; other fish; and fish oil supplements were assessed via food frequency questionnaire during and 6 months (mo) after therapy. Less than 5% of pts reported fish oil supplements. We estimated hazard ratios (HR) and 95% confidence intervals (CI) for disease-free, recurrence-free (RFS), and overall (OS) survival, adjusting for clinical, sociodemographic, and lifestyle factors. In a subset of 510 pts, we examined whether tumor COX2 expression modified the association between LCN-3 FA intake and DFS. Results: We observed 343 CC recurrences and 305 deaths (7 y median follow-up). Higher LCN-3 FA intake after CC diagnosis was associated with 28% improved DFS [HR highest (Q4) vs. lowest (Q1) quartile: 0.72 (0.54, 0.97); p: 0.03]. The association was driven by CC recurrence [RFS - HR Q4 vs. Q1: 0.68 (0.50, 0.94); p: 0.01; OS - HR Q4 vs. Q1: 0.79 (0.56, 1.13); p: 0.17]. Among pts with moderate/high COX2 expression tumors, LCN-3 FA intake was associated with 80% improved DFS [HR Q4 vs. Q1: 0.2 (0.06-0.76); p: 0.006]; there was no association among pts with absent/low COX2 expression (p-interaction: 0.13). When we examined dietary sources of LCN-3 FA, pts who consumed dark meat fish ≥ 2 times/mo vs. none had longer DFS [HR: 0.64 (0.48, 0.86); p: 0.007), RFS [HR: 0.61 (0.44, 0.83); p: 0.005], and OS [HR: 0.68 (0.48, 0.97); p: 0.05]. No other fish or seafood after CC diagnosis was associated with DFS, RFS, or OS. Conclusions: LCN-3 FA and dark meat fish intake after CC diagnosis may be associated with improved DFS, especially among pts with high COX2 expression tumors. Support: K07CA197077, U10CA180821, U10CA180882. ClinicalTrials.gov: NCT00003835.