BACKGROUND: Higher levels of physical activity are associated with lower colorectal cancer incidence and mortality, perhaps through influencing energy balance, cellular prostaglandin biosynthesis and systemic inflammation. Although evidence suggests interactive effects of energetics, sedentary lifestyle, and tumor CTNNB1 (beta-catenin) or CDKN1B (p27) status on colon cancer prognosis, interactive effects of physical activity and tumor PTGS2 (the official symbol for cyclooxygenase-2) status on clinical outcome remain unknown. METHODS: Utilizing molecular pathological epidemiology database of 605 stage I-III colon and rectal cancers in two prospective cohort studies (the Nurse's Health Study and the Health Professionals Follow-up Study), we examined patient survival according to post-diagnosis physical activity and tumor PTGS2 status (with 382 PTGS2-positive and 223 PTGS2-negative tumors by immunohistochemistry). Cox proportional hazards models were used to calculate colorectal cancer-specific mortality hazard ratio (HR), adjusting for clinical and other tumor variables including microsatellite instability status. RESULTS: Among PTGS2-positive cases, compared with the least active first quartile, the multivariate HRs (95% confidence interval) were 0.30 (0.14-0.62) for the second, 0.38 (0.20-0.71) for the third, and 0.18 (0.08-0.41) for the fourth quartile of physical activity level (Ptrend=0.0002). In contrast, among PTGS2-negative cases, physical activity level was not significantly associated with survival (Ptrend =0.84; Pinteraction=0.024, between physical activity and tumor PTGS2 status). CONCLUSIONS: Post-diagnosis physical activity is associated with better survival among patients with PTGS2-positive tumors, but not among patients with PTGS2-negative tumors. Impact: Immunohistochemical PTGS2 expression in colorectal carcinoma may serve as a predictive biomarker in pathology practice, which may predict stronger benefit from exercise.