Background: The obesity paradox –i.e. inverse associations between body mass index (BMI) and mortality – has been reported in patients with cancer, heart failure, and diabetes. However, the influence of cardiorespiratory fitness (CRF) on this relationship is not well established. This study assesses the association of BMI and CRF with all-cause mortality among cancer patients. Methods: The Henry Ford (HF) FIT Project is a retrospective cohort study of 69,885 consecutive patients who underwent physician-referred exercise stress testing from 1991 through 2009. Cancer diagnosis was identified through linkage to the HF tumor registry. We included patients 40-70 years old, with BMI recorded, at time of exercise test, with a history of cancer > 6 months prior. BMI was categorized as normal (18.5-24.9kg/m²), overweight (25-29.9kg/m²), or obese ( > = 30kg/m2). All-cause mortality was obtained from the National Death Index. Because of a significant interaction between BMI and cancer type, patients with breast or prostate cancer were excluded. Multivariable adjusted Cox proportional hazard models were used to evaluate the association of CRF andBMI with all-cause mortality; adjusted for age at exercise test, sex, diabetes, smoking, cancer stage, and time from cancer diagnosis to exercise test. Results: Included were 676 patients with a mean age of 58 years (SD 7.5), 51% female, 70% White, 25% Black, with a median of 4.8 years from diagnosis to exercise test and median follow up time of 10.3 years. Among patients achieving < 10 METs, those who are overweight and obese had a lower risk of mortality HR 0.47 (95% CI 0.25,0.86) and HR 0.44 (95% CI 0.26, 0.74, respectively), compared to those with normal BMI. Among patients with METs > = 10, those who were overweight had the lowest risk of all-cause mortality (HR 0.23, 95% CI 0.09-0.62) compared to normal weight, while no statistically significant different risk of mortality was observed when comparing those who are obese to normal weight (HR 0.37, 95% CI 0.13-1.06). In an analysis combining BMI and fitness groups (four categories), those with BMI > = 25 and METs > = 10 had the lowest risk of all-cause mortality (Table). Conclusions: In non-breast/non-prostate cancer patients, increased BMI is associated with improved overall survival in those with METs < 10, while a U-shaped relationship between BMI and all-cause mortality exists among those with METs > = 10.