in geriatrics and oncology. For elderly patients with cancer, its impact is significant in terms of quality of life, morbidity and mortality, nevertheless, it is under-diagnosed and under-treated. Our hypothesis is that certain socio-demographic, clinical and treatmentrelated characteristics are associated with depression during chemotherapy and identification would allow early detection of patients needing specific treatment. Objectives: To identify prognostic factors associated with depressive symptoms among cancer patients aged 70 years and older, during chemotherapy. To assess whether the efficacy and safety of chemotherapy are associated with depression. Methods: A regional prospective multicentre cohort, of cancer patients aged 70 years and older, receiving first-line chemotherapy and a comprehensive geriatric assessment at baseline and before the fourth treatment cycle. Depressive symptoms were measured by the Geriatric Depression Scale (GDS). Prognostic and associated factors were sought using a logistic regression model. The complete case analysis was compared with an analysis after multiple imputation, due to missing data. Results:Among the 344 patientswithGDS at the start of chemotherapy, 260 had a second assessment before the fourth treatment cycle. The prevalence of depressive symptoms at fourth cycle was 45%. Independent predictors of depression were malnutrition (OR =5.1, p=0.014), risk of malnutrition (OR= 1.6, p=0.014) and depressive symptoms at baseline (OR = 6.7, p b0.001). Parameter estimates were similar after multiple imputation. Patients who lived alone had a higher risk of depressive symptoms, but this result was not confirmed after multiple imputation. Effective chemotherapy was associated with a lower risk of depression especially after multiple imputation (OR =0.42, p=0.018). Conclusion: We highlight the prognostic role of depressive symptoms and nutritional status at baseline on the occurrence of subsequent depression during chemotherapy, which has not been identified previously. This can be usefull to identify cancer patients at risk for depression and to start as early as possible specific treatments with demonstrated benefits. The role of chemotherapy on depressive symptoms requires further study.