Background: The widely studied inflammatory markers, neutrophil lymphocyte ratio (NLR), platelet lymphocyte ratio (PLR) and lymphocyte monocyte ratio (LMR), are associated with survival outcomes in patients with various cancers. However, little is known about the value of these markers in older cancer patients, especially in relation to traditional prognostic factors, such as age and Karnofsky Performance Status (KPS), and a geriatric assessment (GA) derived scale that is predictive of overall survival (OS) [Nishijima, J Geriatr Oncol. in press]. Methods: Our sample includes 144 patients age ≥ 65 years with solid tumor (Carolina Senior Registry NCT01137825) who completed a GA within 3 months of their date of diagnosis in 2010 - 2014 and had pretreatment CBC with differential. Patients were followed for death from any cause for a median of 2.5 years, during which 54 patients died. NLR was dichotomized at the upper limit of normal 3.5 [Forget, BMC Res Notes 2017] while PLR and LMR were dichotomized at the median. The 3-item GA-derived prognostic scale (score ranging 0-3) consisted of (1) "limitation in walking several blocks", (2) "limitation in shopping", and (3) "≥ 5% unintentional weight loss in 6 months". Univariable and multivariable Cox proportional hazards models evaluated whether NLR, PLR and LMR were independently predictive of OS. Results: Median age was 72 years, 53% had breast cancer, 27% had stage 4 cancer, 14% had KPS < 80, 11% received less intensive treatment than standard treatment for stage and 39% had NLR > 3.5. In the univariable survival analyses, higher NLR (hazard ratio (HR) = 5.08, 95% CI; 2.85-9.07, p < 0.001, 2-year OS; 43% vs 86%), higher PLR (HR = 2.10, 95% CI; 1.20-3.67, p = 0.009, 2-year OS; 60% vs 79%) and lower LMR (HR = 2.11, 95% CI; 1.21-3.66, p = 0.008, 2-year OS; 58% vs 80%) were associated with poor OS. Only NLR remained a significant predictor of OS (HR = 2.16, 95% CI; 1.10-4.25, p = 0.025) after accounting for cancer type, stage, age, KPS, treatment intensity and GA-derived prognostic scale. Conclusions: NLR > 3.5 was found to be predictive of poorer OS in older adults with cancer, independent of GA-derived prognostic scale and traditional prognostic factors; further validation in external datasets is warranted.