161 Background: Body composition has emerged as an important prognostic factor in cancer patients. Especially, skeletal mass depletion has been associated with poor performance status, toxicity of chemotherapy and shortened survival in cancer patients. However, the impact of pre-treatment skeletal muscle index on survival or toxicity in metastatic gastric cancer patients remains uncertain. Methods: In this retrospective study, we reviewed 98 metastatic gastric cancer (mGC) patients who received S-1 based combination chemotherapy as first-line treatment from April 2006 to March 2013. Pre-treatment skeletal muscle mass was quantified by CT cross sectional area at the third lumbar vertebrae and evaluated as lumbar skeletal muscle index (SMI) (cm2/m2) after normalization for stature (m2). Patients were categorized into 2 groups depending on initial SMI: 35 patients with SMI ≤ 40 and 63 patients with SMI > 40. Results: Median overall survival was significantly shorter in the SMI ≤ 40 group than in the SMI >40 group (439 days versus 565 days; p= 0.03). Progression free survival was also better in the SMI> 40 group without statistical significance (175 days versus 151 days; p= 0.17). Toxicity (grade 3 or 4) was more common in the SMI ≤ 40 group than in the SMI >40 group. (41.1% versus 11.1%; p=0.001). In multivariate analysis, performance status of 2 (HR 4.711, 95%CI 1.065 to 20.832, p=0.04), presence of primary tumor (HR 2.322, 95%CI 1.007 to 5.357, p=0.04) and pre-treatment SMI (HR 2.525, 95%CI 1.145 to 5.568, p=0.02) were independent prognostic factors for OS. Conclusions: The present study suggests that skeletal muscle depletion at the initiation of first-line chemotherapy might be an independent prognostic factor for mGC patients.