OBJECTIVE: To determine whether preoperative anthropometric and computed tomography (CT) measurements of body composition can predict postoperative morbidity and mortality in patients with gastric or esophageal cancer. MATERIALS AND METHODS: This was a retrospective study in which we reviewed the medical records and abdominal CT scans of patients with gastric or esophageal cancer who underwent surgery in 2015 at a cancer center. CT scans performed during routine preoperative evaluation were retrospectively assessed to measure the area of lean body mass at the level of the third lumbar vertebra, as well as the area of visceral and subcutaneous fat. RESULTS: Seventy patients were included in the study. The mean age was 59.9 years (range, 33–82 years), and 47 patients (67.1%) were men. The mean postoperative follow-up period was 14.9 months. Neither postoperative morbidity nor postoperative mortality correlated significantly with gender, age, the type of primary tumor, the presence of comorbidities, smoking status, body mass index, nutritional status, or visceral fat area. The survival rate was higher for patients with normal lean body mass than for those with low lean body mass (hazard ratio = 0.116; 95% confidence interval: 0.015–0.906; p = 0.040). CONCLUSION: Our data suggest that lean body mass can be a relevant prognostic factor in patients with gastric or esophageal cancer, and that CT measurements should be included in the routine preoperative evaluation, because it may provide information that aids nutritional and clinical care for these patients.

Keywords: Gastrointestinal neoplasms; Stomach neoplasms; Esophageal neoplasms; Body composition; Body fat distribution; Tomography, X-ray computed.