(18)F-2-deoxy-2-fluoro-glucose Positron Emission Tomography (FDG-PET) has been recently proposed as a promising cancer-screening test. However, the validity of FDG-PET in cancer screening has not been evaluated. We investigated the sensitivity of FDG-PET compared with upper gastric endoscopy in gastric cancer screening for asymptomatic individuals. A total of 2861 consecutive subjects (1600 men and 1261 women) who were asymptomatic and who underwent both FDG-PET and upper gastrointestinal endoscopy between 1 February 2004 and 31 January 2005 were included in this study. Both endoscopists and a radiologist were unaware of the results of the other diagnostic tests. The FDG-PET images were examined using criteria determined by the pattern of FDG accumulation. Sensitivity and specificity of FDG-PET were calculated compared with endoscopic diagnosis as the gold standard. Among 2861 subjects enrolled in the study, there were 20 subjects with gastric cancer, of whom 18 were T1 in depth of cancer invasion. Positive FDG-PET results were obtained only in 2 of the 20 cancer subjects. The calculated sensitivity and specificity for overall gastric cancers were 10.0% (95% confidence interval (CI): 1.2-31.7%) and 99.2% (95% CI: 98.8-99.5%), respectively. (18)F-2-deoxy-2-fluoro-glucose Positron Emission Tomography was poorly sensitive for detection of gastric cancer in the early stages.