Purpose: We retrospectively investigated the diagnostic performance of 18F-fluorodeoxyglucose positron emission tomography (PET) and PET/CT for cancer detection in asymptomatic health-check examinees. Materials and Methods: This study consisted of 5091 PET or PET/CT conducted as part of annual health examination at one hospital from March 1998 to February 2008. To find the incidence of cancers, medical records of the subjects were thoroughly reviewed for a follow-up period of one year. The patterns of formal readings of PET and PET/CT were analyzed to assess the sensitivity and specificity for cancer detection. The histopathology and stage of the cancers were evaluated in relation to the results of PET. Results: Eighty-six cancers (1.7%) were diagnosed within one year after PET or PET/CT. When PET and PET/CT results were combined, the sensitivity was 48.8% and specificity was 81.1% for cancer detection. PET only had a sensitivity of 46.2% and a specificity of 81.4%, and PET/CT only had a sensitivity of 75.0% and a specificity of 78.5% respectively. There were no significant differences in cancer site, stage and histopathology between PET positive and PET negative cancers. In 19.3% of formal readings of PET and PET/CT, further evaluation to exclude malignancy or significant disease was recommended. Head and neck area and upper gastrointestinal tract were commonly recommended sites for further evaluation. Conclusions: PET and PET/CT showed moderate performance for detecting cancers in asymptomatic adults in this study. More experience and further investigation are needed to overcome limitations of PET and PET/CT for cancer screening.