OBJECTIVE: To compare the probability of cancer in a solitary pulmonary nodule using standard criteria with Bayesian analysis and result of 2- [F-18] fluoro-2-deoxy-D-glucose-positron emission tomographic (FDG-PET) scan. SETTING: A university hospital and a teaching Veteran Affairs Medical Center. METHODS: Retrospective analysis of 52 patients who had undergone both CT scan of the chest and a FDG-PET scan for evaluation of a solitary pulmonary nodule. FDG-PET scan was classified as abnormal or normal. Utilizing Bayesian analysis, the probability of cancer using "standard criteria" available in the literature, based on patient's age, history of previous malignancy, smoking history, size and edge of nodule, and presence or absence of calcification were calculated and compared to the probability of cancer based on an abnormal or normal FDG-PET scan. Histologic study of the nodules was the gold standard. RESULTS: The likelihood ratios for malignancy in a solitary pulmonary nodule with an abnormal FDG-PET scan was 7.11 (95% confidence interval [CI], 6.36 to 7.96), suggesting a high probability for malignancy, and 0.06 (95% CI, 0.05 to 0.07) when the PET scan was normal, suggesting a high probability for benign nodule. FDG-PET scan as a single test alone was more accurate than the standard criteria and standard criteria plus PET scan in correctly classifying nodules as malignant or benign. CONCLUSION: FDG-PET scan as a single test was a better predictor of malignancy in solitary pulmonary nodules than the standard criteria using Bayesian analysis. FDG-PET scan can be a useful adjunct test in the evaluation of solitary pulmonary nodules.