BACKGROUND: The causes of solitary pulmonary nodule are many, but the main concern is whether the nodule is benign or malignant. Because a solitary pulmonary nodule is the initial manifestation of the majority of lung cancer, accurate clinical and radiologic interpretation is important. Bayes' theorem is a simple method of combining clinical and radiologic findings to estimate the probability that a nodule in an individual patients is malignant. We estimated the probability of malignancy of solitary pulmonary nodules with a specific combination of features by Byesian approach. METHOD: One hundred and eighty patients with solitary pulmonary nodules were identified from multi-center analysis. The hospital records of these patients were reviewed and patient age, smoking history, original radiologic findings, and diagnosis of the solitary pulmonary nodules were recorded. The diagnosis of solitary pulmonary nodule was established pathologically in all patients. We used to Bayes' theorem to devise a simple scheme for estimation the likelihood that a solitary pulmonary nodule is malignant based on radiological and clinical characteristics. RESULTS: In Patients characteristics, the probability of malignancy increases with advancing age, peaking in patients older than 66 year of age(LR : 3.65), and higher in patients with smoking history more than 46 pack years(LR : 8.38). In radiological features, the likeklihood ratios were increased with increasing size of the nodule and nodule with lobulated or spiculated margin. CONCLUSION: In conclusion, the likelihood ratios of malignancy may improve the accuracy of the probability of malignancy, and can be a guide of management of solitary pulmonary nodule.