Solitary pulmonary nodule (SPN, intraparenchymal lung mass <3 cm) is often a diagnostic challenge. This study was performed to evaluate the diagnostic accuracy of( 18)F-fluorodeoxyglucose positron emission tomography (FDG PET) in radiologically indeterminate SPN < or =10 mm on spiral CT. Between August 1997 and March 2001, we identified all patients with radiologically indeterminate SPNs < or =10 mm who were referred for FDG PET imaging at the VU University Medical Centre. All PET scans were retrospectively reviewed by an experienced nuclear medicine physician. PET was considered positive in cases with at least moderately enhanced focal uptake, and otherwise as negative. Lesions were considered benign on the basis of histology, no growth during 1.5 years or disappearance within at least 6 months. Thirty-five patients with 36 SPNs < or =10 mm in diameter at clinical presentation were identified (one patient had two metachronous lesions). In 13 of 14 malignant nodules and in two of 22 benign nodules, diagnosis was confirmed by histology. Prevalence of malignancy was 39%. PET imaging correctly identified 30 of 36 small lesions. One lesion proved to be false negative on PET (CT: 10 mm), and in five lesions, PET scans proved to be false positive. Specificity was 77% (17/22; 95% CI: 0.55-0.92), sensitivity 93% (13/14; 95% CI: 0.66-1.0), positive predictive value 72% (13/18; 95% CI: 0.46-0.90) and negative predictive value 94% (17/18; 95% CI: 0.73-1.0). This retrospective study suggests that FDG PET imaging could be a useful tool in differentiating benign from malignant SPNs < or =10 mm in diameter at clinical presentation. Such results may help in the design of larger prospective trials with structured clinical work-up.