The purpose of this study was to document the accuracy of (18)F-fluorodeoxyglucose ((18)F-FDG) positron emission tomography (PET) with sodium iodide detectors in characterizing indeterminate lung nodules or masses and in identifying additional extra-lesional findings. 50 consecutive patients without a confident diagnosis of malignancy on CT underwent (18)FDG PET with and without attenuation correction. The diagnosis of malignancy was made using visual diagnostic criteria, and tumour-to-blood pool ratios were calculated. The final diagnosis was established by surgery, biopsy or long-term follow-up. Any additional findings made at PET were recorded and similarly verified. Using blinded visual diagnostic criteria for the differentiation of malignant from benign nodules, sodium iodide PET achieved a sensitivity of 91% (30 of 33 cases), a specificity of 88% (15 of 17 cases), a positive predictive value for malignancy of 94% (30 of 32 cases) and a negative predictive value of 83% (15 of 18 cases). False positives occurred with active tuberculosis and sarcoidosis. False negatives were a 3 cm bronchoalveolar carcinoma, a 1.3 cm sarcoma metastasis and a 1 cm carcinoma. Use of tumour-to-blood pool ratios did not improve performance. PET suggested the presence of nodal or distant metastases in 13 of 33 patients with a malignant pulmonary lesion. These PET findings were confirmed in 11 patients. These results indicate that sodium iodide PET is an accurate tool for the characterization of indeterminate pulmonary masses or nodules and simultaneously provides non-invasive staging information that can alter patient management in up to one-third of such patients. Performance of sodium iodide PET is comparable with reported results for PET scanners using other detector materials.