Accurate assessment of mediastinal lymph node involvement in patients with non-small-cell lung cancer (NSCLC) is necessary to select patients for direct surgical treatment. The aims of the present study were to assess the feasibility of staging NSCLC with FDG using a dual-headed positron emission tomographic (PET) camera and to compare this non-invasive technique with computed tomography (CT) and lymph node sampling, since both modalities are currently used for staging NSCLC. Thirty-three patients (29 men and 4 women, mean age 60 years) with newly diagnosed NSCLC were studied. In all patients, CT, FDG dual-headed PET and mediastinoscopy were performed within 4 weeks. The results of mediastinoscopy were used to select patients for thoracotomy. For both the assessment of individual lymph node involvement and the patient-based classification, the results of FDG dual-headed PET and CT were compared using the McNemar test. Thirty-one of 187 lymph nodes studied contained tumour metastases. FDG dual-headed PET showed a significantly higher sensitivity (P < 0.001) and specificity (P < 0.001) than CT. FDG dual-headed PET and CT correctly staged 27 and 20 patients, respectively. Due to the significantly higher negative predictive value of FDG dual-headed PET versus CT (P = 0.012), it was a better non-invasive diagnostic tool for selecting patients for surgery. In seven of eight patients, additional intrapulmonary sites of increased uptake were found, which revealed malignancy on histological examination. CT was false-negative in three of these patients. In one patients, increased FDG uptake was caused by an infection. In conclusion, it is possible to stage mediastinal lymph nodes in patients with NSCLC using a dual-headed PET camera. The high negative predictive value of FDG dual-headed PET suggests that mediastinoscopy may be omitted in patients with NSCLC.