RESUMO
PURPOSE: To evaluate the efficacy of FDG-PET in the definition of tumour lung node lesions and to compare the diagnostic validity of CT and FDGPET in the staging of patients with non-small cell lung cancer (NSCLC). MATERIAL AND METHODS: Patients with clinical suspicion of potentially resectable NSCLC (n = 108) were studied by standard procedures in our setting, including fibrobronchoscopy, transthoracic fine-needle aspiration, thoracoabdominal CT and FDG-PET. PET images were analysed by researchers blinded to results of other imaging modalities. Definitive tumour diagnosis was by histopathological study in patients who underwent surgery and by specific imaging methods and biopsy, when available, in patients who did not. Diagnostic accuracy was evaluated by comparing CT/PET results with the definitive diagnosis. RESULTS: In 13% of patients, no FDG-PET findings were observed and the histological study was negative for tumour. In 22% of patients, FDG-PET detected metastatic disease (M0 by CT). For mediastinal involvement, global diagnostic accuracy was 0.90 with FDG-PET and 0.59 with CT. False positive FDG-PET findings were produced by inflammatory conditions and false negative findings by the small size or proximity of lymph nodes to primary tumour. Mediastinal staging by CT and FDG-PET was correct in 56% and 87% of patients, respectively. CT indicated mediastinal invasion in 17% of patients with no FDG-PET finding of mediastinal involvement. Conversely, mediastinal spread was undetected by CT in 14% of patients with FDG-PET findings of mediastinal involvement. CONCLUSIONS: Although complementary, the functional method (FDG-PET) is significantly superior to the structural method (CT) for detection of mediastinal tumour disease.