Positron emission tomography with 18F-fluoro-2-deoxy-d-glucose (FDG-PET) is more accurate than computed tomography for staging of mediastinal (hilar) lymph nodes. In the case of positive findings, tissue sampling of lymph nodes is required. The diagnostic/staging yield of transbronchial needle aspiration (TBNA) following endobronchial ultrasound (EBUS) localisation was assessed in this particular clinical setting. The number of avoided surgical procedures was evaluated. All consecutive patients referred for staging and/or diagnosis of mediastinal FDG-PET positive lesions were included. Data were prospectively collected. TBNA sampling of lymph nodes was performed after EBUS localisation. If no diagnosis was reached, further surgical sampling or adequate follow-up was performed. From January 2003 to June 2004, 33 patients were included. The average number of TBNA samples per patient was 4.2+/-1.5. Cytological or histological diagnoses were obtained in 27 (82%) of the patients, of which 78% were obtained after previous EBUS localisation. In 25 (76%) of the 33 patients, surgical staging procedures were suppressed. In conclusion, transbronchial needle aspiration after endobronchial ultrasound localisation should be considered as a primary method of evaluation of lymph nodes positive by positron emission tomography with 18F-fluoro-2-deoxy-d-glucose, and may replace the majority of surgical mediastinal staging/diagnostic procedures.