We review our experience of transbronchial fine needle aspiration (TBFNA) over a 3-year period. A total of 112 TBFNAs were performed on 95 patients. Four aspirates were from peripheral lung lesions, 20 from non-ulcerated submucosal infiltrative lesions, 19 from mediastinal abnormalities close to the tracheobronchial tree, and the remaining 69 were for staging of bronchogenic carcinoma with apparent mediastinal lymph node spread, evaluated by chest computed tomography (CT). In the 20 submucosal lesions TBFNA reached a sensitivity of 82.3%, providing the only evidence of a malignant process in five cases. With respect to the 19 mediastinal lesions arising in close proximity to the central airways, TBFNA permitted a diagnosis in cases that would otherwise have required more invasive procedures, although the diagnostic sensitivity of the technique in this group of patients was poor (26%). In the mediastinal staging group, the sensitivity was 76.9%, with no false positive results. Complete sensitivity of TBFNA for the detection of disease was 65.8%. We conclude that TBFNA is a reliable and low risk procedure.