Transbronchial needle aspiration (TBNA) was performed as a diagnostic procedure in 91 consecutive patients ultimately proven to have bronchogenic carcinoma. Results of TBNA were compared, in the same patients, to the diagnostic yield of cytologic examination of sputum, endobronchial brushings and washings, and endobronchial/transbronchial biopsy. The diagnostic yield for sputum was 13 percent (10 of 75); brushings, 40 percent (34 of 84); washings, 29 percent (26 of 89); biopsy, 56 percent (42 of 75); and TBNA, 45 percent (41 of 91). Aspirates were positive in 35 percent of patients with adenocarcinoma, 41 percent with squamous cell carcinoma, 52 percent with large cell undifferentiated carcinoma, and 55 percent of patients with small cell carcinoma. Carinal aspirates were positive in 54 percent (6 of 11); paratracheal aspirates, 57 percent (13 of 23); parabronchial aspirates, 39 percent (11 of 28); endobronchial, 78 percent (7 of 9), and peripheral mass or solitary pulmonary nodule, 40 percent (17 of 42). The overall diagnostic yield for brushings, washings, and biopsy was 64 percent. The addition of TBNA increased the yield to 71 percent. Bronchogenic carcinoma was diagnosed solely by TBNA in six patients, all with extrabronchial or extratracheal lesions. We conclude that TBNA increases the diagnostic yield of bronchoscopy, particularly in patients with extratracheal and extrabronchial lesions. An equally important observation is that TBNA fails to contribute significantly to the diagnosis of cancer in patients with lesions readily accessible by conventional bronchoscopic techniques. Exceptions to this observation include occasional patients with necrotic endobronchial tumors, submucosal lesions, and rarely patients with peripheral lung nodules or masses.