In 233 cases in which transthoracic needle aspiration was done at the Mayo Clinic from 1980 through 1983, the cytology slides, tissue fragments and patient histories were reviewed; the original and review diagnoses were compared and correlated with the subsequent clinical course. In most cases, the procedure was performed with an 18-gauge needle under fluoroscopic guidance, primarily in cases with suspected malignant masses that were considered to be not surgically resectable. In 70% of the cases, there was a history of malignancy, and 82% of the malignant lesions were of extrapulmonary origin. Correlation of the original diagnosis with the clinical course yielded 70% (164 cases) true positives, 6% (14 cases) true negatives, 16% (37 cases) false negatives, 0% false positives and 8% (18 cases) indeterminants. In none of the false-negative cases was the slide subsequently read as positive in a blind review. Of the true-positive cases, 12% had positive tissue fragments only, 37% had positive cytology smears only, and 51% had both positive smears and fragments. In 32% of the cases, there were radiologically demonstrable pneumothoraces, and in 12%, placement of a chest tube was required. Hemoptysis occurred in less than 5% of the cases. In summary, transthoracic needle biopsy provides an efficient way to accurately obtain diagnostic tissue, with acceptable minor complications.