Twenty-five percutaneous lung biopsies using a 20-gauge cutting needle and automated biopsy gun (ABG) were performed under CT guidance in 25 patients with thoracic lesions. This procedure was compared with that using a 21-gauge manual aspiration needle in 36 patients (40 examinations, 37 lesions) in terms of success rate, rate of correct diagnosis, mean examination time and rate of complications. Specimens obtained from lung biopsy were graded by a histopathologist according to quality and quantity from 0 to 4 (pathological score). There were no statistically significant differences between the two procedures in terms of success rate, rate of correct diagnosis and rate of complications; only the time required was significantly different. However, sufficient biopsy material and a mean pathological score of G-II 2.8 (that of G-I was 1.9, p less than 0.05) could be obtained by the biopsy procedure using the cutting needle. The above results indicated that aspiration needle biopsy was adequate for lung biopsy, but that a cutting needle and ABG should be used when a good biopsy specimen is needed for tissue diagnosis.