PURPOSE: To determine the diagnostic accuracy and frequency of complications of fluoroscopy-guidedtransthoracic needle biopsy, using an automated gun biopsy system. MATERIALS AND METHODS: We retrospectively reviewed 86 patients who underwent automated gun biopsy between October 1995 and October 1996. An 18-gauge cutting needle was used in all cases. All biopsies were performed under fluoroscopic guidance by one interventional radiologist. RESULTS: Tissue sufficient for histologic diagnosis was obtained in 73 of 86 biopsies(84.9%). Fifty-six lesions were malignent and 30 were benign. Sensitivity and diagnostic accuracy for malignant lesions were 87.5% and 87.5%, respectively while cell type specificity in malignant diagnosis was 91.7%(11/12). Sensitivity and diagnostic accuracy for benign lesions were 80.0% and 73.3%, respectively. Postbioptic pneumothorax occurred in three of 86 biopsies(3.5%), one of which required placement of a chest tube. CONCLUSION: Automated gun biopsy is a simple, safe method for the diagnosis of focal chest lesions. An automated biopsy device offers high diagnostic accuracy in cases involving malignant and benign lesions of the chest, and is particularly useful for determining malignant cell type and specific diagnosis of benign lesions.