OBJECTIVE We have developed a new CT-guided technique using puncture site-down positioning during the biopsy. The goal of our study was to determine the efficacy and safety of this technique for biopsy of lung lesions compared with the standard technique. METHODS Medical records of 236 patients who underwent CT-guided transthoracic needle biopsy were retrospectively evaluated. This study included 89 cases that were biopsied using the standard technique (group A) and 147 cases that were biopsied using the puncture site-down positioning technique (group B). A 20-gauge automated cutting needle without coaxial technique was used in all patients. Medical records were reviewed for lesion size and location, biopsy results, and complications. RESULTS When using the standard technique, the sensitivity for malignant lesions was 96.1%; the sensitivity for benign lesions, 92.1%; and diagnostic accuracy, 94.4%. Thirty-seven patients (41.6%) had pneumothorax, with 16 (18.0%) requiring chest tube placement. When using the puncture site-down positioning technique, the sensitivity for malignant lesions was 95.4%; the sensitivity for benign lesions, 93.3%; and diagnostic accuracy, 94.6%. Nineteen patients (12.9%) had pneumothorax, with four (2.7%) requiring chest tube placement. Other complications were minimal. CONCLUSIONS CT-guided transthoracic needle biopsy using the puncture site-down positioning technique is an effective and safe procedure with a high diagnostic accuracy and low complication rate. This new technique is especially useful in reducing the rate of pneumothorax.