STUDY DESIGN: This retrospective study is to confirm the diagnostic accuracy of percutaneous needle biopsy of vertebral lesions. OBJECTIVES: To evaluate the diagnostic accuracy of percutaneous needle biopsy and the clinical and pathological correlation of vertebral lesions. SUMMARY OF LITERATURE REVIEW: A review of previous articles showed that the diagnostic accuracy of percutaneous needle biopsy of vertebral lesions was between 77-94%. The greatest accuracy was achieved in diagnosis of metastatic disease or infection. MATERIALS AND METHODS: We performed 22 percutaneous needle biopsies in twenty vertebral bodies and two posterior elements under the CT guidance. All biopsies were performed with a 16-gauge Osty-cut bone biopsy needle. There were 1 cervical, 8 thoracic and 13 lumbar biopsies. Transpedicular approach was performed on 13 patients and paraspinal approach on 7 patients. There were 8 malignancies, 9 infections, 2 benign tumors, 2 compression fractures and 1 butterfly vertebra. RESULTS: An accurate diagnosis was made in 18 cases(81.8%) of all cases. Two of the 22 specimens were blood clots and two specimens of tuberculous spondylitis were insufficient for definite diagnosis. Bacteriological studies revealed a causative agent in 43% of pyogenic spondylitis. Histologic diagnosis confirming the clinical suspicion was obtained in 16(80%) of the 20 positive cases. There was no complication. CONCLUSION: Percutaneous needle biopsy under the CT guidance is a safe and accurate method for obtaining a diagnosis of vertebral lesions such as osteolytic lesion and infection.