PURPOSE. To compare computed tomography (CT)- guided core needle biopsy (CNB) with incisional biopsy in diagnosing musculoskeletal lesions. METHODS. 62 men and 50 women aged 12 to 83 (mean, 45) years who underwent a CT-guided CNB were compared with 31 men and 33 women aged 9 to 81 (mean, 53) years who underwent an incisional biopsy. All specimens had final pathology report to compare with. Comparisons were made in terms of (1) diagnostic rate, (2) accuracy in distinguishing benign from malignant lesions, (3) accuracy in distinguishing low- from high-grade sarcomas, (4) accuracy for histological diagnosis, and (5) complication and repeated biopsy rates. RESULTS. The diagnostic rate of CT-guided CNB and incisional biopsy was not significantly different (92.9% vs. 96.9%, p=0.33), nor were the accuracy in distinguishing benign from malignant lesions (100% vs. 98.4%, p=0.37), the accuracy in distinguishing low- from high-grade sarcomas (100% vs. 100%, p=1.00), the accuracy for specific diagnosis (75.9% vs. 85.2%, p=0.17), the repeated biopsy rate (6.3% vs. 4.7%, p=0.75), and the complication rate (0.9% vs. 4.7%, p=0.14). The accuracy for specific diagnosis was higher for bone than soft-tissue lesions for both CT-guided CNB (87.0% vs. 59.5%, p=0.002) and incisional biopsy (87.0% vs. 77.3%, p=0.43). The accuracy of CT-guided CNB for specific diagnosis of benign soft-tissue tumours as well as infection and inflammation was relatively low. CONCLUSION. CT-guided CNB is safe, easy to perform, efficient, and less invasive, and should be considered as a first-line biopsy for musculoskeletal lesions.