Purpose. To evaluate the sensitivity and specificity of core needle biopsy in determining musculoskeletal tumours in our hospital. Methods. Records of 134 patients who underwent core needle biopsy followed by definitive surgery were retrospectively reviewed. Results of the core needle biopsy were compared with those of the final histology. Histology was classified into benign versus malignant, and bony versus soft-tissue lesions. The sensitivity and specificity of core needle biopsy were calculated. Results. Based on final histology, there were 33 bone tumours (3 benign and 30 malignant), 74 soft-tissue tumours (6 benign and 68 malignant), 11 schwannomas (7 benign and 4 malignant), and 16 inflammatory/necrotic (benign) lesions. For 118 (88%) tumours, the biopsy results matched the final histological results. For 7 tumours, biopsy results were non-diagnostic, as the amount of tissue obtained was insufficient. For 9 tumours, biopsy results did not match the final histological results; 5 considered benign but turned out to be malignant, one considered malignant but turned out to be benign, and 3 were correctly identified as malignant but incorrectly subtyped. The sensitivity and specificity of core needle biopsy were 95% (97/102) and 97% (31/32), respectively, assuming that the 7 non-diagnostic tumours were correctly diagnosed. Conclusion. Core needle biopsy is an accurate and reliable diagnostic tool for musculoskeletal tumours if performed by skilled persons and adequate tissue is obtained.