We performed a prospective study of sixty-two patients who were managed with a closed core needle biopsy in an outpatient clinic for a soft-tissue mass or a bone tumor with soft-tissue extension between August 1, 1992, and June 1, 1994. Eight (13 per cent) of the closed core needle biopsies yielded no neoplastic tissue. Two needle biopsies (3 per cent), which were of myxomatous masses, did not allow distinction between a benign and a malignant neoplasm; both masses were extraskeletal myxoid chondrosarcomas. Additionally, the histological grade of four resected specimens (6 per cent) differed from that determined with the closed needle biopsy.The diagnostic accuracy of the closed needle biopsies was 84 per cent (fifty-two of sixty-two). All ten diagnostic errors involved soft-tissue tumors. A retrospective study of a similar cohort of patients, who had open biopsy in an outpatient operating room by the same surgeon in a contemporary period in the same institution and with analysis by the same pathologist, revealed a diagnostic accuracy of 96 per cent (forty-eight of fifty).The hospital charges for the closed core needle biopsy were $1106, compared with $7234 for the open biopsy.We concluded that core needle biopsy can be performed in an outpatient clinic with use of local anesthesia and that it is substantially less expensive and more convenient than open biopsy. This technique has an acceptable but definitely lower rate of accuracy compared with open biopsy, especially for soft-tissue tumors, and it should be used only in a small subset of patients (those who have a large soft-tissue mass or a bone tumor with palpable soft-tissue extension). However, given the small size of the tissue sample, the clinician must recognize possible disadvantages, including a non-diagnostic biopsy, an indeterminate biopsy, or a potential error in the histological grade. These problems are much more likely to occur after core needle biopsy of soft-tissue masses. Because of the potential for errors in diagnosis when core needle biopsy is used, the musculoskeletal oncologist must rely on his or her clinical acumen. When a diagnostic is in reasonable doubt, there is no radiographic confirmation, the biopsy shows no tumor cells, or there is a combination of these findings, operative decisions should be made as if no biopsy had been performed. The management of patients who, after core needle biopsy, have a diagnosis of a bone or soft-tissue tumor, is best carried out by an experienced musculoskeletal oncologist working in close collaboration with an experienced musculoskeletal pathologist.