In 1982, members of the Musculoskeletal Tumor Society, representing sixteen centers for the treatment of bone and soft-tissue cancer, compiled data regarding the hazards associated with 329 biopsies of primary malignant musculoskeletal sarcomas. The investigation showed troubling rates of errors in diagnosis and technique, which resulted in complications and also adversely affected the care of the patients. These data were quite different when the biopsy had been carried out in a treatment center rather than in a referring institution. On the basis of these observations, the Society made a series of recommendations about the technical aspects of the biopsy and stated that, whenever possible, the procedure should be done in a treatment center rather than in a referring institution.In 1992, the Musculoskeletal Tumor Society decided to perform a similar study to determine whether the rates of complications, errors, and deleterious effects related to biopsy had changed. Twenty-five surgeons from twenty-one institutions submitted the cases of 597 patients. The results were essentially the same as those in the earlier study. The rate of diagnostic error for the total series (in which cases from referring institutions and treatment centers were combined) was 17.8 per cent. There was no significant difference in the rate of patients for whom a problem with the biopsy forced the surgeon to carry out a different and often more complex operation or to use adjunctive irradiation or chemotherapy (19.3 per cent in the current study, compared with 18 per cent in the previous one). There was also no significant difference in the percentage of patients who had a change in the outcome, such as the need for a more complex resection that resulted in disability, loss of function, local recurrence, or death, attributable to problems related to the biopsy (10.1 per cent in the current study, compared with 8.5 per cent in the 1982 study). Eighteen patients in the current study had an unnecessary amputation as a result of the biopsy, compared with fifteen in the previous study.Errors, complications, and changes in the course and outcome were two to twelve times greater (p < 0.001) when the biopsy was done in a referring institution instead of in a treatment center.