The evaluation of surgical margin is useful in determining the curative success of surgical treatment of musculoskeletal sarcoma and the degree to which later surgery will be reduced by the preoperative therapy. However, until recently no reliable evaluation method has been developed for these purposes. In this paper we propose a new method for evaluating the surgical margin as drafted in 1989 by the Bone and Soft Tissue Tumor Committee of the Japanese Orthopaedic Association (JOA). In this method, surgical margin was classified into four types based on the distance between the surgical margin and the reactive zone of the tumour. These classifications of surgical margin (in order to surgical extent) are curative wide margin (curative margin), wide margin, marginal margin, and intralesional margin. The surgical margin is said to be curative if the margin is more than 5 cm outside the reactive zone. It is referred to as wide if the margin is less than 5 cm. Similarly, a margin that is in the reactive zone is considered as marginal, and a margin passing through a tumour as intralesional. Moreover, wide margin is classified as adequate (at least 2 cm outside the reactive zone) or inadequate (1 cm). In our evaluation, a "thin" barrier is considered to be a 2-cm thickness of normal tissue, a "thick" barrier as a 3-cm thickness, and joint cartilage is said to be equivalent to a 5-cm thickness. In addition, a surgical margin that is outside a barrier, with normal tissue between the barrier and the reactive zone of the tumour, is considered to be curative. This method was applied in 457 cases (involving 499 surgeries) at the Cancer Institute Hospital to determine clinical significance in the treatment of musculoskeletal sarcoma (1979-1994). From the results of this study we were able to conclude that this evaluation method can be highly useful in clinical practice for establishing optimum surgery. Moreover, we found that the safety margin for high-grade musculoskeletal sarcoma is a curative margin providing an adequate wide margin of 3 cm or more when preoperative therapy is not performed or is not effective, while the safety margin for high-grade sarcoma that responds to preoperative chemo- or radiotherapy seems to be an adequate wide margin of 2 cm. Here, radiotherapy is more effective compared to chemotherapy for reducing surgical margin. An inadequate wide margin, however, combined with radiotherapy, is not enough to ensure local curative success following surgery for musculoskeletal sarcoma.(ABSTRACT TRUNCATED AT 400 WORDS)