OBJECTIVETo investigate the way of diagnosis and therapy of parosteal osteosarcoma.

METHODSA retrospective review was conducted of 48 patients treated at our department between June 1964 and December 2001. The average age of patients in this study was 29.2 years (13 - 47 years). Thirty-two of the patients were female; sixteen were male. The single most common site is the posterior aspect of the distal femur (in 36 patients), followed by the femur shaft (in 6 patients), the proximal tibia (in 2 patients), etc. Nine patients had been operated on before referral to our department.

RESULTSThe average follow-up period from the first operation was 5.2 (0.25 - 24) years. In 36 of the 39 patients in whom a limb-salvage procedure was performed, a segment of the tumor-bearing bone was excised along with the tumor, whereas in 3 patients only the subadjacent cortex was excised with the tumor. In the limb-sparing group, the reconstruction was achieved by means of attenuated tumor bone or allograft in 23 cases, by endoprosthetic replacement in 9 cases, and by allograft replacement in 4 cases. The local resections were wide in 35 cases, and marginal in 13 cases. After marginal surgery, local recurrence occurred in 5/13 patients, whereas it occurred in 3/35 patients treated with wide resection. Pulmonary metastases developed in 6 patients, four patients died, and 2 patients are alive with disease. There were 4 cases of fractures of bone grafts. Four patients developed an infection. Long-term survival rate is 85.8%. For tumors that invaded the medullary canal there was no statistical association with local recurrence or metastasis. There is statistical significance between surgical margin and local recurrence.

CONCLUSIONSWide surgical excision alone is adequate treatment for patients with conventional parosteal osteosarcoma. A tumor-free margin remains the critical factor determining overall prognosis. When a marginal excision was knowingly done to preserve a major neurovascular bundle, the risk of recurrence was less than when it was done to shell-out a presumptively benign lesion. Repeated recurrence probably increases the risk of dedifferentiation and thereby worsens the prognosis. Recurrent lesions with multiple soft-tissue satellite nodules or involvement of the neurovascular structures may however require amputation to provide sufficient local control when a wide margin cannot be achieved. An individualized resection will be performed in the future probably under the help of the advanced technique of image to distinguish the reactive zone from the normal tissue precisely.