OBJECTIVE Successful therapeutic interventions to prevent disease progression in patients with nonmetastatic osteosarcoma have included surgery with adjuvant chemotherapy. Presurgical chemotherapy has been advocated for these patients because of putative improvement in event-free survival (EFS). The advantages of presurgical chemotherapy include early administration of systemic chemotherapy, shrinkage of primary tumor, and pathologic identification of risk groups. The theoretic disadvantage is that it exposes a large tumor burden to marginally effective chemotherapy. The contribution of chemotherapy and surgery timing has not been tested rigorously. METHODS Between 1986 and 1993, we conducted a prospective trial in patients with nonmetastatic osteosarcoma who were assigned randomly to immediate surgery or presurgical chemotherapy. Except for the timing of surgery (week 0 or 10), patients received 44 weeks of identical combination chemotherapy that included high-dose methotrexate with leucovorin rescue, doxorubicin, cisplatin, bleomycin, cyclophosphamide, and dactinomycin. RESULTS One hundred six patients were enrolled onto this study. Six were excluded from analysis. Of the remaining 100 patients, 45 were randomly assigned to immediate chemotherapy, and 55 were randomly assigned to immediate surgery. Sixty-seven patients remain disease-free. At 5 years, the projected EFS +/- SE is 65% +/- 6% (69% +/- 8% for immediate surgery and 61% +/- 8% for presurgical chemotherapy; P =.8). The treatment arms had similar incidence of limb salvage (55% for immediate surgery and 50% for presurgical chemotherapy). CONCLUSIONS Chemotherapy was effective in both treatment groups. There was no advantage in EFS for patients given presurgical chemotherapy.