The SFOP-OS94 randomised multi-centre trial was designed to determine whether preoperative chemotherapy regimen combining high-dose methotrexate courses and etoposide-ifosfamide could improve the proportion of good histologic response (5% viable cells) compared to a regimen based on high-dose methotrexate and doxorubicin, in children/adolescents with localised high-grade limb osteosarcoma. Postoperative chemotherapy was adapted to the histologic response. Overall, 234 patients were randomised between 1994 and 2001. There were 56% good responders in the etoposide-ifosfamide arm versus 39% in the doxorubicin arm (p-value=0.009). With a median follow-up of 77 months, the 5-year event-free survival of the entire population was 62%, slightly greater in the etoposide-ifosfamide arm than in the doxorubicin arm, but the difference was not significant (Hazard Ratio: HR=0.71, 95%CI: 0.5-1.06, p-value=0.09). Five-year overall survival of the entire population was 76%, similar in both arms (HR=0.95, 95%CI: 0.6-1.6, p-value=0.85). Toxicity was manageable with different acute toxicity profiles between treatment arms. No acute toxicity related death was reported. About 43% of the patients in the etoposide-ifosfamide arm were event-free at 3 years without having received any doxorubicin or cisplatin, thus avoiding the risk of long-term cardio- and ototoxicity.