The analysis of the results of two German Pediatric Oncology (GPO) cooperative, neoadjuvant chemotherapy trials after a followup of 7 (COSS-80) and 5 years (COSS-82) allows several conclusions concerning both systemic and local treatment of patients suffering from osteosarcoma. A metastasis free survival rate (MFS) of 59% was reached in the reduced study group of the first study, COSS-80. In addition to size of the primary tumor, the extent of chemotherapy induced devitalisation was very closely related to the probability of survival without systemic recurrence. Following this observation, it was the aim of the next study, COSS-82, to improve the MFS of patients with poorly responding tumors by altering their postoperative chemotherapy regimen. However, this "salvage" approach failed. Moreover, an effort to reduce treatment related toxicity by sparing some patients from the side effects of two particularly toxic drugs, adriamycin (ADR) and cisplatinum (CDDP), by only giving these postoperatively and only after insufficient tumor response to preoperative therapy, failed (MFS of the study arm of COSS-82 45% at 5 years vs. 68% for the control arm with primary use of ADR and CDDP, p less than 0.05). The value of an effective primary chemotherapy is further enhanced by the observation, that en bloc resection of tumors which were poor responders to preoperative therapy was associated with an increased risk of distant metastases when compared with amputation and rotation plasty, while this was not the case for good responders. In conclusion, both systemic tumor control and optimal local therapy require that all effects drugs are to be used as early as possible in the primary treatment of osteosarcoma, in order to enforce maximum tumor cell destruction and hence an optimistic outlook for the individual patient.