We report the results of a French multicentric pilot study of remission induction therapy in metastatic neuroblastoma. Thirty-five successive unselected patients entered the study over 1 year and were treated by alternating sequences of cisplatin/VM-26 (PE) and vincristine/cyclophosphamide/doxorubicin (CADO). Three courses of each sequence were delivered. Disease reevaluation was extensive, with special focus on bone marrow status. Using strict criteria, 24 patients (68%) achieved a good partial response (GPR), which comprised normalization of bone marrow, and ten (28%) achieved a partial response (PR), and one progressed. The overall response rate was 96%. Thirty-two patients underwent surgery, and complete macroscopic removal of the primary was achieved in 21 (65%). After completion of induction and surgery, six patients (17%) were in complete remission (CRm), without evidence of any residual disease; nine (26%) were in very good partial remission (VGPRm; same as CRm except persistence of nonpathologically evaluable improved bone scan), and 19 (51%) were in partial remission (PRm). Toxicity was acceptable, and no treatment-related deaths occurred. These results show no substantial improvement compared with those previously reported with similar but nonalternating regimens. We advocate a two-category concept (response, remission) to describe initial therapy results in metastatic neuroblastoma and emphasize the need to assess bone marrow by an extensive evaluation.