Purpose: To assess the impact of addition of vinblastine (VLB) to standard chemotherapy induction courses, based on the NHL-BFM90 protocol, and then weekly for a total duration of treatment of one year in children with high-risk anaplastic large cell lymphoma (ALCL).Design: The ALCL99-VLB trial is a randomized trial involving most European groups and the Japanese group. This ALCL99-VLB trial is part of the factorial design ALCL99 trial also aiming to assess efficacy and safety of two methotrexate doses and administration schedules (ALCL99-MTX trial already reported). Chemotherapy was based on the NHL-BFM90 protocol. All patients received a 5-day cytoreductive prephase followed by 6 induction courses given every 21 days. Children with high-risk ALCL, defined as disease with mediastinal, lung, liver or spleen involvement or biopsy proven skin disease, were randomized before the second induction course to receive either no vinblastine (No-VLB arm, reference arm) or one vinblastine injection (6 mg/m²) during the five latter induction courses and then weekly as maintenance treatment (VLB-arm). The duration of chemotherapy was 4 months in the No-VLB-arm and 1 year in the VLB-arm. We report here the results of the final analysis, performed after 3 interim analyses, with a 2-sided p-value equal to .0412 and 95.88% confidence intervals [95.88%CI]. The main end-point was the Event-Free survival (EFS). The main analysis was performed on the intent-to-treat population.Results: Overall 217 patients (93% ALK+) were recruited between 11/1999 and 05/2006 in the VLB-trial: 110 in the VLB-arm and 107 in the No-VLB-arm. A major violation of protocol treatment was observed in 4 VLB-patients and 1 No-VLB patient respectively. In the VLB-arm, 100 patients started VLB maintenance treatment; their median total treatment duration was 53 weeks (range, 21 to 87 weeks) and the median dose intensity during maintenance treatment was approximately 4.9 mg/m2/week. At the time of this final analysis, all patients had a least 2-year follow-up, with a median follow-up equal to 4.7 years. A total of 66 events were reported, including 65 relapses or progressions and 1 treatment-related death as first event. No second malignancy has been reported so far. Compared to No-VLB patients, VLB-patients had a significantly lower risk of event during the first year, Hazard Ratio of event, HR=0.31 [0.15 to 0.67], p=.002, whereas they had an increased risk of event after the first year, HR=5.0 [1.65 to 15.0], p=.003. Consequently, EFS rates differed greatly at 1 year (74% in the No-VLB-arm versus 91% in the VLB-arm) whereas they were very similar at 2 years (70% versus 74%). Over the whole follow-up period, the EFS curves did not differ significantly, HR=0.91 [0.55 to 1.51], p=.70. Results were very similar after exclusion of the 6 patients for whom the diagnosis of ALCL was rejected after review. There was no significant heterogeneity in the VLB-effect on the EFS between the two MTX groups (p=.85) or between the countries (p=.29). There was no significant impact of the addition of VLB on the complete remission rate (85% versus 86%, p=.82) or on the overall survival (p=.61, with 95% versus 92% at 2 years). Toxicity during induction courses was very similar in both arms. Only 3 patients stopped VLB maintenance treatment due to toxicity.Conclusion: The addition of vinblastine during induction and as maintenance treatment significantly delayed occurrence of relapses but, with this duration of treatment of 1 year, did not reduce the risk of failure resulting in the absence of benefit when considering the overall EFS curves or the 2-year EFS.
Read full abstract