Relapsed Acute Myeloid Leukemia in Children and Adolescents: Interim Report of the International Randomised Phase III Study Relapsed AML 2001/01.

Relapse occurs in 30–40% of newly diagnosed AML patients, with long-term survival in 20%. Aiming at improved outcome, we initiated a prospective, randomised study for relapsed AML, excluding AML M3 and those >18 years of age at initial diagnosis. FLAG is being used for 2 consecutive courses: fludarabine 30 mg/m2/day x 5, cytarabine 2 g/m2/day x 5, G-CSF 200 μg/m2/dose for 6 days, starting day -1. Liposomal daunorubicin (DaunoXome, DNX) is a new anthracycline with potentially less cardiotoxicity. Therefore, DNX at 60 mg/m2/day on days 1, 3 and 5 was randomly added or not to the first course of FLAG. Main objectives are to determine the efficacy and toxicity of DNX when added to FLAG, and the long-term outcome in a large group of relapsed AML patients. Thirteen groups worldwide are enrolling patients. More than 400 patients were registered by March 2006. This planned 2^nd interim analysis with blinded efficacy data concerns 322 eligible and evaluable patients with first relapsed AML, of whom 250 (78%) were actually randomised. Fifty-two percent of patients relapsed early (<1 year from initial diagnosis). The majority (84%) concerned isolated bone marrow relapse, with central nervous system involvement in 6% of all patients. Dominating FAB types are M2 with auer rods, M4 without eosinophils and M5. Poor response to the 1st course of therapy (>20% of blasts in the BM shortly before the 2nd course), was seen in 23% of patients, more often in early relapses (31%) than in late relapses (15%). Early death occurred in 6% of patients. Complete remission (CR) was achieved in 63% of patients after 2 courses, and they have a probability of survival at 3 years (3-yr pSurv.) of 47% compared to 33% for the total group and 8% for patients not achieving CR. Compared to early relapses, patients with late relapse had higher CR rates (76 vs 51%), and higher 3-yr pSurv.: 42% vs 23%. Similarly, patients with either t(8;21) or inv(16) had a significantly better outcome. Death in continuous CR occurred in 8.6% of 322 patients, without excess of deaths in one treatment arm. Nearly all patients in CR have been transplanted, the majority with a matched unrelated donor. There was significant grade III/IV toxicity, but no unexpected toxicity, and no clinically relevant differences between the arms with and without DNX, especially not in cardiotoxicity. In conclusion, it is feasible to perform a large randomised pediatric study in a very international setting. DNX added to FLAG does not result in major additional toxicity, but follow-up of cardiotoxicity should be extended as planned in this protocol. Late relapses do better in terms of CR and overall survival, as well as patients with t(8;21) or inv(16), but early relapses achieving CR have a realistic chance of survival as well with currently 23% of them in continuous CR. The study is ongoing until 360 eligible and fully evaluable patients have been randomised, to answer the question whether liposomal daunorubicin improves outcome in pediatric relapsed AML.