Due to progressive therapy intensification in the four consecutive studies AML-BFM 78, 83, 93 and 98, prognosis for children with acute myeloid leukemia (AML) has improved steadily. In spite of the intensified therapy, rates of morbidity and mortality have remained unchanged or have even decreased. Against the background that about 40% of the patients still die from immediate causes of an underlying disease relapse or of nonresponse, it seems to be justifiable to intensify therapy - especially for high-risk patients - which on its parts will require an optimization of supportive measures. As the present risk stratification into standard- (SR) and high-risk (HR) patients has proved effective, we will pursue the risk-adapted therapy strategy.

The aim of the study is to improve prognosis in children with AML by intensification of cytostatic therapy and to evaluate by randomisation the equivalence of a prophylactic central nervous system (CNS) irradiation with a total dose of 18 Gy versus 12 Gy.

During the last decade, prognosis in acute myelogenous leukemia (AML) in childhood has improved considerably, but still 30% of the children experience a relapse of disease and further 10% fail to respond sufficiently to the present therapies. A further intensification of therapy might improve the overall survival of these children, but possible, implicit side effects have to be considered carefully. Increase in dose intensification of the proven, effective anthracyclines will be limited by the risk of cumulative cardiotoxicity. A liposomal formulation of daunorubicin may offer a possibility to increase dosage, at least partially, without causing cumulative cardiotoxicity. Objective one of this randomised study is to ascertain if this dose increase improves therapy response and overall survival at acceptable toxicity. The previous experiences with liposomal daunorubicin, gathered from the relapse studies AML-BFM Rez 97 and International Therapy Study Relapsed AML 2001/012 as well as from the pilot study AML-BFM 2002P, have shown that the induction therapy is feasible in clinical centers with experience in AML therapy without leading to a marked increase of toxicity or prolongation of granulocytopenia.

First results of study AML-BFM 98 have shown that the patients of the standard risk (SR) group did not benefit from an additional, second induction (HAM). On this account we did not reintroduce this second induction course in the present study AML-BFM 2004. However, SR patients will take part in the randomisation of initial therapy with a general view to achieving higher effectiveness.

For patients of the high-risk group, the administration of 2-chloro-2-deoxyadenosine (2-CDA) will be integrated in the first phase of consolidation to achieve an even higher intensification. It could be shown that 2-CDA possesses good antileukemic activity in pediatric and adult AML. In a phase-II study it could also be demonstrated that 2-CDA has good effectiveness in combination with cytarabine. Results of phase-II studies conducted at St. Jude Children's Hospital, Memphis, showed that 2-CDA has good effectiveness especially in children with monoblastic leukemias (FAB M4/M5). Consequently, this intensification for high-risk patients, who present in more than half of the cases with monoblastic leukemias (FAB M4/5), may allow further improvement of therapy for this cohort. The pilot study AML-BFM 2002P confirmed that the study design was practicable without increasing significantly the risk of higher toxicity. However, median duration of aplasia was significantly prolonged in comparison to that of the AI (cytarabine, idarubicin)-block. Objective two of this study is to determine by randomisation if an improvement of efficiency is possible.

Study AML-BFM 98 has already focussed on the question of whether or not doses of CNS irradiation of 12 Gy and 18 Gy are equivalent with regard to their capacity of reducing the risk of relapse. As the results of study AML-BFM 87 confirmed the necessity of CNS irradiation, but did not reveal the necessary minimum dose, this randomisation has been implemented in order to prevent, as far as possible, late sequelae of CNS irradiation by reducing the radiation dose (= objective three).

As the number of patients of study AML-BFM 98 was not sufficient to resolve this question, this randomised analysis has been extended for a second period and will therefore be continued in the current study AML-BFM-2004.

Besides the intensification of cytostatic therapy, study AML-BFM 2004 seeks to optimise the quality of supportive therapy by implementing measures of quality assurance. This demands an up-to-date, complete documentation of each therapy phase. In studies AML-BFM 93 and 98, about 12% of deaths were due to primary complications such as leukostasis syndrome, haemorrhage or severe infections (4%), infections in aplasia before achieving remission (4%) or infections in remission (4%). Maybe the lives of even more children will be saved in the future by improved standards for the prevention of primary complications. Further, the efficacy of chemotherapy could be improved by less delays in therapy which are often due to infections or other complications.

Inclusion Criteria:

• Age from >0 to \