Systematic enrolment of children and adolescents with acute lymphoblastic leukaemia (ALL) into clinical trials has allowed the establishment of prognostic parameters derived from initial diagnostic findings. More important, these trials have significantly contributed to the reduction of disease recurrence as much as to the reduction of acute and late side effects. Some problems that are related to the specificity of the parameters used for risk assessment were not overcome: high tumour load by white blood cell count (WBC), age and (rare) cytogenetic subtypes (e.g. t9;22) may characterise a significant proportion of children and adolescents with high-risk ALL. Most patients who will eventually relapse do not present with characteristic features at initial diagnosis. It appears feasible through careful response assessment to identify these patients at risk of relapse, who present initially without specific features. Earlier trials of the ALL-BFM (Berlin/Frankfurt/Münster) study group and others have demonstrated that inadequate leukaemic blast reduction in the peripheral blood or bone marrow after the first few days of therapy is highly predictive of treatment failure. Using clone-specific polymerase chain reaction-based detection of minimal residual disease (MRD) as done in trial AIEOP-BFM ALL 2000 allowed a close surveillance of specific treatment elements when applied in MRD positive patients. This may facilitate innovative chemotherapy approaches and a more rational use of allogeneic haematopoetic stem cell transplantation. In addition, genetic signatures of treatment response or failure have been identified.