Modern treatment for acute lymphoblastic leukaemia (ALL) achieves long term survival in two thirds of children, many of whom are truly cured. Recent improvements have occurred as a result of progressive intensification of treatment, particularly during the first 6 months from diagnosis. This article reviews the means of identifying children for whom standard therapy is not appropriate, and the aspects of standard therapy requiring further refinement, in particular central nervous system (CNS) directed treatment and continuing (maintenance) therapy, a concept unique to ALL. Modern methods for identification of minimal residual disease afford the hope that it may become possible to identify both children who have received sufficient treatment and, conversely, those at subsequent risk of relapse. Selection of patients for alternative therapy, for example marrow transplantation in first remission is difficult and this area needs further study. Treatment for the one third of children who relapse remains unsatisfactory and when successful is only achieved with significant morbidity. Most long-term survivors of childhood leukaemia are well and problem-free but a significant number have problems with learning, concentration, growth and development; the risk of second neoplasms remains unclear. Recent years have seen undeniable progress in the treatment of ALL but there is a continuing need to develop effective forms of treatment which have less potential for damaging late effects.