Three decades of chemotherapy for childhood cancer: from cure 'at any cost' to cure 'at least cost'.

The prognosis for childhood malignancy has improved substantially over the last 30 years. This has been principally due to the use of chemotherapy, centralization of care and treatment by standard protocols. With increasing numbers of children and adolescents cured of cancer there has been an increased awareness of the late effects of therapy. The recent recognition of disease-specific prognostic factors has permitted the subclassification of tumours into good and bad risk. The aim for those malignancies associated with a good outcome is the reduction of late adverse effects, eg intellectual impairment in children with acute lymphoblastic leukemia receiving central nervous system irradiation. In those diseases which have a poor prognosis with present therapy, new innovative and intensive regimens are being investigated in randomized trials. Neuroblastoma is a chemosensitive tumour but there has been little success in translating response into improvements in long-term cure. In poor prognosis neuroblastoma conventional combination chemotherapy, high dose therapy with autologous bone marrow rescue as consolidation and more recently high dose, or high dose rapid schedule, chemotherapy as initial treatment have been sequentially evaluated. In paediatric brain tumours only now are well designed phase II studies being undertaken and problems associated with the classification and evaluation of response to therapy are being addressed. In the past 30 years much of the success with childhood cancer has been largely empirical and the hope for the future is that there will be a more scientific approach to the chemotherapy of childhood malignancy.