Nowadays about two-thirds of children with acute lymphoblastic leukemia (ALL) can be cured with chemotherapy, but one-third die from the disease. The clinical response of leukemic cells to chemotherapy is roughly due to two factors: the effective drug levels reaching the cells and the resistance of these cells to the drugs. The clinical value of cellular drug resistance in children with ALL is not known. We developed an in vitro assay to study drug resistance in these children. In this article, the main results obtained with this MTT assay on samples from 137 children with ALL are summarized: (1) patients whose cells are resistant to several drugs at initial diagnosis have a poor prognosis; (2) relapsed leukemias show a considerable drug resistance which might partly explain the poor prognosis. Relapsed cases differ in their type and degree of resistance; (3) the poor outcome of high risk groups as defined by age and immunophenotype can partly be explained by specific patterns of drug resistance; (4) P-glycoprotein-mediated multidrug resistance is not an important cause of resistance in childhood ALL; and (5) no relation exists between the activities of the purine enzymes HGPRT, 5'NT, ADA, and PNP and drug resistance in childhood ALL. The conclusion is that in vitro drug resistance data have clinical relevance and can be used to develop more effective and less toxic treatment strategies in childhood ALL.