Until recently, the etiology of childhood acute lymphoblastic leukemia (ALL) has remained relatively elusive. Several studies have established a time frame for the development of ALL which could lead to the identification of specific exposures linked to leukemogenesis from the generation of the initial leukemic clone until clinical diagnosis. Utilizing newborn screening ('Guthrie') cards, leukemic clones have been detected retrospectively in dried blood spots using two different PCR-based approaches: (i) the amplification of patient/leukemia-specific breakpoint fusion sequences of rearranged oncogenes; and (ii) the amplification of clonal immunoglobulin heavy chain gene (IgH) or T cell receptor (TcR) gene rearrangements. These studies support the hypothesis that a large proportion of childhood ALL cases arise in utero. In several studies, a long latency period from the generation in utero of the initial ALL clone to clinical diagnosis, indicates that additional genetic events are required for the full development of the leukemia phenotype, potentially from postnatal exposures (e.g. infections). The identification of leukemia-associated translocations in umbilical cord blood samples of healthy newborns, suggest that in the future children may be identified prospectively who have an increased risk of developing leukemia.