With the exception of a small percentage of cases attributable to hereditary cancer syndromes (eg, familial retinoblastoma) or genetic syndromes (Down syndrome), the etiology of most childhood cancers is unknown. Recent epidemiologic studies have focused on the prenatal period and have investigated potential associations with parental age, cigarette smoking, birth weight of the child, parental occupational exposures, and specific environmental exposures. The following challenges lie ahead for future epidemiologic studies of childhood cancer: 1) improvement of diagnostic classification; 2) improved methods for exposure assessment; 3) evaluation of data from molecular biology to generate biologically derived hypotheses; and 4) incorporation of markers of genetic susceptibility when feasible.