Childhood cancer has steadily increased in incidence over the past 60 years, and is the second leading cause of death among children in developed countries. Most childhood cancers have an unknown etiology but are hypothesized to originate in utero due to perinatal, familial, and environmental risk factors that have yet to be identified. We propose to conduct the largest population-based cohort studies to date to identify perinatal, familial, and neighborhood risk factors for specific childhood cancers using comprehensive data for the entire population of Sweden. Perinatal, familial, and neighborhood factors will be identified from the Swedish Birth Registry and census data for all individuals born in Sweden from 1973 through 2008 (N=~3.5 million children). These data will be linked to the Swedish Cancer Registry for complete ascertainment of the most common childhood cancers (including leukemias, brain tumors, lymphomas, neuroblastomas, Wilms tumors, soft tissue sarcomas, retinoblastomas, bone tumors, and hepatoblastomas), diagnosed from birth through 2009. The specific aims are to: 1) Examine the association between perinatal, familial, and neighborhood characteristics and the risk of specific childhood cancers and histologic subtypes; 2) Explore potential interactions among perinatal, familial, and neighborhood characteristics with respect to the risk of specific childhood cancers; and 3) Characterize time trends in the risk of specific childhood cancers during the study period. This cost-effective study will advance the understanding of childhood cancer etiologies, facilitate the identification of high-risk infants, ad ultimately may enable earlier preventive and therapeutic interventions to reduce childhood cancer incidence and mortality.