A tumor registry is essential in the systematic collection and analysis of tumor data in a defined population. Population-based tumor registries have been in operation for more than 30 years in Hiroshima and Nagasaki. While the tumor registry data have previously been used for many site-specific cancer studies in atomic bomb survivors, overall analysis of tumor registry based cancer incidence data among the atomic bomb survivors has not been undertaken for some time. However, a recent improvement in the tumor registries has made it possible to analyze most recent cancer incidence data among the Life Span Study population. Preliminary analysis under way indicates a potential power of the data. There are a few methodological problems associated with the use of the incidence data, especially relating to consistency of data and migration. Several measures taken to resolve such problems are discussed.