The Life Span Study (LSS) cohort of 94,000 survivors of the Hiroshima and Nagasaki atomic bombings is being studied in collaboration with the Radiation Effects Research Foundation (RERF). Cohort studies are used to quantify radiation dose response and its dependence on histological subtype of tumor, age at exposure, sex, age at observation, and time following exposure. Mortality follow-up of this cohort is currently complete through 1997. There have been 9,335 deaths from solid cancer during the 47-year follow-up, with 19% of them occurring in the latest 7 years. An estimated 5% of the solid cancer deaths are attributable to radiation exposure. The excess rates of radiation-related cancers have increased throughout the study period while relative risk is highest among those survivors exposed as children. Since the survivors exposed before age 20 years comprise a large portion (41%) of the cohort and are mostly alive, the number of cancer deaths per year will continue to increase in the next 15 years. The risk of cancer and non-cancer outcomes will continue to be the focus of ongoing and future research. Following the first comprehensive LSS cancer incidence report published several years ago, we are preparing a new solid cancer incidence report. The new incidence series comprises over 12,000 first primary cancers, adding 3,500 new cases that occurred since the first report. Among various site-specific cancer studies, case ascertainment has been completed in studies of tumors of the thyroid and ovary, and manuscripts are in preparation. Case ascertainment is nearing completion for lung cancer and in progress for lymphoid cancer. Among several new findings is the excess risk of nervous system tumors, especially of schwannoma. The data indicate that exposure to even moderate doses, i.e., less than 1 Gy, is associated with an elevated risk of nervous system tumors. A nested case-control study undertaken to clarify the role of hepatitis B and C infection in radiation-related liver cancer risk showed a synergistic interaction between radiation and hepatitis C infection. Analysis of joint effects of radiation and smoking for lung cancer demonstrated important confounding effects of smoking on the radiation risk because of significant differences in smoking behavior among different birth cohorts in Japan. The data also showed that smoking and radiation have an additive effect on lung cancer. Analysis of updated breast cancer incidence data showed that the radiation-related cancer risk is highest among women exposed at ages 0-19 years and lowest among those exposed after age 40, with little variation by exposure between ages 0 and 20, 20 and 40, and above 40, suggesting that breast cell differentiation associated with full-term pregnancy and reduced hormonal stimulation with menopause, both act to modify the radiation effect. A corollary is that secular changes in reproductive history may be partially responsible for the marked difference in breast cancer risk observed between women who were under or over age 20 in 1945 at the time of the bombings. A multidisciplinary study of hormonal assays found pre-diagnostic levels of free estradiol, both at premenopausal and post-menopausal ages, to be a significant predictor of breast cancer. A subsequent international pooled analysis of data from nine studies, including ours, found a significant level of consistency across studies in this respect. A new, expanded A-bomb survivor study is in progress with twice as many breast cancer cases, utilizing more recent stored serum samples and measurement of additional hormones. A multidisciplinary case-case approach is in progress to study genetic susceptibility to radiation-related breast and ovarian cancers. This follows the previously reported phenomenon of extremely high, dose-specific relative risks for early-onset breast cancer (before age 35), which suggests increased sensitivity to radiation among a genetically predisposed population subgroup. In this study, early and late-onset breast cancer cases are compared and BRCA 1/2 and other candidate genes in breast cancer pathways are investigated using archived tissues. Another molecular epidemiological study underway involves thyroid cancer and assesses rearrangements of RET proto-oncogenes (RET-PTC1 and 3) in relation to age at exposure and tumor latency. Construction of a database that incorporates all previous LSS mail survey data is underway. When completed, this database will allow an easy access to a large amount of information of epidemiological interest, i.e., tobacco use, alcohol intake, diet, reproductive history, occupations, medical history, etc. Using some of these data, our recent analysis showed that green tea consumption is virtually unrelated to cancer risk.
