Search

In this article, we describe recent methodological enhancements and findings from the dose reconstruction component of a study of health risks among U.S. radiologic technologists. An earlier version of the dosimetry published in 2006 used physical and statistical models, literature-reported exposure measurements for the years before 1960, and archival personnel monitoring badge data from cohort members through 1984. The data and models previously described were used to estimate annual occupational radiation doses for 90,000 radiological technologists, incorporating information about each individual's employment practices based on a baseline survey conducted in the mid-1980s. The dosimetry methods presented here, while using many of the same methods as before, now estimate 2.23 million annual badge doses (personal dose equivalent) for the years 1916–1997 for 110,374 technologists, but with numerous methodological improvements. Every technologist's annual dose is estimated as a probability density function to reflect uncertainty about the true dose. Multiple realizations of the entire cohort distribution were derived to account for shared uncertainties and possible biases in the input data and assumptions used. Major improvements in the dosimetry methods from the earlier version include: A substantial increase in the number of cohort member annual badge dose measurements; Additional information on individual apron usage obtained from surveys conducted in the mid-1990s and mid-2000s; Refined modeling to develop lognormal annual badge dose probability density functions using censored data regression models; Refinements of cohort-based annual badge probability density functions to reflect individual work patterns and practices reported on questionnaires and to more accurately assess minimum detection limits; and Extensive refinements in organ dose conversion coefficients to account for uncertainties in radiographic machine settings for the radiographic techniques employed. For organ dose estimation, we rely on well-researched assumptions about critical exposure-related variables and their changes over the decades, including the peak kilovoltage and filtration typically used in conducting radiographic examinations, and the usual body location for wearing radiation monitoring badges, the latter based on both literature and national recommendations. We have derived organ dose conversion coefficients based on air-kerma weighting of photon fluences from published X-ray spectra and derived energy-dependent transmission factors for protective lead aprons of different thicknesses. Findings are presented on estimated organ doses for 12 organs and tissues: red bone marrow, female breast, thyroid, brain, lung, heart, colon, ovary, testes, skin of trunk, skin of head and neck and arms, and lens of the eye.

As genome-wide association studies of breast cancer are replicating findings and refinement studies are narrowing the signal location, additional efforts are necessary to elucidate the underlying functional relationships. One approach is to evaluate variation in risk by genotype based on known breast carcinogens, such as ionizing radiation. Given the public health concerns associated with recent increases in medical radiation exposure, this approach may also identify potentially susceptible subpopulations. We examined interaction between 27 newly identified breast cancer risk alleles (identified within the NCI Cancer Genetic Markers of Susceptibility and the Breast Cancer Association Consortium genome-wide association studies) and occupational and medical diagnostic radiation exposure among 859 cases and 1083 controls nested within the United States Radiologic Technologists cohort. We did not find significant variation in the radiation-related breast cancer risk for the variant in RAD51L1 (rs10483813) on 14q24.1 as we had hypothesized. In exploratory analyses, we found that the radiation-associated breast cancer risk varied significantly by linked markers in 5p12 (rs930395, rs10941679, rs2067980 and rs4415084) in the mitochondrial ribosomal protein S30 (MRPS30) gene (P(interaction) = 0.04). Chance, however, may explain these findings, and as such, these results need to be confirmed in other populations with low to moderate levels of radiation exposure. Even though a complete understanding of the way(s) in which these variants may increase breast cancer risk remains elusive, this approach may yield clues for further investigation.

Ionizing radiation-associated breast cancer risk appears to be modified by timing of reproductive events such as age at radiation exposure, parity, age at first live birth, and age at menopause. However, potential breast cancer risk modification of low to moderate radiation dose by polymorphic estrogen metabolism-related gene variants has not been routinely investigated. We assessed breast cancer risk of 12 candidate variants in 12 genes involved in steroid metabolism, catabolism, binding, or receptor functions in a study of 859 cases and 1,083 controls within the US radiologic technologists (USRT) cohort. Using cumulative breast dose estimates from a detailed assessment of occupational and personal diagnostic ionizing radiation exposure, we investigated the joint effects of genotype on the risk of breast cancer. In multivariate analyses, we observed a significantly decreased risk of breast cancer associated with the CYP3A4 M445T minor allele (rs4986910, OR = 0.3; 95% CI 0.1-0.9). We found a borderline increased breast cancer risk with having both minor alleles of CYP1B1 V432L (rs1056836, CC vs. GG, OR = 1.2; 95% CI 0.9-1.6). Assuming a recessive model, the minor allele of CYP1B1 V432L significantly increased the dose-response relationship between personal diagnostic X-ray exposure and breast cancer risk, adjusted for cumulative occupational radiation dose (p (interaction) = 0.03) and had a similar joint effect for cumulative occupational radiation dose adjusted for personal diagnostic X-ray exposure (p (interaction) = 0.06). We found suggestive evidence that common variants in selected estrogen metabolizing genes may modify the association between ionizing radiation exposure and breast cancer risk.

Measurement of chromosome translocations in peripheral blood lymphocytes has been used to quantify prior exposure to ionizing radiation, including for workers exposed to low, chronic doses. We assessed translocation frequencies in a subset of U.S. radiologic technologists to substantiate ionizing radiation dose estimates developed for 110,418 technologists who worked between 1916 and 1984. From 3,441 cohort members known to have begun working before 1950, we selected a sample of 152, stratified by estimated cumulative dose, over-sampling from higher-dose categories and excluding persons with a prior cancer diagnosis, a personal or family history of chromosomal instability disorders, or a current history of smoking. Estimates of film-badge dose ranged from less than 10 cSv to more than 30 cSv. Blood samples, obtained in 2004, were analyzed by fluorescence in situ hybridization (FISH) whole chromosome painting by simultaneously labeling chromosomes 1, 2 and 4 in red and 3, 5 and 6 in green. Translocations were scored in 1800 well-spread metaphase cells and expressed per 100 cell equivalents (CE) per person. Linear Poisson regression models with allowance for overdispersion were used to assess the relationship between estimated occupational red bone marrow absorbed dose in cGy and translocation frequency, adjusted for age, gender and estimated red bone marrow absorbed dose score from personal diagnostic procedures. We observed 0.09 excess translocations per 100 CE per cGy red bone marrow dose (95% CI: -0.01, 0.2; P = 0.07), which is similar to the expected estimate based on previous cytogenetic studies (0.05 excess translocations per 100 CE per cGy). Despite uncertainty in the estimates of occupational red bone marrow absorbed doses, we found good general agreement between the doses and translocation frequencies, lending support to the credibility of the dose assessment for this large cohort of U.S. radiologic technologists.

Dosimetic uncertainties, particularly those that are shared among subgroups of a study population, can bias, distort or reduce the slope or significance of a dose response. Exposure estimates in studies of health risks from environmental radiation exposures are generally highly uncertain and thus, susceptible to these methodological limitations. An analysis was published in 2008 concerning radiation-related thyroid nodule prevalence in a study population of 2,994 villagers under the age of 21 years old between August 1949 and September 1962 and who lived downwind from the Semi-palatinsk Nuclear Test Site in Kazakhstan. This dose-response analysis identified a statistically significant association between thyroid nodule prevalence and reconstructed doses of fallout-related internal and external radiation to the thyroid gland; however, the effects of dosimetric uncertainty were not evaluated since the doses were simple point “best estimates”. In this work, we revised the 2008 study by a comprehensive treatment of dosimetric uncertainties. Our present analysis improves upon the previous study, specifically by accounting for shared and unshared uncertainties in dose estimation and risk analysis, and differs from the 2008 analysis in the following ways: 1. The study population size was reduced from 2,994 to 2,376 subjects, removing 618 persons with uncertain residence histories; 2. Simulation of multiple population dose sets (vectors) was performed using a two-dimensional Monte Carlo dose estimation method; and 3. A Bayesian model averaging approach was employed for evaluating the dose response, explicitly accounting for large and complex uncertainty in dose estimation. The results were compared against conventional regression techniques. The Bayesian approach utilizes 5,000 independent realizations of population dose vectors, each of which corresponds to a set of conditional individual median internal and external doses for the 2,376 subjects. These 5,000 population dose vectors reflect uncertainties in dosimetric parameters, partly shared and partly independent, among individual members of the study population. Risk estimates for thyroid nodules from internal irradiation were higher than those published in 2008, which results, to the best of our knowledge, from explicitly accounting for dose uncertainty. In contrast to earlier findings, the use of Bayesian methods led to the conclusion that the biological effectiveness for internal and external dose was similar. Estimates of excess relative risk per unit dose (ERR/Gy) for males (177 thyroid nodule cases) were almost 30 times those for females (571 cases) and were similar to those reported for thyroid cancers related to childhood exposures to external and internal sources in other studies. For confirmed cases of papillary thyroid cancers (3 in males, 18 in females), the ERR/Gy was also comparable to risk estimates from other studies, but not significantly different from zero. These findings represent the first reported dose response for a radiation epidemiologic study considering all known sources of shared and unshared errors in dose estimation and using a Bayesian model averaging (BMA) method for analysis of the dose response.

The assessment of potential benefits versus harms from mammographic examinations as described in the controversial breast cancer screening recommendations of the U.S. Preventive Task Force included limited consideration of absorbed dose to the fibroglandular tissue of the breast (glandular tissue dose), the tissue at risk for breast cancer. Epidemiological studies on cancer risks associated with diagnostic radiological examinations often lack accurate information on glandular tissue dose, and there is a clear need for better estimates of these doses. Our objective was to develop a quantitative summary of glandular tissue doses from mammography by considering sources of variation over time in key parameters, including imaging protocols, X-ray target materials, voltage, filtration, incident air kerma, compressed breast thickness, and breast composition. We estimated the minimum, maximum and mean values for glandular tissue dose for populations of exposed women within 5-year periods from 1960 to the present, with the minimum to maximum range likely including 90% to 95% of the entirety of the dose range from mammography in North America and Europe. Glandular tissue dose from a single view in mammography is presently about 2 mGy, about one-sixth the dose in the 1960s. The ratio of our estimates of maximum to minimum glandular tissue doses for average-size breasts was about 100 in the 1960s compared to a ratio of about 5 in recent years. Findings from our analysis provide quantitative information on glandular tissue doses from mammographic examinations that can be used in epidemiological studies of breast cancer.

In this paper, we describe recent methodological enhancements and findings from the dose reconstruction component of a study of cancer risks among U.S. radiologic technologists. An earlier version of the dosimetry published in 2006 (Simon et al., Radiat. Res. 166, 174-192, 2006) used physical and statistical models, literature-reported exposure measurements for the years before 1960, and archival personnel monitoring badge data from cohort members through 1984. The data and models were used to estimate unknown occupational radiation doses for 90,000 radiological technologists, incorporating information about each individual's employment practices based on a survey conducted in the mid-1980s. The dosimetry methods presented here, while using many of the same methods as before, now estimate annual and cumulative occupational badge doses (personal dose equivalent) to about 110,000 technologists for each year worked from 1916 to 2006, but with numerous methodological improvements. This dosimetry, using much more comprehensive information on individual use of protection aprons, estimates radiation absorbed doses to 12 organs and tissues (red bone marrow, ovary, colon, brain, lung, heart, female breast, skin of trunk, skin of head and neck and arms, testes, thyroid and lens of the eye). Every technologist's annual dose is estimated as a probability density function (pdf) to account for shared and unshared uncertainties. Major improvements in the dosimetry methods include a substantial increase in the number of cohort member annual badge dose measurements, additional information on individual apron use obtained from surveys conducted in the 1990s and 2005, refined modeling to develop annual badge dose pdfs using Tobit regression, refinements of cohort-based annual badge pdfs to delineate exposures of highly and minimally exposed individuals and to assess minimal detectable limits more accurately, and extensive refinements in organ dose conversion coefficients to account for uncertainties in radiographic techniques employed. For organ dose estimation, we rely on well-researched assumptions about critical exposure-related variables and their changes over the decades, including the peak kilovoltage and filtration typically used in conducting radiographic examinations and the usual body location for wearing radiation monitoring badges. We have derived organ dose conversion coefficients based on air-kerma weighting of photon fluences from published X-ray spectra and derived energy-dependent transmission factors for protective aprons of different thicknesses. We tailor bone marrow dose estimates to individual cohort members by using an individual-specific body mass index correction factor. To our knowledge the models and reconstructed doses presented herein represent the most comprehensive dose reconstructions undertaken for a cohort of medical radiation workers.

Ionizing radiation, an established breast cancer risk factor, has been shown to induce oxidative damage and chronic inflammation. Polymorphic variation in oxidative stress and inflammatory-mediated pathway genes may modify radiation-related breast cancer risk.We estimated breast cancer risk for 28 common variants in 16 candidate genes involved in these pathways among 859 breast cancer cases and 1,083 controls nested within the US Radiologic Technologists cohort. We estimated associations between occupational and personal diagnostic radiation exposures with breast cancer by modeling the odds ratio (OR) as a linear function in logistic regression models and assessed heterogeneity of the dose-response across genotypes.There was suggestive evidence of an interaction between the rs5277 variant in PTGS2 and radiation-related breast cancer risk. The excess OR (EOR)/Gy from occupational radiation exposure = 5.5 (95%CI 1.2-12.5) for the GG genotype versus EOR/Gy0 (95%CI0-3.8) and EOR/Gy0 (95%CI0-14.8) for the GC and CC genotypes, respectively, (p (interaction) = 0.04). The association between radiation and breast cancer was not modified by other SNPs examined.This study suggests that variation in PTGS2 may modify the breast cancer risk from occupational radiation exposure, but replication in other populations is needed to confirm this result.

Abstract In this paper, we describe recent methodological enhancements and findings from the dose reconstruction component of a study of cancer risks among U.S. radiologic technologists. An earlier version of the dosimetry published in 2006 (Simon et al., Radiat. Res. 166, 174-192, 2006) used physical and statistical models, literature-reported exposure measurements for the years before 1960, and archival personnel monitoring badge data from cohort members through 1984. The data and models were used to estimate unknown occupational radiation doses for 90,000 radiological technologists, incorporating information about each individual's employment practices based on a survey conducted in the mid-1980s. The dosimetry methods presented here, while using many of the same methods as before, now estimate annual and cumulative occupational badge doses (personal dose equivalent) to about 110,000 technologists for each year worked from 1916 to 2006, but with numerous methodological improvements. This dosimetry, using much more comprehensive information on individual use of protection aprons, estimates radiation absorbed doses to 12 organs and tissues (red bone marrow, ovary, colon, brain, lung, heart, female breast, skin of trunk, skin of head and neck and arms, testes, thyroid and lens of the eye). Every technologist's annual dose is estimated as a probability density function (pdf) to account for shared and unshared uncertainties. Major improvements in the dosimetry methods include a substantial increase in the number of cohort member annual badge dose measurements, additional information on individual apron use obtained from surveys conducted in the 1990s and 2005, refined modeling to develop annual badge dose pdfs using Tobit regression, refinements of cohort-based annual badge pdfs to delineate exposures of highly and minimally exposed individuals and to assess minimal detectable limits more accurately, and extensive refinements in organ dose conversion coefficients to account for uncertainties in radiographic techniques employed. For organ dose estimation, we rely on well-researched assumptions about critical exposure-related variables and their changes over the decades, including the peak kilovoltage and filtration typically used in conducting radiographic examinations and the usual body location for wearing radiation monitoring badges. We have derived organ dose conversion coefficients based on air-kerma weighting of photon fluences from published X-ray spectra and derived energy-dependent transmission factors for protective aprons of different thicknesses. We tailor bone marrow dose estimates to individual cohort members by using an individual-specific body mass index correction factor. To our knowledge the models and reconstructed doses presented herein represent the most comprehensive dose reconstructions undertaken for a cohort of medical radiation workers.

Exposure to ionizing radiation has been consistently associated with increased risk of female breast cancer. Although the majority of DNA damage caused by ionizing radiation is corrected by the base-excision repair pathway, certain types of multiple-base damage can only be repaired through the nucleotide excision repair pathway. In a nested case–control study of breast cancer in US radiologic technologists exposed to low levels of ionizing radiation (858 cases, 1,083 controls), we examined whether risk of breast cancer conferred by radiation was modified by nucleotide excision gene polymorphisms ERCC2 (XPD) rs13181, ERCC4 (XPF) rs1800067 and rs1800124, ERCC5 (XPG) rs1047769 and rs17655; and ERCC6 rs2228526. Of the 6 ERCC variants examined, only ERCC5 rs17655 showed a borderline main effect association with breast cancer risk (ORGC = 1.1, ORCC = 1.3; p-trend = 0.08), with some indication that individuals carrying the C allele variant were more susceptible to the effects of occupational radiation (EOR/GyGG = 1.0, 95% CI =

Genome-wide association studies are discovering relationships between single-nucleotide polymorphisms and breast cancer, but the functions of these single-nucleotide polymorphisms are unknown and environmental exposures are likely to be important. We assessed whether breast cancer risk single-nucleotide polymorphisms interacted with ionizing radiation, a known breast carcinogen, among 859 cases and 1,083 controls nested in the U.S. Radiologic Technologists cohort. Among 11 Breast Cancer Association Consortium risk single-nucleotide polymorphisms, we found that the genotype-associated breast cancer risk varied significantly by radiation dose for rs2107425 in the H19 gene (Pinteraction = 0.001). H19 is a maternally expressed imprinted mRNA that is closely involved in regulating the IGF2 gene and could exert its influence by this or by some other radiation-related pathway. (Cancer Epidemiol Biomarkers Prev 2008;17(8):2007–11)

The study aim was to determine the risk of cataract among radiologic technologists with respect to occupational and nonoccupational exposures to ionizing radiation and to personal characteristics. A prospective cohort of 35,705 cataract-free US radiologic technologists aged 24-44 years was followed for nearly 20 years (1983-2004) by using two follow-up questionnaires. During the study period, 2,382 cataracts and 647 cataract extractions were reported. Cigarette smoking for >= 5 pack-years; body mass index of >= 25 kg/m(2); and history of diabetes, hypertension, hypercholesterolemia, or arthritis at baseline were significantly (p = 3 x-rays to the face/neck was associated with a hazard ratio of cataract of 1.25 (95% confidence interval: 1.06, 1.47). For workers in the highest category (mean, 60 mGy) versus lowest category (mean, 5 mGy) of occupational dose to the lens of the eye, the adjusted hazard ratio of cataract was 1.18 (95% confidence interval: 0.99, 1.40). Findings challenge the National Council on Radiation Protection and International Commission on Radiological Protection assumptions that the lowest cumulative ionizing radiation dose to the lens of the eye that can produce a progressive cataract is approximately 2 Gy, and they support the hypothesis that the lowest cataractogenic dose in humans is substantially less than previously thought.

Background: Although genes involved in apoptosis pathways and DNA repair pathways are both essential for maintaining genomic integrity, genetic variants in DNA repair have been thought to increase susceptibility to radiation carcinogenesis, but similar hypotheses have not generally been raised about apoptosis genes. For this reason, potential modification of the relationship between ionizing radiation exposure and breast cancer risk by polymorphic apoptosis gene variants have not been investigated among radiation-exposed women. Methods: In a case-control study of 859 cases and 1,083 controls within the U.S. Radiologic Technologists cohort, we assessed breast cancer risk with respect to 16 candidate variants in eight genes involved in apoptosis, inflammation, and proliferation. Using carefully reconstructed cumulative breast dose estimates from occupational and personal diagnostic ionizing radiation, we also investigated the joint effects of these polymorphisms on the risk of breast cancer. Results: In multivariate analyses, we observed a significantly decreased risk of breast cancer associated with the homozygous minor allele of CASP8 D302H [rs1045485, odds ratio (OR), 0.3; 95% confidence interval (95% CI), 0.1-0.8]. We found a significantly increased breast cancer risk with increasing minor alleles for IL1A A114S (rs17561); heterozygote OR 1.2 (95% CI, 1.0-1.4) and homozygote OR 1.5 (95% CI, 1.1-2.0), Ptrend = 0.008. Assuming a dominant genetic model, IL1A A114S significantly modified the dose-response relationship between cumulative personal diagnostic radiation and breast cancer risk, adjusted for occupational dose (Pinteraction = 0.004). Conclusion: The U.S. Radiologic Technologists breast cancer study provided a unique opportunity to examine the joint effects of common genetic variation and ionizing radiation exposure to the breast using detailed occupational and personal diagnostic dose data. We found evidence of effect modification of the radiation and breast cancer dose-response relationship that should be confirmed in studies with more cases and controls and quantified radiation breast doses in the low-to-moderate range. (Cancer Epidemiol Biomarkers Prev 2007;16(10):2000–7)

High-dose ionizing radiation exposure to the breast and rare autosomal dominant genes have been linked with increased breast cancer risk, but the role of low-to-moderate doses from protracted radiation exposure in breast cancer risk and its potential modification by polymorphisms in DNA repair genes has not been previously investigated among large numbers of radiation-exposed women with detailed exposure data. Using carefully reconstructed estimates of cumulative breast doses from occupational and personal diagnostic ionizing radiation, we investigated the potential modification of radiation-related breast cancer risk by 55 candidate single nucleotide polymorphisms in 17 genes involved in base excision or DNA double-strand break repair among 859 cases and 1083 controls from the United States Radiologic Technologists (USRT) cohort. In multivariable analyses, WRN V114I (rs2230009) significantly modified the association between cumulative occupational breast dose and risk of breast cancer (adjusted for personal diagnostic exposure) (p = 0.04) and BRCA1 D652N (rs4986850), PRKDC IVS15 + 6C > T (rs1231202), PRKDC IVS34 + 39T > C (rs8178097) and PRKDC IVS31 − 634C > A (rs10109984) significantly altered the personal diagnostic radiation exposure-response relationship (adjusted for occupational dose) (p ≤ 0.05). None of the remaining 50 SNPs significantly modified breast cancer radiation dose-response relationships. The USRT genetic study provided a unique opportunity to examine the joint effects of common genetic variation and ionizing radiation exposure on breast cancer risk using detailed occupational and personal diagnostic exposure data. The suggestive evidence found for modification of radiation-related breast cancer risk for 5 of the 55 SNPs evaluated requires confirmation in larger studies of women with quantified radiation breast doses in the low-to-moderate range. © 2007 Wiley-Liss, Inc.

Settlements near the Semipalatinsk Test Site (SNTS) in northeastern Kazakhstan were exposed to radioactive fallout during 1949–1962. Thyroid disease prevalence among 2994 residents of eight villages was ascertained by ultrasound screening. Malignancy was determined by cytopathology. Individual thyroid doses from external and internal radiation sources were reconstructed from fallout deposition patterns, residential histories and diet, including childhood milk consumption. Point estimates of individual external and internal dose averaged 0.04 Gy (range 0–0.65) and 0.31 Gy (0–9.6), respectively, with a Pearson correlation coefficient of 0.46. Ultrasound-detected thyroid nodule prevalence was 18% and 39% among males and females, respectively. It was significantly and independently associated with both external and internal dose, the main study finding. The estimated relative biological effectiveness of internal compared to external radiation dose was 0.33, with 95% confidence bounds of 0.09–3.11. Prevalence of papillary cancer was 0.9% and was not significantly associated with radiation dose. In terms of excess relative risk per unit dose, our dose–response findings for nodule prevalence are comparable to those from populations exposed to medical X rays and to acute radiation from the Hiroshima and Nagasaki atomic bombings.

Data have been collected and physical and statistical models have been constructed to estimate unknown occupational radiation doses among 90,000 members of the U.S. Radiologic Technologists cohort who responded to a baseline questionnaire during the mid-1980s. Since the availability of radiation dose data differed by calendar period, different models were developed and applied for years worked before 1960, 1960- 1976 and 1977-1984. The dose estimation used available film-badge measurements (approximately 350,000) for individual cohort members, information provided by the technologists on their work history and protection practices, and measurement and other data derived from the literature. The dosimetry model estimates annual and cumulative occupational badge doses (personal dose equivalent) for each technologist for each year worked from 1916 through 1984 as well as absorbed doses to organs and tissues including bone marrow, female breast, thyroid, ovary, testes, lung and skin. Assumptions have been made about critical variables including average energy of X rays, use of protective aprons, position of film badges, and minimum detectable doses. Uncertainty of badge and organ doses was characterized for each year of each technologist's working career. Monte Carlo methods were used to generate estimates of cumulative organ doses for preliminary cancer risk analyses. The models and predictions presented here, while continuing to be modified and improved, represent one of the most comprehensive dose reconstructions undertaken to date for a large cohort of medical radiation workers.

Most information on radiation-related cancer risk comes from the Life Span Study (LSS) of the Japanese atomic bomb survivors. Stomach cancer mortality rates are much higher in Japan than in the U.S., making the applicability of LSS findings to the U.S. population uncertain. A unique cohort of U.S. patients who were irradiated for peptic ulcer to control gastric secretion provides a different perspective on risk. Cancer mortality data were analyzed and relative risks estimated for 3719 subjects treated by radiotherapy (mean stomach dose 14.8 Gy) and/or by surgery and medication during the period 1936-1965 and followed through 1997 (average 25 years). Compared to the U.S. rates, stomach cancer mortality was significantly increased for irradiated and nonirradiated patients (observed/expected = 3.20 and 1.52, respectively). We observed strong evidence of exposure-related excess mortality from cancer of the stomach (RR 2.6, 95% CI 1.3, 5.1), pancreas (RR 2.7, 95% CI 1.5, 5.1), and lung (RR 1.5, 95% CI 1.1, 2.1), with commensurate radiation dose responses in analyses that included nonexposed patients. However, the dose responses for these cancers were not significant when restricted to exposed patients. Our excess relative risk per gray estimate of 0.20 at doses