Dose response and temporal patterns of radiation-associated solid cancer risks.

Findings of the Life Span Study (LSS) cohort of atomic-bomb survivors are a primary source for quantitative risk estimates that underlie radiation protection. Because of the size and length of follow-up, the LSS provides considerable information on both the nature of the dose response and on how radiation-associated excess risks vary with age, age at exposure, sex, and other factors. Our current analyses extend the mortality follow-up by 7 y (through 1997) and add 8 y (through 1995) to the incidence follow-up. During the follow-up periods there have been a total of about 9,300 solid cancer deaths and almost 12,200 incident cases. As outlined in this presentation, while discussing issues related to the shape of the dose response and low dose risks in some detail, the new reports consider temporal patterns in greater detail than has been done previously. As we have reported, the LSS solid cancer dose response is well described by simple linear dose response over the 0 to 2 Sv range (with some leveling off at higher estimated doses). This remains the case with the extended follow-up. Although LSS is often referred to as a high dose study, about 75% of the 50,000 cohort members with doses in excess of 5 mSv have dose estimates in a range of direct interest for radiation protection (0-200 mSv). Analyses of data limited to this low dose range provide direct evidence of a significant solid cancer dose response with a risk per unit dose that is consistent with that seen for the full dose range. Previous LSS reports have focused on descriptions of the solid cancer excess risks in which the excess relative risk varies with age at exposure and sex. In addition to the age at exposure effects, our current analyses suggest excess relative risks also vary with age (at death or diagnosis). Excess relative risks are higher for those exposed earlier in life, with attained age-specific risks changing by about 20% per decade, but tend to decrease with increasing attained age, roughly in proportion to (1/attained-age)1.5, for any age at exposure. Despite the decreasing relative risk, excess rates have increased rapidly throughout the study period with some indication, especially for the incidence data, that attained-age-specific rates are higher for those exposed at younger ages. Simple comparisons of site-specific excess risks are used to illustrate how the interpretation of age-at-exposure effects on excess relative risks or excess rates is complicated by changes in baseline rates with birth cohort or time period.

Radiation-related cancer risks at low doses among atomic bomb survivors.

To clarify the information in the Radiation Effects Research Foundation data regarding cancer risks of low radiation doses, we focus on survivors with doses less than 0.5 Sv. For reasons indicated, we also restrict attention mainly to survivors within 3, 000 m of the hypocenter of the bombs. Analysis is of solid cancer incidence from 1958-1994, involving 7,000 cancer cases among 50,000 survivors in that dose and distance range. The results provide useful risk estimates for doses as low as 0.05-0.1 Sv, which are not overestimated by linear risk estimates computed from the wider dose ranges 0-2 Sv or 0-4 Sv. There is a statistically significant risk in the range 0-0.1 Sv, and an upper confidence limit on any possible threshold is computed as 0.06 Sv. It is indicated that modification of the neutron dose estimates currently under consideration would not markedly change the conclusions.

Radiation dose dependences in the atomic bomb survivor cancer mortality data: a model-free visualization.

Chomentowski, M., Kellerer, A. M. and Pierce, D. A. Radiation Dose Dependences in the Atomic Bomb Survivor Cancer Mortality Data: A Model-Free Visualization. The standard approach to obtaining nominal risk coefficients for radiation-related cancer involves fitting linear or linear-quadratic dose-response functions. This is usually complemented by a more direct visualization where the data are subdivided into distinct dose categories and the effect level is quantified for each of these categories. Such model-free computations, however, can be quite dependent on the arbitrary choice of the cutpoints in dose. The method proposed here largely avoids this arbitrariness by choosing a dose category width-constant on a log scale-to obtain the desired degree of smoothing, and then superimposing results for all placements of the resulting log-dose grids. The method is applied to leukemia and solid cancer mortality of the A-bomb survivors.

A model for radiation-related cancer suggested by atomic bomb survivor data.

The age-time patterns of excess cancer risk among A-bomb survivors followed up by the Radiation Effects Research Foundation inevitably carry implications regarding the mechanisms of radiation-related cancer. It has recently been found, quite surprisingly and in contrast to impressions given by the relative risks, that for most solid cancers the excess incidence rates themselves depend very little on age at exposure or time since exposure, but mainly on attained age. This paper investigates a mechanistic model that conforms to these age-time patterns. The essence of the model, which is highly idealized, is that: (a) a cancer is caused by mutations that accumulate in a stem cell throughout life, essentially the Armitage-Doll multistage model, and (b) radiation is a general mutagen that can cause virtually any of these mutations. Although postulate (b) departs from previous modeling considerations, the extent to which it explains various aspects of the data in substantial detail is remarkable. A strength of the model is that, similarly to the Armitage-Doll multistage model but differently from many others, it predicts characteristic age-time patterns of excess rates rather independently of its parameter values. The consequence of (a) is that, in Armitage-Doll fashion, background rates increase throughout life as a power of age. The consequence of (b) is that excess absolute rates do not depend on age at exposure and increase with age to a power one less than that of the background rates. Thus the excess relative risk, which is the ratio of these rates, decreases throughout life as 1/age with no dependence on age at exposure. Although this age pattern of relative risks corresponds closely to the RERF data for solid cancer, the interpretation of such a description is quite different from the usual one in which age at exposure plays a primary role.

Studies of the mortality of atomic bomb survivors. Report 12, part II. Noncancer mortality: 1950-1990.

This report updates the data on noncancer mortality for 86,572 atomic bomb survivors with dose estimates in the Radiation Effects Research Foundation's Life Span Study cohort. The primary analyses are based on more than 27,000 noncancer disease deaths that occurred in the cohort between October 1, 1950, and December 31, 1990, 30% more than in the previous report. The present analyses strengthen earlier findings of a statistically significant increase in noncancer disease death rates with radiation dose. Increasing trends are observed for diseases of the circulatory, digestive and respiratory systems. Rates for those exposed to 1 Sv are elevated about 10%, a relative increase that is considerably smaller than that for cancer. However, estimates of the number of radiation-related noncancer deaths in the cohort to date (140 to 280) are 50 to 100% of the number for solid cancer. The data do not yet clarify the shape of the dose response. There is no significant evidence against linearity, but the data are statistically consistent with curvilinear dose-response functions that posit essentially zero risk for doses below 0.5 Sv. Similarly, while the data are consistent with substantial variation in the excess relative risk with age at exposure or attained age, there is no statistically significant dependence on these factors. In view of the small relative risks and the lack of understanding of biological mechanisms, we emphasize consideration of whether the findings could be explained by misclassification, confounding or selection effects. Based on available data, we conclude that such factors are unlikely to fully explain the observed dose response. A significant dose response is also seen for deaths from blood diseases with an excess relative risk that is several times greater than that seen for solid cancer. Particular attention is paid to the possibility that this apparent effect is a consequence of the attribution of leukemia or other cancer deaths to noncancer blood diseases. We find that misclassification does not explain this excess risk. As in earlier reports, suicide rates tend to decrease with increasing dose.

Studies of the mortality of atomic bomb survivors. Report 12, Part I. Cancer: 1950-1990.

This continues the series of periodic general reports on cancer mortality in the cohort of A-bomb survivors followed by the Radiation Effects Research Foundation. The follow-up is extended by the 5 years 1986-1990, and analysis includes an additional 10,500 survivors with recently estimated radiation doses. Together these extensions add about 550,000 person-years of follow-up. The cohort analyzed consists of 86,572 subjects, of which about 60% have dose estimates of at least 0.005 Sv. During 1950-1990 there have been 3086 and 4741 cancer deaths for the less than and greater than 0.005 Sv groups, respectively. It is estimated that among these there have been approximately 420 excess cancer deaths during 1950-1990, of which about 85 were due to leukemia. For cancers other than leukemia (solid cancers), about 25% of the excess deaths in 1950-1990 occurred during the last 5 years; for those exposed as children this figure is nearly 50%. For leukemia only about 3% of the excess deaths in 1950-1990 occurred in the last 5 years. Whereas most of the excess for leukemia occurred in the first 15 years after exposure, for solid cancers the pattern of excess risk is apparently more like a life-long elevation of the natural age-specific cancer risk. Taking advantage of the lengthening follow-up, increased attention is given to clarifying temporal patterns of the excess cancer risk. Emphasis is placed on describing these patterns in terms of absolute excess risk, as well as relative risk. For example: (a) although it is becoming clearer that the excess relative risk for those exposed as children has declined over the follow-up, the excess absolute risk has increased rapidly with time; and (b) although the excess relative risk at a given age depends substantially on sex and age at exposure, the age-specific excess absolute risk depends little on these factors. The primary estimates of excess risk are now given as specific to sex and age at exposure, and these include projections of dose-specific lifetime risks for this cohort. The excess lifetime risk per sievert for solid cancers for those exposed at age 30 is estimated at 0.10 and 0.14 for males and females, respectively. Those exposed at age 50 have about one-third these risks. Projection of lifetime risks for those exposed at age 10 is more uncertain. Under a reasonable set of assumptions, estimates for this group range from about 1.0-1.8 times the estimates for those exposed at age 30. The excess life-time risk for leukemia at 1 Sv for those exposed at either 10 or 30 years is estimated as about 0.015 and 0.008 for males and females, respectively. Those exposed at age 50 have about two-thirds that risk. Excess risks for solid cancer appear quite linear up to about 3 Sv, but for leukemia apparent nonlinearity in dose results in risks at 0.1 Sv estimated at about 1/20 of those for 1.0 Sv. Site-specific risk estimates are given, but it is urged that great care be taken in interpreting these, because most of their variation can be explained simply by imprecision in the estimates.

Joint analysis of site-specific cancer risks for the atomic bomb survivors.

Statistical methods are presented for joint analysis of site-specific cancer risks for the atomic bomb survivors. Previous analyses of these data have been made either without regard to cancer type, excluding leukemia, or separately for types or classes of cancers. Clearly, analyses without regard to cancer type are less than satisfactory. The primary advantages of joint, rather than separate, analyses are that: (1) models can be fitted with some parameters common to cancer types and others type-specific; (2) significance tests can be used to compare type-specific risks; and (3) through consideration of more comprehensive models, a clearer understanding may be obtained of the modifying effects of sex, age at exposure, and time since exposure. Joint analysis is straightforward, entailing primarily the incorporation of another factor, cancer type, in the usual cross-tabulation of the data for analysis. The use of these methods is illustrated in an analysis of three classes of cancer studied by the National Research Council's BEIR V Committee: digestive, respiratory, and other solid tumors. Based on this analysis, some criticism is made of the BEIR V preferred models. The proposed methods are applicable to models for either relative or absolute risks, and results using both types of models are given. Although some of the gains from joint analysis are apparent from the results here, it will be important to use these methods with a more suitable choice of cancer classes and for cancer incidence data where the diagnoses are more accurate.

Understanding HCFA-2552-92 cost reporting forms.

The new cost reporting requirements necessitated by the new Medicare capital regulations may require up to 160 hours of additional data collection. Authors Powell and Pierce detail the anticipated changes to the Medicare cost reporting worksheets used to collect the data.

Analysis of time and age patterns in cancer risk for A-bomb survivors.

This report has two aims: (1) to describe and analyze the age/time patterns of excess cancer risk in the atomic bomb survivor cohort followed up by the Radiation Effects Research Foundation (RERF), and (2) to describe statistical methods which are used in RERF's analyses of data on mortality and morbidity in the cohort. In contrast to previous analyses of the cohort cancer mortality data, substantial use is made of Japanese national cancer rates for the purpose of investigation of the age/time variations in excess risk. This analysis considers mortality from all cancers except leukemia as a group. Primary attention is given to description in terms of the age-specific excess relative risk, but the importance of appropriate descriptions of the absolute excess risk is also emphasized. When models for the excess risk allow variation with age and time, both constant relative and absolute excess risk models provide similar fits to the data. Previous reports have indicated that for a given age at exposure and sex, the excess age-specific relative risk is remarkably constant throughout the current follow-up period. Statistical analysis here indicates that for those less than about 35 years of age at exposure there is no departure from this pattern, beyond ordinary sampling variation. For those over about 35 years of age at exposure, there is modest evidence of an increasing trend in the excess relative risk, which could plausibly be attributed to effects related to minimal latent period. Some brief consideration is given to modeling the absolute excess risk as the product of an age-at-exposure and time-since-exposure effect. Interpretation of these results, particularly in regard to projections beyond the current follow-up, is discussed.

The shape of the cancer mortality dose-response curve for the A-bomb survivors.

The shape of the dose-response curve for cancer mortality in the A-bomb survivor data is analyzed in the context of linear-quadratic models. Results are given for all cancers except leukemia as a group, for leukemia, and for combined inferences assuming common curvature. Since there is substantial information aside from these data suggesting a dose-response curve with upward curvature, the emphasis here is not on estimating the best-fitting dose-response curve, but rather on assessing the maximum curvature under linear-quadratic models which is consistent with the data. The apparent shape of the dose-response curve is substantially affected by imprecision in the dose estimates, and methods are applied to correct for this. The extent of curvature can be expressed as the factor by which linear risk estimates from these data should be divided to arrive at appropriate estimates of risk at low doses. Influential committees have in the past recommended ranges of 1.5-4 and of 2-10 for such a factor. Results here suggest that values greater than about 2.0-2.5 are at least moderately inconsistent with these data, within the context of linear-quadratic models. It is emphasized, however, that there is little direct information in these data regarding risks following low doses; the inferences here depend strongly on the assumption of a linear-quadratic model.

Allowing for dose-estimation errors for the A-bomb survivor data.

Unless allowances are made, random errors in radiation dose estimates cause underestimation of linear risk estimates and distort the shape of dose-response curves. These errors also result in spurious associations between radiogenic endpoints, exaggerating possible variations in individual sensitivity to radiation. Statistical methods have been developed which reduce these biases, based on assumptions regarding the nature and magnitude of dose-estimation errors. Some understanding of the underlying statistical basis for these methods is necessary to both those interested in interpreting radiogenic effects and those interested in the dosimetry system. This paper discusses the basic statistical issues and their implications, presents some statistical methods to deal with the problem, and indicates the sensitivity of certain results to assumptions about the magnitude of the dose-estimation errors.

Intron-mediated enhancement of heterologous gene expression in maize.

Chimeric genes containing the coding sequence for bacterial chloramphenicol acetyl transferase (CAT) have been introduced by electroporation into maize protoplasts (Black Mexican Sweet) and transient expression monitored by enzyme assays. Levels of CAT expression were enhanced 12-fold and 20-fold respectively by the inclusion of maize alcohol dehydrogenase-1 introns 2 and 6 in the chimeric construct. This enhancement was seen when the intron was placed within the 5' translated region but not when it was located upstream of the promoter or within the 3' untranslated region. Deletion of exon sequences adjacent to intron 2 abolished its ability to mediate enhancement of CAT gene expression. Northern analysis of protoplasts electroporated with intron constructs revealed elevated levels of CAT mRNA. However, this elevation was insufficient to account for the increased enzyme activity. One explanation of these results is that splicing affects both the quantity and quality of mRNA.

Allowing for random errors in radiation dose estimates for the atomic bomb survivor data.

The presence of random errors in the individual radiation dose estimates for the A-bomb survivors causes underestimation of radiation effects in dose-response analyses, and also distorts the shape of dose-response curves. Statistical methods are presented which will adjust for these biases, provided that a valid statistical model for the dose estimation errors is used. Emphasis is on clarifying some rather subtle statistical issues. For most of this development the distinction between radiation dose and exposure is not critical. The proposed methods involve downward adjustment of dose estimates, but this does not imply that the dosimetry system is faulty. Rather, this is a part of the dose-response analysis required to remove biases in the risk estimates. The primary focus of this report is on linear dose-response models, but methods for linear-quadratic models are also considered briefly. Some plausible models for the dose estimation errors are considered, which have typical errors in a range of 30-40% of the true values, and sensitivity analysis of the resulting bias corrections is provided. It is found that for these error models the resulting estimates of excess cancer risk based on linear models are about 6-17% greater than estimates that make no allowance for dose estimation errors. This increase in risk estimates is reduced to about 4-11% if, as has often been done recently, survivors with dose estimates above 4 Gy are eliminated from the analysis.