At the request of NASA, the National Research Council’s (NRC’s) Committee for Evaluation of Space Radiation Cancer Risk Model1 reviewed a number of changes that NASA proposes to make to its model for estimating the risk of radiation-induced cancer in astronauts. The NASA model in current use was last updated in 2005, and the proposed model would incorporate recent research directed at improving the quantification and understanding of the health risks posed by the space radiation environment. NASA’s proposed model is defined by the 2011 NASA report Space Radiation Cancer Risk Projections and Uncertainties—2010 (Cucinotta et al., 2011). The committee’s evaluation is based primarily on this source, which is referred to hereafter as the 2011 NASA report, with mention of specific sections or tables cited more formally as Cucinotta et al. (2011).
The overall process for estimating cancer risks due to low linear energy transfer (LET)2 radiation exposure has been fully described in reports by a number of organizations. They include, more recently:
• The “BEIR VII Phase 2” report from the NRC’s Committee on Biological Effects of Ionizing Radiation (BEIR) (NRC, 2006);3
• Studies of Radiation and Cancer from the United Nations Scientific Committee on the Effects of Atomic Radiation (UNSCEAR, 2006),
• The 2007 Recommendations of the International Commission on Radiological Protection (ICRP), ICRP Publication 103 (ICRP, 2007); and
• The Environmental Protection Agency’s (EPA’s) report EPA Radiogenic Cancer Risk Models and Projections for the U.S. Population (EPA, 2011).
The approaches described in the reports from all of these expert groups are quite similar. NASA’s proposed space radiation cancer risk assessment model calculates, as its main output, age- and gender-specific risk of exposure-induced death (REID) for use in the estimation of mission and astronaut-specific cancer risk. The model also calculates the associated uncertainties in REID.
1 Biographical information about the members of the committee is presented in Appendix B.
2 See Appendix C, “Glossary and Acronyms,” for definitions of terms and acronyms.
3 The BEIR VII Phase 2 report is the most recent in a series of reports by NRC committees dealing with ionizing radiation; these are widely known as the BEIR reports.
The general approach for estimating risk and uncertainty in the proposed model is broadly similar to that used for the current (2005) NASA model and is based on recommendations by the National Council on Radiation Protection and Measurements (NCRP, 2000, 2006). However, NASA’s proposed model has significant changes with respect to the following: the integration of new findings and methods into its components by taking into account newer epidemiological data and analyses, new radiobiological data indicating that quality factors differ for leukemia and solid cancers, an improved method for specifying quality factors in terms of radiation track structure concepts as opposed to the previous approach based on linear energy transfer, the development of a new solar particle event (SPE) model, and the updates to galactic cosmic ray (GCR) and shielding transport models. The newer epidemiological information includes updates to the cancer incidence rates from the life span study (LSS) of the Japanese atomic bomb survivors (Preston et al., 2007), transferred to the U.S. population and converted to cancer mortality rates from U.S. population statistics. In addition, the proposed model provides an alternative analysis applicable to lifetime never-smokers (NSs). Details of the uncertainty analysis in the model have also been updated and revised.
NASA’s proposed model and associated uncertainties are complex in their formulation and as such require a very clear and precise set of descriptions. The committee found the 2011 NASA report challenging to review largely because of the lack of clarity in the model descriptions and derivation of the various parameters used. The committee requested some clarifications from NASA throughout its review and was able to resolve many, but not all, of the ambiguities in the written description.
PROPOSED MODEL—OVERALL CONCLUSION
In considering NASA’s proposed model as a whole, the committee noted that the general approach to estimating cancer risks from exposure to low-LET radiation follows that utilized by ICRP, NCRP, EPA, and BEIR VII, and as such is state of the art. The specific data incorporated into NASA’s proposed model are generally appropriate, with some exceptions, noted below, relating to new data that have become available since the development of the model or additional data sets that were already available and not selected for use by NASA. There remains a need for development of additional data to enhance the current approach and to reduce uncertainty in the model; specific needs have been identified by the committee. The committee has some concerns about specific model components, particularly related to the change to an “incidence-mortality” approach for calculating mortality and to the risk-transfer approach used by NASA. The question of the effectiveness of the combination of the several modules into the proposed integrated model was most appropriately answered by the committee’s observing of a live demonstration by NASA of the application of the model for assessing risk to astronauts under some selected specific mission conditions. This demonstration showed that the model was indeed an integrated one—something that was not immediately apparent from the rather complex descriptions provided in the 2011 NASA report. The committee’s overall evaluation is that NASA’s proposed model represents a definite improvement over the current one. However, the committee urges that the necessary improvements identified in the specific recommendations provided below be incorporated before the proposed integrated model is implemented.
NASA’s proposed model is composed of a number of components or modules that separately address highly distinct aspects of radiation risk and uncertainty. The committee assessed each of the individual components of the model as well as the integrated model as a whole. The key results of its evaluations are summarized below. Possible improvements to components of the model and to the integrated model are provided, together with recommendations for addressing gaps in the model. In some cases, specific research is identified that could help NASA address gaps and/or uncertainties in its proposed model for cancer risk projections. The specific research identified is not necessarily a comprehensive list but is intended to include efforts that would have a significant impact and at the same time would be feasible to undertake within the short to medium term (less than 5 years). The recommendations provided in this Summary address those areas for which the committee perceived more substantial gaps or issues. The model components are discussed in more detail in the main body of the report (see Chapter 2), which contains advice in addition to the major recommendations and conclusions. It is the integrated model that will actually be implemented by NASA, and so it is also assessed in detail in Chapter 2 of this report, particularly with regard to the integration methodology.
PROPOSED MODEL—ASSESSMENT OF COMPONENTS
Tissue-Specific Particle Spectra
The committee considers that the radiation environment and shielding transport models used in NASA’s proposed model are a major step forward compared to previous models used. This is especially the case for the statistical solar particle event model. The current models have been developed by making extensive use of available data and rigorous mathematical analyses. The uncertainties conservatively allocated to the space physics parameters (i.e., environment and shielding transport models) are deemed to be adequate at this time, considering that the space physics uncertainty is only a minor contributor to the overall cancer risk assessment. Although further research in this area could reduce the uncertainty, the law of diminishing returns may prevail.
Given the above considerations, the committee does not recommend any specific research to improve the proposed model for tissue-specific particle spectra at this time. However, in this report the committee has identified several specific research areas that could improve the proposed environment models for tissue-specific particle spectra, including additional statistical analysis of the radial dependence of SPE intensity and solar-cycle dependence of SPE frequency and extreme events. The estimates could be further improved by adding physics-based studies of particle transport using the current picture of the heliosphere and its magnetic fields. Particle transport in the interplanetary medium is determined by its electric and magnetic fields. Theoretical and numerical studies of particle trajectories would certainly result in improved transport models and smaller uncertainties in the environmental estimates, but would involve a major effort and a change in modeling approach. NASA would need to weigh the added value of such an approach to its model outputs.
Cancer Risk Projection Model for Low-LET Exposures
A major change proposed in NASA’s model is to use the “incidence-mortality” approach used by BEIR VII (NRC, 2006) for the development of a REID. For this approach, risk coefficients from LSS cancer incidence models are converted into cancer mortality risks. A major reason for the use of the LSS cancer incidence data is that these are likely to be more accurate with respect to diagnosis than are mortality data, which suffer from misclassification of causes on death certificates. The approach results in considerable changes in the REID estimates, particularly in the pattern with age at exposure, and the committee considers this to be an improvement for site-specific cancer mortality estimation.
Recommendation: Before NASA implements its proposed major change to the “incidence-mortality” approach, the committee recommends that NASA conduct more research into the specific patterns of the underlying epidemiological biases that drive these changes. The committee also highlights a specific problem with the method of estimating the mortality probability from the ratio of cancer mortality to incidence as developed by the BEIR VII report published by the National Research Council in 2006 and proposed for use by NASA. In response, the committee recommends that NASA consider alternative methods for improved estimation of mortality probabilities for each cancer site. For example, as presented in its 2011 report EPA Radiogenic Cancer Risk Models and Projections for the U.S. Population, the Environmental Protection Agency has developed an alternative approach for breast cancer mortality estimation, and this could serve as a suitable approach to be applied by NASA.
Transfer of Cancer Risk Estimates from the Japanese to the U.S. Population
Because underlying cancer incidence rates for some cancer sites differ greatly between the Japanese and the U.S. populations, risk estimates based on an excess relative risk (ERR) model can give REID values very different from those based on an excess absolute risk (EAR) model. A number of organizations and committees (ICRP, the
National Council on Radiation Protection and Measurements [NCRP], BEIR VII) have recommended that a site-specific weighted average of the ERR and EAR models be used. The proposed NASA approach follows BEIR VII (NRC, 2006) in calculating a weighted average with uncertain weights and generally follows the recommended BEIR VII weights.
Recommendation: Because there are some deviations in NASA’s proposed model from the weights recommended by BEIR VII for the excess relative risk and excess absolute risk models, the committee recommends that NASA provide additional justification for these alternative weights.
Dose and Dose Rate Effectiveness Factor
A dose and dose rate effectiveness factor (DDREF) value is applied, when appropriate, to reduce the LSS-based cancer risk coefficients for protracted exposures. A median value of 1.75 was selected by NASA for its proposed model, based on an assessment made by the National Institutes of Health (NIH) for a previous estimate and its uncertainty (NIH, 2003). For its proposed model, NASA assumed that the DDREF applies only to low-LET radiations and consequently that there is no dependence of space radiation risks on dose rate. Differences in risks between space radiation charged particles and gamma rays at low dose rate are encompassed entirely within the quality factor, QF, discussed below. A number of publications issued since the NIH report are relevant to this issue, and although these were discussed in the 2011 NASA report, they were not used by NASA in its choice of DDREF or in the associated uncertainty analysis. These studies include the Mayak workers study (Shilnikova et al., 2003), the third analysis of the United Kingdom’s National Registry for Radiation Workers (Muirhead et al., 2009), and the 15-country nuclear workers study (Cardis et al., 2007), together with the review of these studies and comparison with the life span study by Jacob et al. (2009).
Conclusion: Although the proposed NASA approach for estimating a DDREF describes a number of limitations in these newer epidemiological studies and in the BEIR VII DDREF methodology, the justification given for preferring the older approach taken by the National Institutes of Health in 2003 is that it is close to the average of various recommended values of slightly less than 2. The use of this average value is somewhat problematic, given that the recommended values used to derive this average are not independent and thus applying equal weights to these is not justifiable.
Recommendation: The committee agrees with the use of an uncertainty approach for estimating DDREF, but it recommends that NASA use a central value and distribution that better accounts for the recent epidemiological and laboratory animal data.
Risk Models for Never-Smokers
The issue of the smoking status of astronauts and the potential implications for risk projections for smoking-related cancers are important, and it is appropriate that this should be investigated. Most astronauts are non-smokers, which would likely lower the risk projections for astronauts compared to estimates for the general population (a mix of never- and ever-smokers).
Recommendation: The proposed NASA approach for estimating lung cancer risks for astronauts who are never-smokers is limited and does not consider competing risks. Thus, the committee recommends that the NASA approach be developed further, given the important impact that it has on reducing estimated risk. The revised approach should use survival probabilities for competing risks that are specific to never-smokers. Further, the committee recommends that NASA make no changes at this time in the proposed model to include other smoking-related cancers. The data are not sufficiently robust for use in the modification of the REID estimate.
Uncertainties in Low-LET Cancer Risk Model and Overall Uncertainties in Cancer Risk Projections for High-LET Exposures
The 2011 NASA report addresses risk estimates and their uncertainties associated with exposure to low-LET radiation. Uncertainties are important because risk protection involves the use of safety factors, and NASA sets radiation permissible exposure limits (PELs) based on the 95 percent confidence limit that takes into account the uncertainties in risk projection models (NASA, 2005).
Uncertainty Limits and Methodology
Conclusion: Uncertainty limits on radiation-related risk reflect information about anticipated environmental radiation dose levels and accumulated knowledge about the relationship between radiation dose and cancer risk. For the approach used by NASA, more information, if available, might reduce statistical uncertainty and, assuming that the new information did not increase the central risk estimate, lower the upper 95 percent uncertainty bound criterion used by NASA to evaluate the acceptability of activity-related mortality risk.
Maximum Likelihood and Empirical Bayes Estimates
In the 2011 NASA report’s description of the proposed model, the discussion of the use of a maximum likelihood estimate (MLE) and/or empirical Bayes (EB) estimate of site-specific ERR per sievert is ambiguous with respect to the specific approach that was used in specific instances. For example, the site-specific EB estimate of ERR per sievert for kidney cancer (0.40) would be similar to the MLE (also 0.40 for this particular organ site), with a lower estimated standard error (0.19) compared to the MLE standard error of 0.32.
Recommendation: On the assumption that the empirical Bayes approach has been used in NASA’s proposed model, the committee recommends that the authors ensure that the off-diagonal covariance information has been taken into account. If the EB approach has not been used, either this fact should be stated in the text of the 2011 NASA report (Cucinotta et al., 2011) or the references to the EB approach should be removed from the text.
Uncertainty in the Value of the Quality Factor
The uncertainty analysis in NASA’s proposed model reveals that the value of the quality factor (QF, as defined in NASA’s proposed model) is the largest contributor to the uncertainty of REID, introducing about a 3.4-fold uncertainty in risk. Additional analysis by NASA (Cucinotta et al., 2011) using its proposed model finds that this component could be reduced to a 2.8-fold uncertainty if two of the track structure parameters were constrained to a fixed algebraic relationship to one another (such that the Z*2/β2 position of the maximum value of QF is held fixed). In this context, the committee notes that different values of QF are used for leukemia and solid cancers based on recent studies using animal tumor models.
Conclusion: According to NASA’s proposed model, the observation that the use of a fixed relationship between two track structure parameters reduces the uncertainty is a potentially valuable finding that may provide a method to reduce uncertainty in estimations of the risk of exposure-induced death. However, little indication is given in the 2011 NASA report as to why such a fixed position might be justified or expected. The committee suggests that further investigations into the validity and usefulness of this approach would be worthwhile.
Radiation Quality and Track Structure Risk Cross Section
The main parameter used to specify radiation quality is Z*2/β2, where Z* is the effective charge number of the particle and β its speed relative to the speed of light. Z*2/β2 replaces LET used in the conventional quality factor definition, and also by NASA in its current model. However, three additional empirical parameters (κ, Σ0/αγ and m) are introduced to define the quality factor-risk relationships as a function of Z*2/β2. For NASA’s proposed model, values for these parameters have been selected by comparison with experimentally observed variations in relative biological effectiveness (RBE) for different types of radiation for various cellular biological effects and for selected cancer types. While this approach is broadly appropriate for the proposed model parameters, the committee was unable to determine from the 2011 NASA report or from inquiries how the particular parameter values were selected.
Recommendation: The committee recommends that NASA make a detailed comparison of the relative biological effectiveness versus Z*2/β2 dependence of the experimental data with the proposed form and parameters of the quality factor, QF, equation in order to improve the transparency of the basis for the selection of the proposed parameter values for the model and to provide guidance for future research to test, validate, modify, and/or extend the parameterization. This analysis needs to include the defined selection of different values for parameters κ and Σ0/αγ for ions of Z ≤ 4 compared to all ions of higher charge.
Conclusion: In the proposed model, different maximum values of quality factor, QF, are assumed for leukemia (maximum 10) and for solid tumors (maximum 40). This is a change from the current NASA risk model. The committee agrees that it is reasonable to make such a distinction on the basis of the limited animal and human data available.
NASA’s proposed model defines a quantity that is analogous to “effective dose” as defined by ICRP, but it uses different gender-specific sets of normalized tissue weighting factors (wT) to match the estimated risks to the various tissues in representative space radiation environments. NASA proposes to use this as a summary quantity for mission operational purposes and, in NASA’s proposed model, it is simply termed “effective dose.” Effective dose is, strictly speaking, a quantity defined by ICRP that includes the ICRP-defined specification of numerical values for weighting factors and sex-averaging. If considerably different tissue weighting factors and radiation quality specifications are used and “effective dose” is evaluated without sex-averaging, it is problematic for the resulting quantity still to be termed “effective dose,” and the unit sievert given to its numerical values.
The committee believes that the NASA description of the proposed model would be improved by the use of terminology and notation that distinguish NASA-defined quantities (especially the quantity termed “effective dose”) from quantities defined by ICRP.
Non-Cancer Effects (Tissue Reactions)
In its proposed approach to estimating the safe days in deep space, NASA has used a 3 percent REID for fatal cancer as the limit. In its current model, NASA also considers dose limits for non-cancer effects—lens, skin, blood-forming organs, heart, and central nervous system. For example, “career limits for the heart are intended to limit the REID for heart disease to be below approximately 3 to 5 percent, and are expected to be largely age and sex independent” (NASA, 2005, p. 65). It was further assumed by NASA that the limits established would restrict mortality values for these non-cancer effects to less than the risk level for cancer mortality. The cancer and non-cancer risks were not combined into a single REID. More recent data have led ICRP to reconsider the threshold dose values particularly for the cardiovascular system (and cataracts) (see ICRP, 2011). It is concluded by ICRP
(2011) that a threshold absorbed dose of 0.5 Gy should be considered for cardiovascular disease (and cataracts) for acute and for fractionated/protracted exposures. It is appreciated by ICRP that these values have a degree of uncertainty associated with them.
Conclusion: The revised value for the threshold dose value proposed by ICRP suggests that NASA may need to consider how it might account for cardiovascular disease in its calculations of dose limits. However, it is noted that to date there exists very little of the information on relative biological effectiveness for non-cancer effects that is needed for estimates of risks posed by exposure to space radiation.
Delayed effects pertinent to the assessment of risk principally relate to observations whereby ongoing radiation-induced genomic instability is expressed, even at long times after radiation exposure. Such effects could have important implications for radiation protection in view of current notions of the multistep mutational processes involved in carcinogenesis. An early induced change in subsequent and ongoing mutation rates in irradiated somatic cells could accelerate this process.
Conclusion: There are conflicting reports on the generality of the phenomenon of radiation-induced delayed genomic instability and some question about variation in the susceptibilities of cells from different individuals with regard to this effect. Thus, the committee concludes that it is appropriate that genomic instability not be incorporated into NASA’s proposed model, in agreement with the proposed NASA approach. However, the committee considers that further investigation of the phenomenon is certainly warranted.
Non-targeted effects (NTEs) largely refer to the so-called bystander effects, by which responses can be produced in an unirradiated cell as a result of the transfer of a signal from an irradiated cell. For high atomic number and energy (HZE) radiations, doses that may be received by astronauts are very non-uniform in the sense that some cells will be traversed by the primary particle itself, whereas other cells will not be traversed; thus, an NTE is also a phenomenon that is of considerable interest.
Conclusion: Although the 2011 NASA report (Cucinotta et al., 2011) contains an extended discussion on non-targeted effects and their potential impact on risk estimates, NASA appropriately chose not to include these NTEs in its proposed model at this time. Little is known in qualitative or quantitative terms of the contribution of these NTEs directly related to radiation-induced carcinogenesis, but the committee believes that studies to elucidate any such relevance should be encouraged.
It is recognized that there are qualitative differences in the nature of the initial energy depositions and hence in initial chemical, biochemical, and biological damages from different types of ionizing radiation. Differences are particularly great between low-LET gamma rays and the wide variety of high-LET heavy ions in space radiation. This may lead to observed differences in responses of cells, tissues, and organisms such as differences in spectra of mutations and chromosome aberrations, altered gene-expression patterns, and different spectra and latencies for carcinogenesis. There is some experimental evidence for qualitative differences at each of the above levels of biological effect. As a result, it may not be entirely appropriate to apply universal values for quality factors as quantitative scaling factors, based on empirical data such as RBE that assume similar underlying biological processes.
The committee notes that this is an area in which experiments quantifying types, frequencies, and latencies of various cancers—for example, lung, colon, and breast cancer, with further study of liver cancer and leukemia—are sorely needed for radiations of varying LET, especially for high-LET particles at low particle fluences such as
occur in space. Furthermore, the committee suggests that the tumor studies should be coupled with appropriate mechanistic investigations to provide an understanding of the underlying carcinogenic processes.
Probabilistic Risk Assessment
The committee notes that the risk projections discussed in NASA’s proposed space radiation cancer risk assessment model and uncertainties are not presented or intended as being based on a probabilistic risk assessment (PRA) approach. NASA’s proposed model is a health-effects model intended to provide estimates of cancer risk and uncertainties for defined space radiation exposure scenarios. More generally, however, the cancer risk to astronauts is dependent on much more than a defined scenario model of health effects, with engineered barriers, in the space radiation environment. Experience with full-scope PRAs of complex systems indicates the importance of accounting for the “what can go wrong during actual operations” scenarios, as such scenarios generally drive the overall risk. Thus, the committee suggests that comprehensive, mission-specific PRAs also be considered so as to enable accountability for the “what can go wrong” scenarios in the overall risk projections.
Cardis, E., Vrijheid, M., Blettner, M., Gilbert, E., Hakama, M., Hill, C., Howe, G., Kaldor, J., Muirhead, C.R., Schubauer-Berigan, M., and Yoshimura, T., et al. 2007. The 15-Country Collaborative Study of Cancer Risk among Radiation Workers in the Nuclear Industry: Estimates of Radiation-Related Cancer Risks. Radiation Research 167(4):396-416.
Cucinotta, F.A., Kim, M.-H.Y., and Chappell, L.J. 2011. Space Radiation Cancer Risk Projections and Uncertainties—2010. NASA/TP-2011-216155. NASA Johnson Space Center, Houston, Tex. July.
EPA (Environmental Protection Agency). 2011. EPA Radiogenic Cancer Risk Models and Projections for the U.S. Population. U.S. Environmental Protection Agency, Washington, D.C.
ICRP (International Commission on Radiological Protection). 2007. The 2007 Recommendations of the International Commission on Radiological Protection. ICRP Publication 103. Ann ICRP 37 (2-4). International Commission on Radiological Protection, Ottawa, Ontario, Canada.
ICRP. 2011. Early and Late Effects of Radiation in Normal Tissues and Organs: Threshold Doses for Tissue Reactions in a Radiation Protection Context. Draft Report for Consultation. ICRP Ref 4844-6029-7736. International Commission on Radiological Protection, Ottawa, Ontario, Canada. January 20.
Jacob, P., Ruhm, W., Walsh, L., Blettner, M., Hammer, G., and Zeeb, H. 2009. Is cancer risk of radiation workers larger than expected? Occupational and Environmental Medicine 66:789-796.
Muirhead, C.R., O’Hagan, J.A., Haylock, R.G.E., Phillipson, M.A., Willcock, T., Berridge, G.L.C., and Zhang, W. 2009. Mortality and cancer incidence following occupational radiation exposure: Third analysis of the National Registry for Radiation Workers. British Journal of Cancer 100:206-212.
NASA (National Aeronautics and Space Administration). 2005. NASA Space Flight Human System Standard, Volume 1: Crew Health. NASA-STD-3001. (Approved 03-05-2007) NASA, Washington, D.C.
NCRP (National Council on Radiation Protection and Measurements). 2000. Radiation Protection Guidance for Activities in Low-Earth Orbit. NCRP Report No. 132. NCRP, Bethesda, Md.
NCRP. 2006. Information Needed to Make Radiation Protection Recommendations for Space Missions Beyond Low-Earth Orbit. NCRP Report No. 153. NCRP, Bethesda, Md.
NIH (National Institutes of Health). 2003. Report of the NCI-CDC Working Group to Revise the 1985 NIH Radioepidemiological Tables. NIH Publication No. 03-5387. Bethesda, Md.
NRC (National Research Council). 2006. Health Risks from Exposure to Low Levels of Ionizing Radiation: BEIR VII Phase 2. The National Academies Press, Washington, D.C.
Preston, D.L., Ron, E., Tokuoka, S., Funamoto, S., Nishi N., Soda, M., Mabuchi, K., and Kodama, K. 2007. Solid cancer incidence in atomic bomb survivors: 1958-1998. Radiation Research 168:1-64.
Shilnikova, N.S., Preston, D.L., Ron, E., Gilbert, E.S., Vassilenko, E.K., Romanov, S.A., Kuznetsova, I.S., Sokolnikov, M.E., Okatenko, P.V., Kreslov, V.V., and Koshurnikova, N.A. 2003. Cancer mortality risk among workers at the Mayak Nuclear Complex. Radiation Research 159(6):787-798.
UNSCEAR (United Nations Scientific Committee on the Effects of Atomic Radiation). 2006. Studies of Radiation and Cancer. Report to the General Assembly, with Scientific Annexes A and B. United Nations. New York.