Epidemiological studies of natural sources of radiation and childhood cancer : current challenges and future perspectives

Journal of Radiological Protection ACCEPTED MANUSCRIPT Epidemiological studies of natural sources of radiation and childhood cancer: current challenges and future perspectives To cite this article before publication: Antonella Mazzei-Abba et al 2019 J. Radiol. Prot. in press https://doi.org/10.1088/1361-6498/ab5a38 Manuscript version: Accepted Manuscript Accepted Manuscript is “the version of the article accepted for publication including all changes made as a result of the peer review process, and which may also include the addition to the article by IOP Publishing of a header, an article ID, a cover sheet and/or an ‘Accepted Manuscript’ watermark, but excluding any other editing, typesetting or other changes made by IOP Publishing and/or its licensors” This Accepted Manuscript is © 2019 Society for Radiological Protection. Published on behalf of SRP by IOP Publishing Limited. All rights reserved.. During the embargo period (the 12 month period from the publication of the Version of Record of this article), the Accepted Manuscript is fully protected by copyright and cannot be reused or reposted elsewhere. As the Version of Record of this article is going to be / has been published on a subscription basis, this Accepted Manuscript is available for reuse under a CC BY-NC-ND 3.0 licence after the 12 month embargo period. After the embargo period, everyone is permitted to use copy and redistribute this article for non-commercial purposes only, provided that they adhere to all the terms of the licence https://creativecommons.org/licences/by-nc-nd/3.0 Although reasonable endeavours have been taken to obtain all necessary permissions from third parties to include their copyrighted content within this article, their full citation and copyright line may not be present in this Accepted Manuscript version. Before using any content from this article, please refer to the Version of Record on IOPscience once published for full citation and copyright details, as permissions will likely be required. All third party content is fully copyright protected, unless specifically stated otherwise in the figure caption in the Version of Record. View the article online for updates and enhancements. This is the accepted manuscript published in https://doi.org/10.1088/1361-6498/ab5a38 Published 19 February 2020, no 40. Journal of Radiological Protection 1 Review article Epidemiological studies of natural sources of radiation and childhood cancer: current challenges and future perspectives Antonella Mazzei-Abba1, Christophe L. Folly1, Astrid Coste1, Richard Wakeford2, Mark P. Little3, Ole Raaschou-Nielsen4, Gerry Kendall5, Denis Hémon6, Atte Nikkilä7, Claudia Spix8, Anssi Auvinen9, Ben D. Spycher1 1 Institute of Social and Preventive Medicine (ISPM), University of Bern, Bern, Switzerland 2 Centre for Occupational and Environmental Health, Institute of Population Health, University of Manchester, UK 3 Radiation Epidemiology Branch, Division of Cancer Epidemiology and Genetics, National Cancer Institute, Bethesda, USA 4 Institute of Cancer Epidemiology, Danish Cancer Society, Denmark 5 Cancer Epidemiology Unit, Nuffield Department of Population Health, University of Oxford, UK 6 Inserm, UMR 1153 Epidemiology and Biostatistics Sorbonne Paris Cité Research Center (CRESS), Epidémiologie des cancers de l’enfant et de l’adolescent Team (EPICEA), Villejuif, F-94807, France; Paris Descartes University, Sorbonne Paris Cité, France 7 Tampere Center for Child Health Research, Tampere University and Tampere University Hospital, Tampere, Finland 8 German Childhood Cancer Registry, Institute for Medical Biostatistics, Epidemiology and Informatics, University Medical Center Mainz, Germany 9 Unit of Health Sciences, Faculty of Social Sciences, Tampere University, Tampere, Finland and STUK - Radiation and Nuclear Safety Authority, Helsinki, Finland Corresponding author: Ben D Spycher Institute of Social and Preventive Medicine (ISPM), University of Bern, Mittelstrasse 43, 3012 Bern, Switzerland. E-mail: ben.spy[email protected].ch Tel: +41 31 631 33 46 Fax: +41 31 631 35 20 Page 1 of 35 AUTHOR SUBMITTED MANUSCRIPT - JRP-101695.R1 Journal of Radiological Protection 2 Abbreviations ALAcute leukaemia ALL – Acute lymphoid leukaemia AML – Acute myeloid leukaemia ANLL – Acute non-lymphoblastic leukaemia CNS – Central Nervous System GB – Great Britain LSS – Life Span Study RBM – Red bone marrow Page 2 of 35AUTHOR SUBMITTED MANUSCRIPT - JRP-101695.R1 Journal of Radiological Protection 3 Abstract Empirical estimation of cancer risks in children associated with low-dose ionizing radiation (<100 mSv) remains a challenge. The main reason is that the required combination of large sample sizes with accurate and comprehensive exposure assessment is difficult to achieve. An international scientific workshop “Childhood cancer and background radiation” organised by the Institute of Social and Preventive Medicine of the University of Bern brought together researchers in this field to evaluate how epidemiological studies on background radiation and childhood cancer can best improve understanding of the effects of low-dose ionising radiation. This review summarises and evaluates the findings of the existing studies in the light of their methodological differences, identifies key limitations and challenges and proposes ways forward. Large childhood cancer registries, such as those in Great Britain, France and Germany, now allow the conducting of studies that should have sufficient statistical power to detect the effects predicted by standard risk models. Nevertheless, larger studies or pooled studies will be needed to investigate disease subgroups. The main challenge is to accurately assess children’s individual exposure to radiation from natural sources and from other sources, as well as potentially confounding non-radiation exposures, in such large study populations. For this, the study groups should learn from each other to improve exposure estimation and develop new ways to validate exposure models with personal dosimetry. Keywords: childhood cancer, background ionising radiation, exposure assessment, record-based study Page 3 of 35 AUTHOR SUBMITTED MANUSCRIPT - JRP-101695.R1 Journal of Radiological Protection 4 Introduction Exposure to high doses of ionizing radiation is known to increase the risk of cancer. Standard risk models based on data from the Life Span Study (LSS) of atomic bomb survivors from Hiroshima and Nagasaki are broadly consistent with a linear/linear-quadratic increase in cancer risk with dose. The excess relative risk (ERR) per gray is modified by sex, age at exposure and time since exposure (1, 2). These variations are pronounced for leukaemia: ERR/Gy is highest after exposure in childhood and reaches a peak some 5 years after exposure (with ERR estimates of about 50 per Sv), declining thereafter (3). Evidence from other studies is also consistent with a higher risk of radiation inducedcancer after exposure during childhood compared to exposure in later life for various cancer types including leukaemia, thyroid, skin, breast and brain cancer (4). The empirical estimation of excess cancer risks associated with low doses (<100 mGy low-LET radiation) is more difficult due to sample size requirements and the challenge of reliable dosimetry. However, a recent pooled analysis of nine cohort studies with individual dosimetry, including over 260,000 people exposed to low doses during childhood from medical exposure and from the atomic bombs, found evidence of excess risks associated with doses of less than 50 mSv for acute leukaemia (5). The pooled analysis included a large cohort study from the UK that reported that cumulative doses to the red bone marrow (RBM) of about 50 mGy and to the brain of about 60 mGy (2-3 head CT scans) might almost triple the risk of leukaemia and brain tumours (6). A recent nationwide cohort study in the Netherlands included 168,394 children who received one or more CT scans also reported that brain doses of about 20–50 mGy may increase brain tumour risk (7), but no association was observed for leukaemia. However, results from studies of paediatric CT scans need to be interpreted with caution, because of the potential for reverse causation and confounding by indication (8, 9). Children’s heightened susceptibility combined with short latency periods suggest that a meaningful proportion of leukaemia cases in children, and possibly also of central nervous system (CNS) tumours, might be caused by exposure to natural sources of radiation. Indeed, based on standard risk models, studies from Great Britain (GB) and France estimate this proportion to be up to about 20% (10, 11), and in Finland estimates were about 5% (unpublished results), albeit all with large uncertainties. Page 4 of 35AUTHOR SUBMITTED MANUSCRIPT - JRP-101695.R1 Journal of Radiological Protection 5 However, in the LSS, which commenced just over 5 years after the bombings and upon which the standard leukaemia risk models are based, only four cases of leukaemia occurred among survivors with attained age <10 years and their RBM doses were >1 Sv (12). So, caution is required in applying these risk models to children receiving very low annual doses. Most of the previous ecological studies investigating associations between childhood leukaemia and naturally occurring sources of ionising radiation have found positive associations for radon (13-15) while for gamma radiation and cosmic rays results have been inconsistent (16-22). Early case-control studies of the association between natural sources of radiation and childhood leukaemia were underpowered and have reported mixed results (23-26). The largest of these, the UK Childhood Cancer Study, included over 2000 cases of childhood cancer and reported weak evidence of a negative association between childhood leukaemia and measured radon concentrations (25) but no evidence of an association with measured gamma dose rates (26). However, the proportion of eligible subjects participating in the measurements was low and varied by socio-economic status. Because exposure to these sources is ubiquitous and variation in cumulative doses received by children of similar age are small, large sample sizes are needed to detect the small predicted risk. Given the rarity of childhood cancer, the only way to achieve such sample sizes is by combining data over long periods of systematic cancer registration. In the last decade, several nationwide record-based studies in Europe, also referred to as registry-based or register-based studies, have investigated associations between childhood cancer and natural sources of radiation including gamma radiation (with or without the cosmic component) (27-31) and domestic radon (27, 30, 32, 33). In contrast to questionnaireor interview-based studies, record-based studies rely for the most part on comprehensive data compiled systematically for the entire population and do not require any active participation by study members. Exposure prediction models are used to estimate residential exposure to different sources of background radiation. The Institute of Social and Preventive Medicine of the University of Bern, Switzerland, organized the international scientific workshop “Childhood cancer and background radiation” on June 6th, 2018. The aim of the workshop was to bring together researchers in the field and interested parties from Page 5 of 35 AUTHOR SUBMITTED MANUSCRIPT - JRP-101695.R1 Journal of Radiological Protection 6 around the world to discuss how epidemiological studies on background radiation and childhood cancer can improve understanding of the effects of low-dose ionising radiation. The studies presented at the workshop represent all studies that have been conducted based on nationwide registration of childhood cancers. The purpose of this review is to describe the findings of the existing studies and their methodological differences, identify limitations and challenges, and propose ways forward in this area of research. In the first section, we describe the methods and findings of the studies. The second section highlights the main methodological challenges, providing an inside view from the authors and presenters of the workshop. Finally, we conclude with future perspectives and recommendations for further research. Review of recent record-based studies In this section, we briefly summarize each of the record-based studies on natural sources of radiation and childhood cancer in chronological order of their publication (27-33). An overview of the methodological characteristics and findings is presented in Tables 1 and 2. More details on methods of assessing exposure to gamma radiation and residential radon are provided in Tables 3 and 4. All studies used cancer registries with high completeness to identify cases (31, 34-38). The first one was a Danish case-control study that examined domestic radon exposure (33). It included 2,400 childhood malignancies (leukaemia, CNS, and malignant lymphoma) diagnosed in 1968-1994. Control children were selected from the Danish Central Population Registry matching on sex and year of birth. Exposure assessment covered all residences in which the child had lived between birth and diagnosis (or equivalent date). Domestic radon exposure was estimated using a regression model developed from measurements in the living rooms of 3,116 Danish dwellings, with predictors including geographical region, soil type, and house characteristics (39). The study found a relative risk (RR) of 1.56 per cumulative exposure of 103 Bq/m3-years (95% confidence interval (CI): 1.05, 2.30) for acute lymphoblastic leukaemia (ALL). No association was observed for childhood acute nonlymphoblastic leukaemia (ANLL) or brain/CNS tumours, with RRs of 0.75 (95% CI: 0.34, 1.62) and 0.92 (95% CI: 0.69, 1.22) per 103 Bq/m3-years, respectively. The exposure model performed relatively Page 6 of 35AUTHOR SUBMITTED MANUSCRIPT - JRP-101695.R1 Journal of Radiological Protection 7 well in predicting radon concentrations in a test sample of 758 independent measurements (coefficient of determination R2 = 0.45) (39). Background gamma radiation was not assessed. A large record-based case-control study from Great Britain (GB: England, Wales and Scotland) investigated associations of indoor radon and gamma exposure with various childhood cancers (30). It included 27,447 children diagnosed with cancer during 1980–2006 of which 9,058 were childhood leukaemia. For each case, a control was selected from the same birth register matching for sex and date of birth (within six months). Radiation exposures were estimated for mother’s residence at the child’s birth. Exposure to gamma radiation was estimated using the County District mean dose rates based on 2,283 indoor measurements made throughout GB. Exposure to radon was estimated using a predictive map based on approximately 400,000 measurements in homes throughout GB. Cumulative doses to the RBM since conception were calculated assuming residential exposures at the same dose rate as the residence of birth. The authors reported a RR for childhood leukaemia of 1.12 (95% CI: 1.03, 1.22) per mSv cumulative equivalent dose to the RBM from terrestrial gamma and 1.03 (95% CI: 0.96, 1.11) for RBM dose from domestic radon. In Switzerland, two studies, one on radon and another on gamma radiation, were conducted using data from a census-based cohort study (29, 32). Cases of childhood cancer were identified through probabilistic record linkage with the Swiss Childhood Cancer Registry (SCCR). Exposure to residential radon was estimated using a prediction model based on 35,706 indoor measurements and soil and building characteristics (tectonic units, soil texture, floor level and degree of urbanization) as predictors. In internal validation, the radon model had a relatively low R2 of 0.2. Outdoor dose rates from terrestrial gamma and cosmic radiation were estimated using a map developed by interpolation based on a diverse set of measurements including airborne spectrometry (40). Change of residence between censuses, but not full lifetime residential history, was taken into account to calculate timevarying cumulative exposure. The study on radon exposure included 997 cases of childhood cancer and found no evidence of an association, neither for all cancers combined, nor for leukaemia nor CNS tumours. The study on exposure to gamma radiation included 1,782 cases and found evidence of associations for leukaemia and CNS tumours: for both diagnostic groups a RR of about 1.04 (95% CI: 1.00, 1.08) per mSv cumulative whole-body dose was estimated. Page 7 of 35 AUTHOR SUBMITTED MANUSCRIPT - JRP-101695.R1 Journal of Radiological Protection 8 A nationwide case-control study in Finland investigated association between exposure to gamma radiation and childhood leukaemia. It included 1,093 cases diagnosed over the period 1990-2011 (28). Three controls per case, individually matched on year of birth and sex, were selected from the national population register. Exposure was assessed using a map of terrestrial gamma radiation dose rate and a map of Chernobyl fallout. For cases with partially unknown residential history, municipal averages of terrestrial gamma dose rates were used. Exposure assessment accounted for type of building regarding shielding and the radiation from the building materials. Full residential history was available to calculate cumulative dose to the RBM. Overall, there was no evidence of an association for childhood leukaemia (RR 1.01, 95% CI: 0.97, 1.05 for 10 nSv/h increase in average equivalent dose rate to RBM). In subgroup analyses, leukaemia diagnosed at ages 2-6 years was associated with cumulative dose to the RBM (RR 1.27, 95% CI: 1.01, 1.60 per mSv). In France, a nationwide case-control study investigated the association of childhood acute leukaemia and background radiation (27). It included 2,761 cases diagnosed during 2002-2007 and 30,000 controls sampled from a national dataset of households. Exposure to both radon and gamma radiation was based on cokriging models that combined indoor measurements – 17,404 for gamma (41) and 10,843 for radon (42) – with a map of geogenic radon/uranium potential (R²=0.32 for radon; R²=0.65 for gamma). Exposure was assessed based on residence at time of diagnosis. Cumulative doses to the RBM from radon and gamma radiation were calculated assuming constant place of residence since birth. The authors reported an RR for childhood acute leukaemia of 1.00 (95% CI: 0.98, 1.01) per nSv/h of gamma radiation and 0.98 (95% CI: 0.90, 1.07) per 100 Bq/m3 of radon concentration. Recently, two ecological studies were conducted in France and Germany. In France, the ecological study assessed cancer risks and exposure across 36,326 municipalities and included 9,056 cases diagnosed between 1990 and 2009. Results were published in parallel with the aforementioned casecontrol study (27) and used the same exposure models, but exposure was determined at the town centre for radon and at municipality-level means for gamma. Among the 30,000 controls in the casecontrol study, the municipality-based estimates of exposure used in the ecological study correlated strongly with the estimates of exposure based on residential addresses: r = 0.975 for gamma exposure and r = 0.991 for radon exposure. The results were consistent in both studies and neither showed any Page 8 of 35AUTHOR SUBMITTED MANUSCRIPT - JRP-101695.R1 Journal of Radiological Protection 15 The best performance was found for a linear model based on weighted sums of gamma dose rates among neighbouring measurement points and other simple models (53), which might be used in future studies. In France, the model used to estimate gamma exposure was based on Warnery et al. (41) and validated against an independent set of 8,839 dwelling indoor measurements (54). A relatively good correlation (r = 0.59) between estimates and measurements was observed, but there was a significant difference in mean dose rates (76 vs 55 nSv/h), possibly reflecting a difference between the dental surgeries and veterinary clinics, where the measurements used for model development had been made, and dwellings. In a sensitivity analysis, using an exposure model based on these 8,839 measurements within dwellings (unpublished results) rather than on the 17,404 ones used in the published analyses (27, 54), the findings of the study were unchanged. To date, there has been no validation of the exposure models used in the reviewed studies based on personal dosimetry in children. Such a study could help better understand the errors of the exposure models (55). Neglected sources of exposure Doses from medical uses of radiation and ingested radionuclides have been largely neglected in studies of cancer risks from background radiation, because data acquisition is exceedingly difficult, particularly without active participation of the study population. The Finnish study evaluated various hypothetical bias scenarios due to doses from CT scan examinations, but results were not materially affected (28). To the extent that omitted exposures correlate with the exposures that were assessed and included in regression models, estimated dose response relationships may be biased. Given a likely correlation between exposure to residential radon and exposure to terrestrial gamma radiation (30) , (organ-specific) doses from these sources should be combined, but dose conversion models for radon exposure are not well established as yet. Detailed information regarding possible correlations between these exposures and doses from ingested radionuclides or from medical radiation is lacking. A correlation of exposure to gamma radiation or domestic radon with doses from ingested radionuclides is plausible as the latter may also depend on local or regional concentrations of naturally occurring radionuclides. Such correlations would be more likely if consumed food products are grown locally Page 15 of 35 AUTHOR SUBMITTED MANUSCRIPT - JRP-101695.R1 Journal of Radiological Protection 16 and drinking water is sourced from local aquifers. To some extent homeostatic control of 40K concentrations will reduce the variation in doses from internal emitters in the body, but such mechanisms do not apply to other radionuclides (56). Residential mobility and timing of exposure assessment Another important question that arises in studies of childhood cancer and natural sources of radiation is whether complete residential histories are required for accurate assessment of cancer risks. Radiation doses from gamma and radon exposure are received continuously over the whole lifetime at dose rates that are approximately constant at a given residential location (although radon remediation measures could substantially reduce radon exposure). The extent to which cancer risks at a given attained age depend on doses received at earlier ages remains unclear. Existing models suggest that these relationships differ considerably between cancer types (57). In the absence of an agreed alternative weighting scheme, cumulative doses are calculated by (unweighted) integration of doserates from conception (or birth) to attained age. Ideally, this calculation should be based on full residential history. However, such data are only available in a few, mainly Nordic countries. In Finland, Nikkilä et al. examined the effects of incomplete residential histories on studies of background radiation (58). About 48% of cases and controls had lived only at one address and those who had relocated generally only moved short distances (median 4km, mean 40km) resulting in small differences in exposure levels between successive addresses. Similarly, Demoury et al. found only about 34% of children moved to another municipality between birth and diagnosis, and that there was a high correlation between exposures at birth and at diagnosis or at inclusion in control group: the Pearson correlation coefficient was 0.86 for radon and 0.89 for gamma radiation (27). Of the childhood cancer cases in the study from GB, 44% had not moved residence between birth and diagnosis, and about two-thirds were living at diagnosis within 2 km of their birth address (43). Thus, in the absence of data on full address histories, the estimation from a single address, despite introducing measurement error, should still capture a large proportion of exposure variability between individuals. Page 16 of 35AUTHOR SUBMITTED MANUSCRIPT - JRP-101695.R1 Journal of Radiological Protection 17 Potential confounding Potential confounders comprise non-radiation risk factors that are associated with the primary disease endpoint and with determinants of radiation exposure from natural sources, such as residential location, dwelling characteristics and inhabitants’ living habits. Some factors for which a link with childhood cancer is supported by the literature and that might be associated with radiation exposure (29, 59) include traffic-related air pollution (60), pesticides (61), exposure to infections (62-66) and socioeconomic status (SES) (59, 67-69). Such association may also exist for other factors discussed in the literature of childhood cancer, for instance: genetic syndromes (47) and birth weight (70). Although all studies had considered some of these factors, it is difficult for a single study to include all (Table 1). The studies from France, Switzerland and Denmark included a broad range of covariates that showed some correlation with gamma or radon exposure. However, these adjustments had little effect on estimates of interest. Overall assessment of potential errors and bias Despite the methodological challenges, record-based studies have potential to detect and quantify childhood cancer risks associated with natural sources of radiation. First, by design, these studies are virtually free of selection bias (assuming complete cancer registries and random sample of representative controls) and the larger studies are adequately powered. Though exposure assessment is difficult, we would argue that the consequences of measurement errors may not be as severe as one might expect. The methods of estimating individual exposure to natural sources of radiation in recordbased studies involve interpolation, smoothing of measurements, and thus have a tendency for regression to the (local) mean. Arguably, therefore, the dominant component of non-systematic exposure measurement error in the discussed studies is of Berkson type, which results in reduced precision and statistical power, rather than of classical type, which would lead to bias towards the null. Furthermore, exposure models that, to a certain extent, smooth out small-scale variation may even improve precision of individual exposure assessment, because children’s true exposure is a timeweighted average of exposures at the various locations (indoors and outdoors) where they spend most of their time. Lastly, the risk of confounding may be minimal, because for most of the suspected risk Page 17 of 35 AUTHOR SUBMITTED MANUSCRIPT - JRP-101695.R1 Journal of Radiological Protection 18 factors neither the correlations between background gamma or radon nor the effects of the latter on the risks of childhood cancer are likely to be strong. Furthermore, in none of the studies did adjustment for potential confounders, such as SES, appreciably alter effect estimates. Despite these grounds for optimism, the excess cancer risks associated with natural background radiation are expected to be small and, consequently, even small biases from unmeasured confounding or measurement error could obscure the true effects. In consequence, the potential for bias should not be neglected. Indeed, the discrepancies between the results of the reviewed studies might suggest that systematic errors are at work in some way. Systematic errors in exposure estimates might occur if the measurements on which these are based are not representative of exposure levels at the locations where the study subjects spend much of their time. However, such errors are unlikely to cause bias in effect estimates unless they differ systematically between cases and controls. Other potential sources of systematic error could include regional differences in cancer registration coverage that correlate with natural sources of radiation levels, neglected exposures, large-scale confounding, ecologic bias or biases associated with aggregating (or over-smoothing) the exposure (71), and sampling variation, among others. Conclusions and future perspectives Recent studies on exposure to natural sources of radiation and childhood cancer have shown conflicting results, which remain to be resolved. As we have outlined, these studies face some common methodological challenges that should be addressed in future research. We propose some steps forward in Box 1. Thanks to the early establishment of national cancer registries in some countries, the challenge of achieving sufficient statistical power can now be met. Nevertheless, still larger studies or the pooling of studies will be needed to investigate disease subgroups. Currently, the greater challenge is to accurately assess children’s exposure for such large study populations. For this, the study groups should learn from each other and join in concerted efforts to improve exposure estimation and look for new ways to validate these models with personal dosimetry. Quantitative analysis of potential biases associated with exposure misclassification and unmeasured confounding could shed light on existing inconsistencies and help study designs in future studies. By Page 18 of 35AUTHOR SUBMITTED MANUSCRIPT - JRP-101695.R1 Journal of Radiological Protection 19 addressing these challenges, we are reasonably confident that studies on exposure to natural sources of radiation and cancer risks in children can provide an evidence base for a better understanding of the effects of low dose ionizing radiation. Page 19 of 35 AUTHOR SUBMITTED MANUSCRIPT - JRP-101695.R1 Journal of Radiological Protection 20 Acknowledgments The authors are grateful to Dr. Ausrele Kesminiene for her contribution to the workshop as a speaker, and to all the participants in the workshop for their comments to the discussion. The authors would like to thank Mathias Egger and Claudia Kühni for chairing the sessions, Christian Kreis and Garyfallos Konstantinoudis for the support on the organization of the workshop. Finally, the authors would like to thank the Institute of Social and Preventive Medicine for providing the venue for the workshop. The authors would also like to thank the two referees for their detailed and helpful comments. Funding The workshop was held in the framework of a project entitled “Low dose ionising radiation and the risk of childhood cancer”, which is funded by the Swiss National Science Foundation (Grant No. 320030_176218). The workshop was generously supported by the Swiss Cancer League and the Swiss Cancer Research foundation (ABD-4465-01-2018). The attendance of Dr. Little was supported in part by the Intramural Research Program of the National Institutes of Health, the National Cancer Institute, Division of Cancer Epidemiology and Genetics. Page 20 of 35AUTHOR SUBMITTED MANUSCRIPT - JRP-101695.R1 Journal of Radiological Protection 21 References 1. UNSCEAR. UNSCEAR Report 2006, Effects of ionizing radiation, Vol. I Annex A, Epidemiological studies of radiation and cancer. United Nations Scientific Committee on the Effects of Atomic Radiation (UNSCEAR), United Nations; 2006. 2. Council NR. Health Effects of Exposure to Low Levels of Ionizing Radiation: BEIR V. Washington, DC: The National Academies Press; 1990. 436 p. 3. Richardson D, Sugiyama H, Nishi N, Sakata R, Shimizu Y, Grant EJ, et al. Ionizing radiation and leukemia mortality among Japanese Atomic Bomb Survivors, 1950-2000. Radiation research. 2009;172(3):368-82. 4. UNSCEAR. UNSCEAR Report 2013, Sources, effects and risks of ionizing radiation, vol. II Annex B, Effects of radiation exposure of children. New York: United Nations Scientific Committee on the Effects of Atomic Radiation (UNSCEAR), United Nations; 2013. 5. Little MP, Wakeford R, Borrego D, French B, Zablotska LB, Adams MJ, et al. Leukaemia and myeloid malignancy among people exposed to low doses (< 100 mSv) of ionising radiation during childhood: a pooled analysis of nine historical cohort studies. Lancet Haematol. 2018;5(8):E346-E58. 6. Pearce MS, Salotti JA, Little MP, McHugh K, Lee C, Kim KP, et al. Radiation exposure from CT scans in childhood and subsequent risk of leukaemia and brain tumours: a retrospective cohort study. Lancet. 2012;380(9840):499-505. 7. Meulepas JM, Ronckers CM, Smets A, Nievelstein RAJ, Gradowska P, Lee C, et al. Radiation Exposure From Pediatric CT Scans and Subsequent Cancer Risk in the Netherlands. Journal of the National Cancer Institute. 2019;111(3):256-63. 8. Journy N, Rehel JL, Ducou Le Pointe H, Lee C, Brisse H, Chateil JF, et al. Are the studies on cancer risk from CT scans biased by indication? Elements of answer from a large-scale cohort study in France. British journal of cancer. 2015;112(1):185-93. 9. Berrington de Gonzalez A, Salotti JA, McHugh K, Little MP, Harbron RW, Lee C, et al. Relationship between paediatric CT scans and subsequent risk of leukaemia and brain tumours: assessment of the impact of underlying conditions. British journal of cancer. 2016;114(4):388-94. 10. Little MP, Wakeford R, Kendall GM. Updated estimates of the proportion of childhood leukaemia incidence in Great Britain that may be caused by natural background ionising radiation. Journal of radiological protection : official journal of the Society for Radiological Protection. 2009;29(4):467-82. 11. Laurent O, Ancelet S, Richardson DB, Hemon D, Ielsch G, Demoury C, et al. Potential impacts of radon, terrestrial gamma and cosmic rays on childhood leukemia in France: a quantitative risk assessment. Radiation and environmental biophysics. 2013;52(2):195-209. 12. Walsh L, Kaiser JC. Multi-model inference of adult and childhood leukaemia excess relative risks based on the Japanese A-bomb survivors mortality data (1950-2000). Radiation and environmental biophysics. 2011;50(1):21-35. 13. Laurier D, Valenty M, Tirmarche M. Radon exposure and the risk of leukemia: a review of epidemiological studies. Health physics. 2001;81(3):272-88. 14. Raaschou-Nielsen O. Indoor radon and childhood leukaemia. Radiation protection dosimetry. 2008;132(2):175-81. 15. Tong J, Qin L, Cao Y, Li J, Zhang J, Nie J, et al. Environmental radon exposure and childhood leukemia. Journal of toxicology and environmental health Part B, Critical reviews. 2012;15(5):332-47. 16. Mason TJ, Miller RW. Cosmic radiation at high altitudes and U.S. cancer mortality, 1950-1969. Radiation research. 1974;60(2):302-6. 17. Tirmarche M, Rannou A, Mollie A, Sauve A. Epidemiological-Study of Regional Cancer Mortality in France and Natural Radiation. Radiation protection dosimetry. 1988;24(1-4):479-82. 18. Hatch M, Susser M. Background gamma radiation and childhood cancers within ten miles of a US nuclear plant. International journal of epidemiology. 1990;19(3):546-52. 19. Muirhead CR, Butland BK, Green BMR, Draper GJ. An Analysis of Childhood Leukemia and Natural Radiation in Britain. Radiation protection dosimetry. 1992;45(1-4):657-60. Page 21 of 35 AUTHOR SUBMITTED MANUSCRIPT - JRP-101695.R1 Journal of Radiological Protection 22 20. Richardson S, Monfort C, Green M, Draper G, Muirhead C. Spatial variation of natural radiation and childhood leukaemia incidence in Great Britain. Statistics in medicine. 1995;14(2122):2487-501. 21. Evrard AS, Hemon D, Billon S, Laurier D, Jougla E, Tirmarche M, et al. Childhood leukemia incidence and exposure to indoor radon, terrestrial and cosmic gamma radiation. Health physics. 2006;90(6):569-79. 22. Auvinen A, Hakama M, Arvela H, Hakulinen T, Rahola T, Suomela M, et al. Fallout from Chernobyl and incidence of childhood leukaemia in Finland, 1976-92. Bmj. 1994;309(6948):151-4. 23. Little MP, Wakeford R, Lubin JH, Kendall GM. The statistical power of epidemiological studies analyzing the relationship between exposure to ionizing radiation and cancer, with special reference to childhood leukemia and natural background radiation. Radiation research. 2010;174(3):387-402. 24. Axelson O, Fredrikson M, Akerblom G, Hardell L. Leukemia in childhood and adolescence and exposure to ionizing radiation in homes built from uranium-containing alum shale concrete. Epidemiology. 2002;13(2):146-50. 25. UKCCS Investigators. The United Kingdom Childhood Cancer Study (UKCCS) of exposure to domestic sources of ionising radiation: I: radon gas. British journal of cancer. 2002;86(11):1721-6. 26. UKCCS Investigators. The United Kingdom Childhood Cancer Study (UKCCS) of exposure to domestic sources of ionising radiation: 2: gamma radiation. British journal of cancer. 2002;86(11):1727-31. 27. Demoury C, Marquant F, Ielsch G, Goujon S, Debayle C, Faure L, et al. Residential Exposure to Natural Background Radiation and Risk of Childhood Acute Leukemia in France, 1990 - 2009. Environmental health perspectives. 2016. 28. Nikkila A, Erme S, Arvela H, Holmgren O, Raitanen J, Lohi O, et al. Background radiation and childhood leukemia: A nationwide register-based case-control study. International journal of cancer Journal international du cancer. 2016;139(9):1975-82. 29. Spycher BD, Lupatsch JE, Zwahlen M, Roosli M, Niggli F, Grotzer MA, et al. Background ionizing radiation and the risk of childhood cancer: a census-based nationwide cohort study. Environmental health perspectives. 2015;123(6):622-8. 30. Kendall GM, Little MP, Wakeford R, Bunch KJ, Miles JC, Vincent TJ, et al. A record-based casecontrol study of natural background radiation and the incidence of childhood leukaemia and other cancers in Great Britain during 1980-2006. Leukemia. 2013;27(1):3-9. 31. Spix C, Grosche B, Bleher M, Kaatsch P, Scholz-Kreisel P, Blettner M. Background gamma radiation and childhood cancer in Germany: an ecological study. Radiation and environmental biophysics. 2017;56(2):127-38. 32. Hauri D, Spycher B, Huss A, Zimmermann F, Grotzer M, von der Weid N, et al. Domestic radon exposure and risk of childhood cancer: a prospective census-based cohort study. Environmental health perspectives. 2013;121(10):1239-44. 33. Raaschou-Nielsen O, Andersen CE, Andersen HP, Gravesen P, Lind M, Schuz J, et al. Domestic radon and childhood cancer in Denmark. Epidemiology. 2008;19(4):536-43. 34. Gjerstorff ML. The Danish Cancer Registry. Scand J Public Health. 2011;39(7 Suppl):42-5. 35. Lacour B, Guyot-Goubin A, Guissou S, Bellec S, Desandes E, Clavel J. Incidence of childhood cancer in France: National Children Cancer Registries, 2000-2004. Eur J Cancer Prev. 2010;19(3):17381. 36. Stiller CA. Childhood cancer in Britain: incidence, survival, mortality. Oxford: Oxford University Press; 2007. 37. Schindler M, Mitter V, Rueegg CS, Bergstrasser E, Gumy-Pause F, Michel G, et al., editors. Death certificate notations in the Swiss Childhood Cancer Registry: Validation of registration procedures and completness. ENCR Scientific Meeting and General Assembly, Towards a harmonised cancer information system in Europe; 2014 12-14 Nov; Ispra, Italy. Ispra, Italy: European Network of Cancer Registries (ENCR); 2014. Page 22 of 35AUTHOR SUBMITTED MANUSCRIPT - JRP-101695.R1 Journal of Radiological Protection 23 38. Leinonen MK, Miettinen J, Heikkinen S, Pitkaniemi J, Malila N. Quality measures of the population-based Finnish Cancer Registry indicate sound data quality for solid malignant tumours. European journal of cancer. 2017;77:31-9. 39. Andersen CE, Raaschou-Nielsen O, Andersen HP, Lind M, Gravesen P, Thomsen BL, et al. Prediction of 222Rn in Danish dwellings using geology and house construction information from central databases. Radiation protection dosimetry. 2007;123(1):83-94. 40. Rybach L, Bachler D, Bucher B, Schwarz G. Radiation doses of Swiss population from external sources. J Environ Radioact. 2002;62(3):277-86. 41. Warnery E, Ielsch G, Lajaunie C, Cale E, Wackernagel H, Debayle C, et al. Indoor terrestrial gamma dose rate mapping in France: a case study using two different geostatistical models. Journal of Environmental Radioactivity. 2015;139:140-8. 42. IRSN IdRedSN-. Analyse spatiale et cartographie du radon sur le territoire métropolitain par l’utilisation de méthodes géostatistiques. Paris; 2012. 43. Kendall GM, Wakeford R, Bunch KJ, Vincent TJ, Little MP. Residential mobility and associated factors in relation to the assessment of exposure to naturally occurring radiation in studies of childhood cancer. Journal of radiological protection : official journal of the Society for Radiological Protection. 2015;35(4):835-68. 44. Kheifets L, Swanson J, Yuan Y, Kusters C, Vergara X. Comparative analyses of studies of childhood leukemia and magnetic fields, radon and gamma radiation. Journal of radiological protection : official journal of the Society for Radiological Protection. 2017;37(2):459-91. 45. Committee to Assess Health Risks from Exposure to Low Levels of Ionizing Radiation. Health risks from exposure to low levels of ionizing radiation, BEIR VII, Phase 2. Washington, D.C.: Board on Radiation Effects - Research Division on Earth and Life Studies - National Research Council of the National Academies; 2006. 46. Stram DO, Kopecky KJ. Power and uncertainty analysis of epidemiological studies of radiation-related disease risk in which dose estimates are based on a complex dosimetry system: some observations. Radiat Res. 2003;160(4):408-17. 47. Wiemels J. Perspectives on the causes of childhood leukemia. Chemico-biological interactions. 2012;196(3):59-67. 48. Kendall GM, Wakeford R, Athanson M, Vincent TJ, Carter EJ, McColl NP, et al. Levels of naturally occurring gamma radiation measured in British homes and their prediction in particular residences. Radiation and environmental biophysics. 2016;55(1):103-24. 49. Kendall GM, Bunch KJ, Miles JC, Vincent TJ, Little MP, Wakeford R, et al. Report of a recordbased case-control study of natural background radiation and incidence of childhood cancer in Great Britain. Documents of the Health Protection Agency HPA-CRCE-045 ed: Health Protection Agency; 2013. 50. Kendall GM, Fell TP, Harrison JD. Dose to red bone marrow of infants, children and adults from radiation of natural origin. Journal of radiological protection : official journal of the Society for Radiological Protection. 2009;29(2):123-38. 51. Harley NH, Robbins ES. Radon and Leukemia in the Danish Study: Another Source of Dose. Health physics. 2009;97(4):343-7. 52. Chernyavskiy P, Kendall GM, Wakeford R, Little MP. Spatial prediction of naturally occurring gamma radiation in Great Britain. J Environ Radioact. 2016;164:300-11. 53. Kendall GM, Chernyavskiy P, Appleton JD, Miles JCH, Wakeford R, Athanson M, et al. Modelling the bimodal distribution of indoor gamma-ray dose-rates in Great Britain. Radiat Environ Biophys. 2018;57(4):321-47. 54. Marquant F, Demoury C, Ielsch G, Laurier D, Hemon D, Clavel J. Response to comment on "Indoor terrestrial gamma dose rate mapping in France: A case study using two different geostatistical models" by Warnery et al. J Environ Radioact. 2018;182:174-6. 55. Adachi N, Adamovitch V, Adjovi Y, Aida K, Akamatsu H, Akiyama S, et al. Measurement and comparison of individual external doses of high-school students living in Japan, France, Poland and Belarus—the ‘D-shuttle’ project—. Journal of Radiological Protection. 2015;36(1):49-66. Page 23 of 35 AUTHOR SUBMITTED MANUSCRIPT - JRP-101695.R1 Journal of Radiological Protection 24 56. Kendall GM, Hughes JS, Oatway WB, Jones AL. Variations in radiation exposures of adults and children in the UK. Journal of radiological protection : official journal of the Society for Radiological Protection. 2006;26(3):257-76. 57. Berrington de Gonzalez A, Iulian Apostoaei A, Veiga LH, Rajaraman P, Thomas BA, Owen Hoffman F, et al. RadRAT: a radiation risk assessment tool for lifetime cancer risk projection. Journal of radiological protection : official journal of the Society for Radiological Protection. 2012;32(3):20522. 58. Nikkila A, Kendall G, Raitanen J, Spycher B, Lohi O, Auvinen A. Effects of incomplete residential histories on studies of environmental exposure with application to childhood leukaemia and background radiation. Environ Res. 2018;166:466-72. 59. Kendall GM, Miles JC, Rees D, Wakeford R, Bunch KJ, Vincent TJ, et al. Variation with socioeconomic status of indoor radon levels in Great Britain: The less affluent have less radon. J Environ Radioact. 2016;164:84-90. 60. Filippini T, Hatch EE, Rothman KJ, Heck JE, Park AS, Crippa A, et al. Association between Outdoor Air Pollution and Childhood Leukemia: A Systematic Review and Dose–Response MetaAnalysis. Environmental health perspectives. 2019;127(4):046002. 61. Van Maele-Fabry G, Gamet-Payrastre L, Lison D. Household exposure to pesticides and risk of leukemia in children and adolescents: Updated systematic review and meta-analysis. International journal of hygiene and environmental health. 2019;222(1):49-67. 62. Greaves M. A causal mechanism for childhood acute lymphoblastic leukaemia. Nature Reviews Cancer. 2018;18(8):471-84. 63. Martin-Lorenzo A, Hauer J, Vicente-Duenas C, Auer F, Gonzalez-Herrero I, Garcia-Ramirez I, et al. Infection Exposure Is a Causal Factor in B-cell Precursor Acute Lymphoblastic Leukemia as a Result of Pax5-Inherited Susceptibility. Cancer discovery. 2015;5(12):1328-43. 64. Rudant J, Lightfoot T, Urayama KY, Petridou E, Dockerty JD, Magnani C, et al. Childhood acute lymphoblastic leukemia and indicators of early immune stimulation: a childhood leukemia international consortium study. American journal of epidemiology. 2015;181(8):549-62. 65. Kinlen LJ. An examination, with a meta-analysis, of studies of childhood leukaemia in relation to population mixing. British journal of cancer. 2012;107(7):1163-8. 66. Urayama KY, Ma X, Selvin S, Metayer C, Chokkalingam AP, Wiemels JL, et al. Early life exposure to infections and risk of childhood acute lymphoblastic leukemia. International journal of cancer Journal international du cancer. 2011;128(7):1632-43. 67. Adam M, Rebholz CE, Egger M, Zwahlen M, Kuehni CE. Childhood leukaemia and socioeconomic status: what is the evidence? Radiation protection dosimetry. 2008;132(2):246-54. 68. Kroll ME, Stiller CA, Murphy MF, Carpenter LM. Childhood leukaemia and socioeconomic status in England and Wales 1976-2005: evidence of higher incidence in relatively affluent communities persists over time. British journal of cancer. 2011;105(11):1783-7. 69. Marquant F, Goujon S, Faure L, Guissou S, Orsi L, Hemon D, et al. Risk of Childhood Cancer and Socio-economic Disparities: Results of the French Nationwide Study Geocap 2002-2010. Paediatric and perinatal epidemiology. 2016;30(6):612-22. 70. Caughey RW, Michels KB. Birth weight and childhood leukemia: a meta-analysis and review of the current evidence. International journal of cancer Journal international du cancer. 2009;124(11):2658-70. 71. Jonathan W, Hilary L. Spatial Aggregation and the Ecological Fallacy. Handbook of Spatial Statistics: CRC Press; 2010. Page 24 of 35AUTHOR SUBMITTED MANUSCRIPT - JRP-101695.R1 IOP Publishing Journal of Radiological Protection 31 Table 2. Relative risks in recent nationwide record-based epidemiological studies on background radiation and childhood leukaemia Source of exposure Country Cases Time-place of exposure Relative risks (95% confidence interval) Leukaemia Central Nervous System tumours Radon concentration (Bq/m3) DK 1,153 Full residential history 1.34 (0.97, 1.85)a/ 0.92 (0.69, 1.22)a/ CH 283 Census 0.90 (0.68,1.19)b/ 1.19 (0.91, 1.57) b/ Radon radiation dose (mSv) GB 9,058 Birth 1.03 (0.96, 1.11) FR 9,056 Diagnosis 1.00 (0.97, 1.02) - Gamma radiation dose (mSv) GB 9,058 Birth 1.12 (1.03, 1.22) FI 1,093 Full residential history 0.97 (0.89, 1.06) - Page 31 of 35 AUTHOR SUBMITTED MANUSCRIPT - JRP-101695.R1 47 48 49 50 51 52 53 54 55 56 57 58 59 60 Journal of Radiological Protection 32 FR 9,056 Diagnosis 1.00 (0.99, 1.01) - CH 530 Census 1.04 (1.00, 1.08)c/ 1.04 (1.00, 1.08)c/ DE 11,447 Diagnosis 1.04 (0.91, 1.20)d/ 1.35 (1.17, 1.57)d/ Radon and background gamma dose combined (mSv) GB 9,058 Birth 1.07 (1.01, 1.13) - FR 9,056 Diagnosis 1.00 (0.99, 1.01) - Note: data are relative risk (95% confidence intervals) per mSv cumulative equivalent dose to the RBM (if not otherwise indicated). Abbreviations: FI Finland, GB Great Britain, FR France, CH Switzerland, DK Denmark and DE Germany. a Per 103 Bq/m3-years b Per 100 Bq/m3 c Per mSv cumulative effective dose (whole body) d RR comparing 1.5 vs 0.5 mSv/a for acute lymphoid leukaemia Page 32 of 35AUTHOR SUBMITTED MANUSCRIPT - JRP-101695.R1 47 48 49 50 51 52 53 54 55 56 57 58 59 60 Journal of Radiological Protection 33 Table 3: Characteristics of models use to assess exposure on gamma radiation in recent record-based epidemiological studies on background radiation and childhood cancer Great Britain Switzerland Finland France Germany Study Kendall et al. 2013 Spycher et al. 2015 Nikkilä et al. 2016 Demoury et al. 2017 Spix et al. 2017 Exposure assessment Indoor dose rates from cosmic and terrestrial sources Outdoor dose rates from cosmic and terrestrial sources Indoor and outdoor dose rates from terrestrial sources Indoor dose rates from cosmic and terrestrial sources Outdoor annual ambient dose rate from terrestrial and cosmic sources Sources 2,283 domestic measurements in Great Britain Airborne spectrometry, 166 in-situ spectrometry measurements, 837 in situ dose rate measurements, and 612 laboratory measurements of rock and soil 346 domestic measurements, a mobile survey with Geiger-counters and spectrometers, Municipal averages of dose rates, Map of Cs-137 activity after Chernobyl nuclear accident, -Terrestrial gamma radiation: 14,124 measurements (8,895 indoor, 5,229 outdoor) and 14,234 TLD measurements in surveillance data. -Telluric gamma radiation: Map of geogenic uranium potential and 97,595 TLD measurements in dentist 1,800 stations in Germany Page 33 of 35 AUTHOR SUBMITTED MANUSCRIPT - JRP-101695.R1 47 48 49 50 51 52 53 54 55 56 57 58 59 60 Journal of Radiological Protection 34 Building material information as an indoor/outdoor factor surgeries and veterinary clinics -Ecological study: average municipality exposure Type of model County districts mean Interpolation using inverse distance weighting Bivariate interpolation Cokriging Interpolation using inverse distance weighting Geographic resolution County District level 2×2 km2 grid map 8 x 8 km grid map 1 × 1 km2 grid map Community level Page 34 of 35AUTHOR SUBMITTED MANUSCRIPT - JRP-101695.R1 47 48 49 50 51 52 53 54 55 56 57 58 59 60 Journal of Radiological Protection 35 Table 4. Characteristics of radon exposure assessment in recent record-based epidemiological studies on background radiation and childhood cancer Denmark Great Britain Switzerland France Study Raaschou-Nielsen et al. 2008 Kendall et al. 2013 Hauri et al. 2013 Demoury et al. 2017 Exposure assessment Domestic radon concentration Domestic radon concentration Domestic radon concentration Domestic radon concentration Sources 3,116 indoor measurements ~400,000 indoor measurements, 35,706 indoor measurements 10,843 measurement of indoor radon Predictors Geographical region, soil type and house characteristics Bedrock and superficial geological characteristics Tectonic units, building information, soil texture, urbanization and floor level concentration and a map of geogenic radon potential Performance R² = 0.45 (tested against independent measurements) R² = 0.34 - 0.40 R² = 0.20 R² = 0.32 Method Linear regression model Log-normal modelling based on measurements grouped by grid square and geological boundaries Log-linear regression model Cokriging model Geographic resolution 1 × 1 km2 grid map 1 × 1 km2 grid map Page 35 of 35 AUTHOR SUBMITTED MANUSCRIPT - JRP-101695.R1 47 48 49 50 51 52 53 54 55 56 57 58 59 60