A generalisation of the method of regression calibration and comparison with Bayesian and frequentist model averaging methods.

For many cancer sites low-dose risks are not known and must be extrapolated from those observed in groups exposed at much higher levels of dose. Measurement error can substantially alter the dose-response shape and hence the extrapolated risk. Even in studies with direct measurement of low-dose exposures measurement error could be substantial in relation to the size of the dose estimates and thereby distort population risk estimates. Recently, there has been considerable attention paid to methods of dealing with shared errors, which are common in many datasets, and particularly important in occupational and environmental settings. In this paper we test Bayesian model averaging (BMA) and frequentist model averaging (FMA) methods, the first of these similar to the so-called Bayesian two-dimensional Monte Carlo (2DMC) method, and both fairly recently proposed, against a very newly proposed modification of the regression calibration method, the extended regression calibration (ERC) method, which is particularly suited to studies in which there is a substantial amount of shared error, and in which there may also be curvature in the true dose response. The quasi-2DMC with BMA method performs well when a linear model is assumed, but very poorly when a linear-quadratic model is assumed, with coverage probabilities both for the linear and quadratic dose coefficients that are under 5% when the magnitude of shared Berkson error is large (50%). For the linear model the bias is generally under 10%. However, using a linear-quadratic model it produces substantially biased (by a factor of 10) estimates of both the linear and quadratic coefficients, with the linear coefficient overestimated and the quadratic coefficient underestimated. FMA performs as well as quasi-2DMC with BMA when a linear model is assumed, and generally much better with a linear-quadratic model, although the coverage probability for the quadratic coefficient is uniformly too high. However both linear and quadratic coefficients have pronounced upward bias, particularly when Berkson error is large. By comparison ERC yields coverage probabilities that are too low when shared and unshared Berkson errors are both large (50%), although otherwise it performs well, and coverage is generally better than the quasi-2DMC with BMA or FMA methods, particularly for the linear-quadratic model. The bias of the predicted relative risk at a variety of doses is generally smallest for ERC, and largest for the quasi-2DMC with BMA and FMA methods (apart from unadjusted regression), with standard regression calibration and Monte Carlo maximum likelihood exhibiting bias in predicted relative risk generally somewhat intermediate between ERC and the other two methods. In general ERC performs best in the scenarios presented, and should be the method of choice in situations where there may be substantial shared error, or suspected curvature in the dose response.

Solar ultraviolet radiation exposure, and incidence of childhood acute lymphocytic leukaemia and non-Hodgkin lymphoma in a US population-based dataset.

BACKGROUND: Acute lymphocytic leukaemia (ALL) and non-Hodgkin lymphoma (NHL) are among the commonest types of childhood cancer. Some previous studies suggested that elevated ultraviolet radiation (UVR) exposures increase ALL risk; many more indicate NHL risk is reduced. METHODS: We assessed age<20 ALL/NHL incidence in Surveillance, Epidemiology and End Results data using AVGLO-derived UVR irradiance/cumulative radiant exposure measures, using quasi-likelihood models accounting for underdispersion, adjusted for age, sex, racial/ethnic group and other county-level socioeconomic variables. RESULTS: There were 30,349 cases of ALL and 8062 of NHL, with significant increasing trends of ALL with UVR irradiance (relative risk (RR) = 1.200/mW/cm2 (95% CI 1.060, 1.359, p = 0.0040)), but significant decreasing trends for NHL (RR = 0.646/mW/cm2 (95% CI 0.512, 0.816, p = 0.0002)). There was a borderline-significant increasing trend of ALL with UVR cumulative radiant exposure (RR = 1.444/MJ/cm2 (95% CI 0.949, 2.197, p = 0.0865)), and significant decreasing trends for NHL (RR = 0.284/MJ/cm2 (95% CI 0.166, 0.485, p < 0.0001)). ALL and NHL trend RR is substantially increased among those aged 0-3. All-age trend RRs are most extreme (increasing for ALL, decreasing for NHL) for Hispanics for both UVR measures. CONCLUSIONS: Our more novel finding, of excess UVR-related ALL risk, is consistent with some previous studies, but is not clear-cut, and in need of replication.

Reconstruction of organ doses for patients undergoing computed tomography examinations in Canada 1992-2019.

We derived the first comprehensive organ dose library for Canadian pediatric and adult patients who underwent computed tomography (CT) scans between 1992 and 2019 to support epidemiological analysis of radiation risk. We calculated organ absorbed doses for Canadian CT patients in two steps. First, we modeled Computed Tomography Dose Index (CTDI) values by patient age, scan body part, and scan year for the scan period between 1992 and 2019 using national survey data conducted in Canada and partially the United Kingdom survey data as surrogates. Second, we converted CTDI values to organ absorbed doses using a library of organ dose conversion coefficients built in an organ dose calculation program, the National Cancer Institute dosimetry system for CT. In result, we created a library of doses delivered to 33 organs and tissues by different patient ages and genders, scan body parts and scan years. In the scan period before 2000, the organs receiving the greatest dose in the head, chest and abdomen-pelvis scans were the active marrow (3.7-15.2 mGy), lungs (54.7-62.8 mGy) and colon (54.9-68.5 mGy), respectively. We observed organ doses reduced by 24% (pediatric head and torso scans, and adult head scans) and 55% (adult torso scans) after 2000. The organ dose library will be used to analyse the risk of radiation exposure from CT scans in the Canadian CT patient cohort.

A generalisation of the method of regression calibration and comparison with Bayesian and frequentist model averaging methods.

For many cancer sites low-dose risks are not known and must be extrapolated from those observed in groups exposed at much higher levels of dose. Measurement error can substantially alter the dose-response shape and hence the extrapolated risk. Even in studies with direct measurement of low-dose exposures measurement error could be substantial in relation to the size of the dose estimates and thereby distort population risk estimates. Recently, there has been considerable attention paid to methods of dealing with shared errors, which are common in many datasets, and particularly important in occupational and environmental settings. In this paper we test Bayesian model averaging (BMA) and frequentist model averaging (FMA) methods, the first of these similar to the so-called Bayesian two-dimensional Monte Carlo (2DMC) method, and both fairly recently proposed, against a very newly proposed modification of the regression calibration method, the extended regression calibration (ERC) method, which is particularly suited to studies in which there is a substantial amount of shared error, and in which there may also be curvature in the true dose response. The quasi-2DMC with BMA method performs well when a linear model is assumed, but very poorly when a linear-quadratic model is assumed, with coverage probabilities both for the linear and quadratic dose coefficients that are under 5% when the magnitude of shared Berkson error is large (50%). For the linear model the bias is generally under 10%. However, using a linear-quadratic model it produces substantially biased (by a factor of 10) estimates of both the linear and quadratic coefficients, with the linear coefficient overestimated and the quadratic coefficient underestimated. FMA performs as well as quasi-2DMC with BMA when a linear model is assumed, and generally much better with a linear-quadratic model, although the coverage probability for the quadratic coefficient is uniformly too high. However both linear and quadratic coefficients have pronounced upward bias, particularly when Berkson error is large. By comparison ERC yields coverage probabilities that are too low when shared and unshared Berkson errors are both large (50%), although otherwise it performs well, and coverage is generally better than the quasi-2DMC with BMA or FMA methods, particularly for the linear-quadratic model. The bias of the predicted relative risk at a variety of doses is generally smallest for ERC, and largest for the quasi-2DMC with BMA and FMA methods (apart from unadjusted regression), with standard regression calibration and Monte Carlo maximum likelihood exhibiting bias in predicted relative risk generally somewhat intermediate between ERC and the other two methods. In general ERC performs best in the scenarios presented, and should be the method of choice in situations where there may be substantial shared error, or suspected curvature in the dose response.

Radiotherapy for non-cancer diseases: benefits and long-term risks.

PURPOSE: The discovery of X-rays was followed by a variety of attempts to treat infectious diseases and various other non-cancer diseases with ionizing radiation, in addition to cancer. There has been a recent resurgence of interest in the use of such radiotherapy for non-cancer diseases. Non-cancer diseases for which use of radiotherapy has currently been proposed include refractory ventricular tachycardia, neurodegenerative diseases (e.g. Alzheimer's disease and dementia), and Coronavirus Disease 2019 (COVID-19) pneumonia, all with ongoing clinical studies that deliver radiation doses of 0.5-25 Gy in a single fraction or in multiple daily fractions. In addition to such non-cancer effects, historical indications predominantly used in some countries (e.g. Germany) include osteoarthritis and degenerative diseases of the bones and joints. This narrative review gives an overview of the biological rationale and ongoing preclinical and clinical studies for radiotherapy proposed for various non-cancer diseases, discusses the plausibility of the proposed biological rationale, and considers the long-term radiation risks of cancer and non-cancer diseases. CONCLUSIONS: A growing body of evidence has suggested that radiation represents a double-edged sword, not only for cancer, but also for non-cancer diseases. At present, clinical evidence has shown some beneficial effects of radiotherapy for ventricular tachycardia, but there is little or no such evidence of radiotherapy for other newly proposed non-cancer diseases (e.g. Alzheimer's disease, COVID-19 pneumonia). Patients with ventricular tachycardia and COVID-19 pneumonia have thus far been treated with radiotherapy when they are an urgent life threat with no efficient alternative treatment, but some survivors may encounter a paradoxical situation where patients were rescued by radiotherapy but then get harmed by radiotherapy. Further studies are needed to justify the clinical use of radiotherapy for non-cancer diseases, and optimize dose to diseased tissue while minimizing dose to healthy tissue.

Radiation Dose Does Not Affect the Predictive Value of Thyroid Biopsy for Diagnosing Papillary Thyroid Cancer in a Belarusian Cohort Exposed to Chernobyl Fallout.

INTRODUCTION: The Chernobyl nuclear accident exposed residents of contaminated territories to substantial quantities of radioiodines and was followed by an increase in thyroid cancer, primarily papillary thyroid cancer (PTC), among exposed children and adolescents. Although thyroid biopsy is an essential component of screening programs following accidental exposure to radioiodines, it is unknown whether the predictive value of biopsy is affected by different levels of environmental exposure. METHODS: A cohort of 11,732 Belarusians aged ≤18 years at the time of the Chernobyl accident with individual thyroid radiation dose estimates was screened at least once 11-22 years later. Paired cytologic conclusions and histopathologic diagnoses were possible for 258 thyroid nodules from 238 cohort members. Cytologic conclusions were divided into five reporting categories, with all follicular lesion aspirates combined into a single indeterminate category. Standard performance indicators, risk of malignancy (ROM), and odds ratios for a correct cytologic conclusion were calculated, both overall and according to quintile of thyroid radiation dose. RESULTS: The arithmetic mean thyroid dose estimate for the study group was 1.73 Gy (range: 0.00-23.64 Gy). The final histopathologic diagnosis was cancer for 136 of 258 biopsies (52.7%; 135 papillary and 1 follicular). The overall ROM was 96.7% for cytologies definite for PTC, 83.7% for suspicious for PTC, 33.0% for indeterminate, 8.1% for benign, and 31.0% for non-diagnostic. The ROM showed little change according to level of radiation exposure. Overall, there was no association between thyroid radiation dose and the odds ratio for a correct cytologic conclusion (p = 0.24). When analyzed according to dose quintile, the odds ratio for a correct conclusion increased two-fold at 0.10-0.29 Gy compared to a dose of 0.00-0.09 Gy and decreased at doses of 0.3-24 Gy (p value for linear trend = 0.99). CONCLUSIONS: At radiation doses received by a cohort of young Belarusians exposed to radioiodines by the Chernobyl accident, the predictive value of thyroid biopsy for diagnosing PTC was not significantly affected by level of radiation exposure.

Minimum latency effects for cancer associated with exposures to radiation or other carcinogens.

BACKGROUND: In estimating radiation-associated cancer risks a fixed period for the minimum latency is often assumed. Two empirical latency functions have been used to model latency, continuously increasing from 0. A stochastic biologically-based approach yields a still more plausible way of describing latency and can be directly estimated from clinical data. METHODS: We derived the parameters for a stochastic biologically-based model from tumour growth data for various cancers, and least-squares fitted the two types of empirical latency function to the stochastic model-predicted cumulative probability. RESULTS: There is wide variation in growth rates among tumours, particularly slow for prostate and thyroid cancer and particularly fast for leukaemia. The slow growth rate for prostate and thyroid tumours implies that the number of tumour cells required for clinical detection cannot greatly exceed 106. For all tumours, both empirical latency functions closely approximated the predicted biological model cumulative probability. CONCLUSIONS: Our results, illustrating use of a stochastic biologically-based model using clinical data not tied to any particular carcinogen, have implications for estimating latency associated with any mutagen. They apply to tumour growth in general, and may be useful for example, in planning screenings for cancer using imaging techniques.

Assessment of Cataract Risk after Diagnostic Head CT Scan Radiation Exposure in Ontario, Canada.

Ionizing radiation is one of the known risk factors for cataract development, however, there is still debate regarding the level of risk after low dose exposures. One of the largest sources of radiation exposure to the lens of the eye is diagnostic CT scans. The aim of this study was to examine whether ionizing radiation associated with head CT scans increases cataract risk in residents of Ontario, Canada. Data were collected from January 1, 1994 to December 31, 2015 (22 years) from anonymized Ontario Health Insurance Plan (OHIP) medical records for over 16 million subjects. A lens dose was estimated for each CT scan using the National Cancer Institute dosimetry system for CT (NCICT) program combined with Canada-specific CTDIvol data. Multivariate Cox proportional hazards analysis was performed with cataract extraction surgery as the primary outcome and lens dose as the main variable of interest, with inclusion of various medical and demographic covariates. Lag periods of 3, 5 and 7 years were incorporated. When lens dose was treated as a continuous variable, hazard ratios (per 100 mGy) ranged from 0.82 (0.80-0.84) to 1.10 (1.09-1.11) depending on the lag period. As a secondary analysis, when individuals were binned based on their total cumulative dose, no significant dose response pattern was observed in the low dose region. Overall, within the bounds of this study, the data do not support an increased risk of vision impairing cataracts after diagnostic head CT scan radiation exposure.

A generalisation of the method of regression calibration and comparison with the Bayesian 2-dimensional Monte Carlo method.

For many cancer sites it is necessary to assess risks from low-dose exposures via extrapolation from groups exposed at moderate and high levels of dose. Measurement error can substantially alter the shape of this relationship and hence the derived population risk estimates. Even in studies with direct measurement of low-dose exposures measurement error could be substantial in relation to the size of the dose estimates and thereby distort population risk estimates. Recently, much attention has been devoted to the issue of shared errors, common in many datasets, and particularly important in occupational settings. In this paper we test a Bayesian model averaging method, the so-called Bayesian two-dimensional Monte Carlo (2DMC) method, that has been fairly recently proposed against a very newly proposed modification of the regression calibration method, which is particularly suited to studies in which there is a substantial amount of shared error, and in which there may also be curvature in the true dose response. We also compared both methods against standard regression calibration and Monte Carlo maximum likelihood. The Bayesian 2DMC method performs poorly, with coverage probabilities both for the linear and quadratic dose coefficients that are under 5%, particularly when the magnitudes of classical and Berkson error are both moderate to large (20%-50%). The method also produces substantially biased (by a factor of 10) estimates of both the linear and quadratic coefficients, with the linear coefficient overestimated and the quadratic coefficient underestimated. By comparison the extended regression calibration method yields coverage probabilities that are too low when shared and unshared Berkson errors are both large (50%), although otherwise it performs well, and coverage is generally better than the Bayesian 2DMC and all other methods. The bias of the predicted relative risk at a variety of doses is generally smallest for extended regression calibration, and largest for the Bayesian 2DMC method (apart from unadjusted regression), with standard regression calibration and Monte Carlo maximum likelihood exhibiting bias in predicted relative risk generally somewhat intermediate between the other two methods.

A generalisation of the method of regression calibration.

There is direct evidence of risks at moderate and high levels of radiation dose for highly radiogenic cancers such as leukaemia and thyroid cancer. For many cancer sites, however, it is necessary to assess risks via extrapolation from groups exposed at moderate and high levels of dose, about which there are substantial uncertainties. Crucial to the resolution of this area of uncertainty is the modelling of the dose-response relationship and the importance of both systematic and random dosimetric errors for analyses in the various exposed groups. It is well recognised that measurement error can alter substantially the shape of this relationship and hence the derived population risk estimates. Particular attention has been devoted to the issue of shared errors, common in many datasets, and particularly important in occupational settings. We propose a modification of the regression calibration method which is particularly suited to studies in which there is a substantial amount of shared error, and in which there may also be curvature in the true dose response. This method can be used in settings where there is a mixture of Berkson and classical error. In fits to synthetic datasets in which there is substantial upward curvature in the true dose response, and varying (and sometimes substantial) amounts of classical and Berkson error, we show that the coverage probabilities of all methods for the linear coefficient [Formula: see text] are near the desired level, irrespective of the magnitudes of assumed Berkson and classical error, whether shared or unshared. However, the coverage probabilities for the quadratic coefficient [Formula: see text] are generally too low for the unadjusted and regression calibration methods, particularly for larger magnitudes of the Berkson error, whether this is shared or unshared. In contrast Monte Carlo maximum likelihood yields coverage probabilities for [Formula: see text] that are uniformly too high. The extended regression calibration method yields coverage probabilities that are too low when shared and unshared Berkson errors are both large, although otherwise it performs well, and coverage is generally better than these other three methods. A notable feature is that for all methods apart from extended regression calibration the estimates of the quadratic coefficient [Formula: see text] are substantially upwardly biased.

Radiation exposure and leukaemia risk among cohorts of persons exposed to low and moderate doses of external ionising radiation in childhood.

BACKGROUND: Many high-dose groups demonstrate increased leukaemia risks, with risk greatest following childhood exposure; risks at low/moderate doses are less clear. METHODS: We conducted a pooled analysis of the major radiation-associated leukaemias (acute myeloid leukaemia (AML) with/without the inclusion of myelodysplastic syndrome (MDS), chronic myeloid leukaemia (CML), acute lymphoblastic leukaemia (ALL)) in ten childhood-exposed groups, including Japanese atomic bomb survivors, four therapeutically irradiated and five diagnostically exposed cohorts, a mixture of incidence and mortality data. Relative/absolute risk Poisson regression models were fitted. RESULTS: Of 365 cases/deaths of leukaemias excluding chronic lymphocytic leukaemia, there were 272 AML/CML/ALL among 310,905 persons (7,641,362 person-years), with mean active bone marrow (ABM) dose of 0.11 Gy (range 0-5.95). We estimated significant (P < 0.005) linear excess relative risks/Gy (ERR/Gy) for: AML (n = 140) = 1.48 (95% CI 0.59-2.85), CML (n = 61) = 1.77 (95% CI 0.38-4.50), and ALL (n = 71) = 6.65 (95% CI 2.79-14.83). There is upward curvature in the dose response for ALL and AML over the full dose range, although at lower doses (<0.5 Gy) curvature for ALL is downwards. DISCUSSION: We found increased ERR/Gy for all major types of radiation-associated leukaemia after childhood exposure to ABM doses that were predominantly (for 99%) <1 Gy, and consistent with our prior analysis focusing on <100 mGy.

A generalisation of the method of regression calibration.

There is direct evidence of risks at moderate and high levels of radiation dose for highly radiogenic cancers such as leukaemia and thyroid cancer. For many cancer sites, however, it is necessary to assess risks via extrapolation from groups exposed at moderate and high levels of dose, about which there are substantial uncertainties. Crucial to the resolution of this area of uncertainty is the modelling of the dose-response relationship and the importance of both systematic and random dosimetric errors for analyses in the various exposed groups. It is well recognised that measurement error can alter substantially the shape of this relationship and hence the derived population risk estimates. Particular attention has been devoted to the issue of shared errors, common in many datasets, and particularly important in occupational settings. We propose a modification of the regression calibration method which is particularly suited to studies in which there is a substantial amount of shared error, and in which there may also be curvature in the true dose response. This method can be used in settings where there is a mixture of Berkson and classical error. In fits to synthetic datasets in which there is substantial upward curvature in the true dose response, and varying (and sometimes substantial) amounts of classical and Berkson error, we show that the coverage probabilities of all methods for the linear coefficient \(\alpha\) are near the desired level, irrespective of the magnitudes of assumed Berkson and classical error, whether shared or unshared. However, the coverage probabilities for the quadratic coefficient \(\beta\) are generally too low for the unadjusted and regression calibration methods, particularly for larger magnitudes of the Berkson error, whether this is shared or unshared. In contrast Monte Carlo maximum likelihood yields coverage probabilities for \(\beta\) that are uniformly too high. The extended regression calibration method yields coverage probabilities that are too low when shared and unshared Berkson errors are both large, although otherwise it performs well, and coverage is generally better than these other three methods. A notable feature is that for all methods apart from extended regression calibration the estimates of the quadratic coefficient \(\beta\) are substantially upwardly biased.

Lympho-hematopoietic malignancies risk after exposure to low dose ionizing radiation during cardiac catheterization in childhood.

Pediatric patients with congenital heart disease (CHD) often undergo low dose ionizing radiation (LDIR) from cardiac catheterization (CC) for the diagnosis and/or treatment of their disease. Although radiation doses from a single CC are usually low, less is known about the long-term radiation associated cancer risks. We aimed to assess the risk of lympho-hematopoietic malignancies in pediatric CHD patients diagnosed or treated with CC. A French cohort of 17,104 children free of cancer who had undergone a first CC from 01/01/2000 to 31/12/2013, before the age of 16 was set up. The follow-up started at the date of the first recorded CC until the exit date, i.e., the date of death, the date of first cancer diagnosis, the date of the 18th birthday, or the 31/12/2015, whichever occurred first. Poisson regression was used to estimate the LDIR associated cancer risk. The median follow-up was 5.9 years, with 110,335 person-years. There were 22,227 CC procedures, yielding an individual active bone marrow (ABM) mean cumulative dose of 3.0 milligray (mGy). Thirty-eight incident lympho-hematopoietic malignancies were observed. When adjusting for attained age, gender and predisposing factors to cancer status, no increased risk was observed for lympho-hematopoietic malignancies RR/mGy = 1.00 (95% CI: 0.88; 1.10). In summary, the risk of lympho-hematopoietic malignancies and lymphoma was not associated to LDIR in pediatric patients with CHD who undergo CC. Further epidemiological studies with greater statistical power are needed to improve the assessment of the dose-risk relationship.

Ionising radiation and cardiovascular disease: systematic review and meta-analysis.

OBJECTIVE: To systematically review and perform a meta-analysis of radiation associated risks of cardiovascular disease in all groups exposed to radiation with individual radiation dose estimates. DESIGN: Systematic review and meta-analysis. MAIN OUTCOME MEASURES: Excess relative risk per unit dose (Gy), estimated by restricted maximum likelihood methods. DATA SOURCES: PubMed and Medline, Embase, Scopus, Web of Science Core collection databases. ELIGIBILITY CRITERIA FOR SELECTING STUDIES: Databases were searched on 6 October 2022, with no limits on date of publication or language. Animal studies and studies without an abstract were excluded. RESULTS: The meta-analysis yielded 93 relevant studies. Relative risk per Gy increased for all cardiovascular disease (excess relative risk per Gy of 0.11 (95% confidence interval 0.08 to 0.14)) and for the four major subtypes of cardiovascular disease (ischaemic heart disease, other heart disease, cerebrovascular disease, all other cardiovascular disease). However, interstudy heterogeneity was noted (P<0.05 for all endpoints except for other heart disease), possibly resulting from interstudy variation in unmeasured confounders or effect modifiers, which is markedly reduced if attention is restricted to higher quality studies or those at moderate doses (<0.5 Gy) or low dose rates (<5 mGy/h). For ischaemic heart disease and all cardiovascular disease, risks were larger per unit dose for lower dose (inverse dose effect) and for fractionated exposures (inverse dose fractionation effect). Population based excess absolute risks are estimated for a number of national populations (Canada, England and Wales, France, Germany, Japan, USA) and range from 2.33% per Gy (95% confidence interval 1.69% to 2.98%) for England and Wales to 3.66% per Gy (2.65% to 4.68%) for Germany, largely reflecting the underlying rates of cardiovascular disease mortality in these populations. Estimated risk of mortality from cardiovascular disease are generally dominated by cerebrovascular disease (around 0.94-1.26% per Gy), with the next largest contribution from ischaemic heart disease (around 0.30-1.20% per Gy). CONCLUSIONS: Results provide evidence supporting a causal association between radiation exposure and cardiovascular disease at high dose, and to a lesser extent at low dose, with some indications of differences in risk between acute and chronic exposures, which require further investigation. The observed heterogeneity complicates a causal interpretation of these findings, although this heterogeneity is much reduced if only higher quality studies or those at moderate doses or low dose rates are considered. Studies are needed to assess in more detail modifications of radiation effect by lifestyle and medical risk factors. SYSTEMATIC REVIEW REGISTRATION: PROSPERO CRD42020202036.

Exposure to the Thyroid from Intake of Radioiodine Isotopes after the Chornobyl Accident. Report I: Revised Doses and Associated Uncertainties for the Ukrainian-American Cohort.

Thyroid doses from intake of radioiodine isotopes (131I, 132Te+132I, and 133I) and associated uncertainties were revised for the 13,204 Ukrainian-American cohort members exposed in childhood and adolescence to fallout from the Chornobyl nuclear power plant accident. The main changes related to the revision of the 131I thyroid activity measured in cohort members, the use of thyroid-mass values specific to the Ukrainian population, and the revision of the 131I ground deposition densities in Ukraine. Uncertainties in doses were assessed considering shared and unshared errors in the parameters of the dosimetry model. Using a Monte-Carlo simulation procedure, 1,000 individual stochastic thyroid doses were calculated for each cohort member. The arithmetic mean of thyroid doses from intake of 131I, 132Te+132I, and 133I for the entire cohort was 0.60 Gy (median = 0.22 Gy). For 9,474 subjects (71.6% of the total), the thyroid doses were less than 0.5 Gy. Thyroid doses for 42 cohort members (0.3% of the total) exceeded 10 Gy while the highest dose was 35 Gy. Intake of 131I contributed around 95% to internal thyroid exposure from radioiodine isotopes. The geometric standard deviation of individual stochastic thyroid doses varied among cohort members from 1.4 to 4.3 with an arithmetic mean of 1.6 and a median of 1.4. It was shown that the contribution of shared errors to the dose uncertainty was small. The revised thyroid doses resulted, in average, in around 40% decrease for cohort members from Zhytomyr Oblast and an increase of around 24% and 35% for the cohort members from Kyiv and Chernihiv Oblast, respectively. Arithmetic mean of TD20 doses for the cohort was around 8% less than that estimated in TD10, 0.60 Gy vs. 0.65 Gy, respectively; however, global median of TD20 doses somewhat increased compared to TD10: 0.22 Gy vs. 0.19 Gy, respectively. The difference between TD10 and TD20 was mainly due to a revision of the individual 131I thyroid activity measured in the cohort members.

Low-dose radiotherapy for COVID-19 pneumonia and cancer: summary of a recent symposium and future perspectives.

The lessons learned from the Coronavirus Disease 2019 (COVID-19) pandemic are numerous. Low dose radiotherapy (LDRT) was used in the pre-antibiotic era as treatment for bacterially/virally associated pneumonia. Motivated in part by these historic clinical and radiobiological data, LDRT for treatment of COVID-19-associated pneumonia was proposed in early 2020. Although there is a large body of epidemiological and experimental data pointing to effects such as cancer at low doses, there is some evidence of beneficial health effects at low doses. It has been hypothesized that low dose radiation could be combined with immune checkpoint therapy to treat cancer. We shall review here some of these old radiobiological and epidemiological data, as well as the newer data on low dose radiation and stimulated immune response and other relevant emerging data. The paper includes a summary of several oral presentations given in a Symposium on "Low dose RT for COVID and other inflammatory diseases" as part of the 67th Annual Meeting of the Radiation Research Society, held virtually 3-6 October 2021.

Breast cancer risk in Ukrainian women exposed to Chornobyl fallout while pregnant or lactating: standardized incidence ratio analysis, 1998 to 2016.

The radiation-related risk of breast cancer among women following the Chornobyl accident remains uncertain. During pregnancy, there is rapid cell proliferation in the breast while radioactive iodine from fallout exposure can concentrate in lactating breast tissues. We conducted a standardized incidence ratio (SIR) analysis of breast cancer in a cohort of 2,631 women who were lactating and/or pregnant at any time during the 2-month period of radioiodine fallout (April 26, 1986-June 30, 1986). There were 37,151 person-years of follow-up, and 26 incident breast cancers were identified through linkage with the National Cancer Registry of Ukraine. Breast cancer rates among pregnant or lactating women were compared to the general population rates, and SIRs were adjusted for oblast, urban/rural, age, and calendar year. The SIR was not significant for women pregnant at the time of the accident (SIR = 0.75; 95% CI 0.44, 1.18) or for women lactating anytime within 2 months of the accident (SIR = 0.96; 95% CI 0.48, 1.68). However, there was a non-significantly elevated risk for women lactating at the time of accident (SIR = 1.30, 95% CI 0.40, 3.01). The increased SIR for breast cancer among lactating women is consistent with the results of a similar study in Belarus and indicates the need to quantify the radiation risk of breast cancer in a larger study of women lactating during the period of fallout exposure.

Alternative stopping rules to limit tree expansion for random forest models.

Random forests are a popular type of machine learning model, which are relatively robust to overfitting, unlike some other machine learning models, and adequately capture non-linear relationships between an outcome of interest and multiple independent variables. There are relatively few adjustable hyperparameters in the standard random forest models, among them the minimum size of the terminal nodes on each tree. The usual stopping rule, as proposed by Breiman, stops tree expansion by limiting the size of the parent nodes, so that a node cannot be split if it has less than a specified number of observations. Recently an alternative stopping criterion has been proposed, stopping tree expansion so that all terminal nodes have at least a minimum number of observations. The present paper proposes three generalisations of this idea, limiting the growth in regression random forests, based on the variance, range, or inter-centile range. The new approaches are applied to diabetes data obtained from the National Health and Nutrition Examination Survey and four other datasets (Tasmanian Abalone data, Boston Housing crime rate data, Los Angeles ozone concentration data, MIT servo data). Empirical analysis presented herein demonstrate that the new stopping rules yield competitive mean square prediction error to standard random forest models. In general, use of the intercentile range statistic to control tree expansion yields much less variation in mean square prediction error, and mean square prediction error is also closer to the optimal. The Fortran code developed is provided in the Supplementary Material.

Low-Dose Extrapolation Factors Implied by Mortality and Incidence Data from the Japanese Atomic Bomb Survivor Life Span Study Data.

Assessment of the effect of low dose and low-dose-rate exposure depends critically on extrapolation from groups exposed at high dose and high-dose rates such as the Japanese atomic bomb survivor data, and has often been achieved via application of a dose and dose-rate effectiveness factor (DDREF). An important component of DDREF is the factor determining the effect of extrapolation of dose, the so-called low-dose extrapolation factor (LDEF). To assess LDEF models linear (or linear quadratic) in dose are often fitted. In this report LDEF is assessed via fitting relative rate models that are linear or linear quadratic in dose to the latest Japanese atomic bomb survivor data on solid cancer, leukemia and circulatory disease mortality (followed from 1950 through 2003) and to data on solid cancer, lung cancer and urinary tract cancer incidence. The uncertainties in LDEF are assessed using parametric bootstrap techniques. Analysis is restricted to survivors with <3 Gy dose. There is modest evidence for upward curvature in dose response in the mortality data. For leukemia and for all solid cancer excluding lung, stomach and breast cancer there is significant curvature (P < 0.05). There is no evidence of curvature for circulatory disease (P > 0.5). The estimate of LDEF for all solid cancer mortality is 1.273 [95% confidence intervals (CI) 0.913, 2.182], for all solid cancer mortality excluding lung cancer, stomach cancer and breast cancer is 2.183 (95% CI 1.090, >100) and for leukemia mortality is 11.447 (95% CI 2.390, >100). For stomach cancer mortality LDEF is modestly raised, 1.077 (95% CI 0.526, >100), while for lung cancer, female breast cancer and circulatory disease mortality the LDEF does not much exceed 1. LDEF for solid cancer incidence is 1.186 (95% CI 0.942, 1.626) and for urinary tract cancer is 1.298 (95% CI <0, 7.723), although for lung cancer LDEF is not elevated, 0.842 (95% CI 0.344, >100).