Skip to main content Skip to navigation
U.S. Transuranium and Uranium Registries Conference Contributions

Health Physics Society Meeting, Phoenix, AZ, July 25-29, 2021

The 2021 Health Physics Society meeting was held in a hybrid format, allowing attendees to participate in-person or virtually. USTUR faculty members gave four presentations. They were also co-authors on two presentations that were given by the USTUR’s DOE program manager and one given by a collaborator at Los Alamos National Laboratory (LANL).

Dr. Šefl presents his follow-up study of a Manhattan Project worker

USTUR Whole-body Case 0680: 53-year follow-up of a Manhattan Project worker

Martin Šefl (USTUR), Maia Avtandilashvili (USTUR), Sergei Y. Tolmachev (USTUR)

This whole-body tissue donor to the United States Transuranium and Uranium Registries (USTUR) was occupationally exposed to a mixture of plutonium compounds via chronic inhalation. This individual was one of 26 Manhattan project workers, informally known as ‘UPPU (You Pee Pu) Club’. He died 53 years post-exposure. At the time of death, 1,765 Bq of 239Pu was retained in the body, of which 39.7% was in the skeleton, 37.5% in the liver, 16.0% in the respiratory tract, and 6.8% in the remaining soft tissues. Nineteen urine, one fecal, and one blood analysis results as well as four in vivo chest measurements were available. The organ activities at the time of death and bioassay data were used to estimate the intake and radiation doses using the Taurus internal dosimetry software. ICRP recommended biokinetic models adequately described the individual’s long-term plutonium retention and excretion. The total cumulative 239Pu intake of 31,716 Bq was estimated; of which, 24,853 Bq (78.4%) were contributed by inhalation of plutonium nitrate and 6,863 Bq (21.6%) of plutonium dioxide. The committed equivalent doses to the red bone marrow, bone surface, liver, lungs, and brain were 0.71 Sv, 6.5 Sv, 8.3 Sv, 3.8 Sv, and 0.068 Sv, respectively. The committed effective dose was 1.22 Sv. [USTUR-0576-21A]

Presentation Slides (PPTX with animation)
Presentation Slides (PDF)

Analysis of long-term retention of plutonium in the respiratory tract tissues of four workers: Bound fraction vs. scar-tissue compartments

Deepesh Poudel (LANL), Maia Avtandilashvili (USTUR), John A. Klumpp (LANL), Luiz Bertelli (LANL), Sergei Y. Tolmachev (USTUR)

Respiratory tract tissues collected from four former nuclear workers involved in various inhalation incidents were analyzed post mortem for plutonium by the United States Transuranium and Uranium Registries. Activities in the upper respiratory tract of these individuals were found to be higher than those predicted using the most recent biokinetic models described in publications of the International Commission on Radiological Protection. An assumption of ‘bound fraction’ of 0.4-4% was able to explain the data from three workers who had inhaled soluble to fairly insoluble forms of plutonium. For the fourth worker who had inhaled high-fired plutonium oxide, a more insoluble form of plutonium, a mechanism other than bound fraction was required to explain the observed retention of plutonium in the respiratory tract tissues. Literature review points to the presence of – and a significant retention of – plutonium activity in the scar tissues of the lungs. This presentation proposes a human respiratory tract model modified with the addition of scar tissue compartments to describe the long-term retention of plutonium in the respiratory tract of these individuals. The transfer rates between the compartments were determined using Markov Chain Monte Carlo analysis of the urinary excretion data, lung counts, and post-mortem measurements of the systemic and respiratory tract compartments, as available. The estimates obtained from modeling these data showed that as much as one-third of the total activity in the lung can be sequestered in scar tissues. [USTUR-0579-21A]

Presentation Slides

Comparison of two methods to estimate skeletal plutonium concentration from limited sets of bones

George Tabatadze (USTUR), Maia Avtandilashvili (USTUR), Sergei Y. Tolmachev (USTUR)

Historically, two calculation methods have been used by the United States Transuranium and Uranium Registries (USTUR) to estimate the actinide skeletal concentration: (i) arithmetic average and (ii) mass-weighted average of concentrations measured in bone samples. Preliminary comparison of skeletal concentrations, estimated for 216 partial-body USTUR cases, using these two methods indicates a statistically significant difference (p <0.05) with the bias of 15% between the estimates. The aim of this research is to determine: (1) which method of skeletal concentration estimate is more accurate for a given (collected) set of bones and (2) among the sets of bones most commonly collected for partial-body donations, which set provides more accurate estimate of the total skeletal concentration using each method. Nineteen whole-body cases with complete skeleton analyses were used to compare the estimates of the skeletal concentration based on different sets of bones with the concentration based on all measured bones from the right side of the skeleton. Out of 19 cases, 239Pu was a primary radionuclide of exposure for 17, and 238Pu for two cases. Five individuals were diagnosed with osteoporosis. Since osteoporosis significantly impacts plutonium distribution in the skeleton, 19 cases were divided in two study groups – osteoporotic (5) and non-osteoporotic (14). These cases were further sub-divided into 11 bone groups, based on a number of bones (2 to 8) and their frequency of collection at autopsies. These groups represent different balance of cortical- and trabecular-bone-rich bone samples. To compare the two methods, for each bone group, the arithmetic (Ca) and weighted (Cw) average plutonium concentrations were calculated and compared to the total skeleton concentrations (Csk). Preliminary results indicated that, for all cases and all bone groups, both Ca and Cw predict Csk within 10% of the best estimate and Cw yields slightly better estimate of the Csk for non-osteoporotic cases; however, it has a higher uncertainty. [USTUR-0577-21A]

Presentation Slides

Latent bone modeling approach to estimate plutonium activity concentration in human skeleton

Joey Y. Zhou (DOE), Maia Avtandilashvili (USTUR), Sergei Y. Tolmachev (USTUR)

The skeleton is a major depository site for plutonium in a human body. In radiation protection, a long-term standing question is: What is the most accurate and precise way to estimate the skeleton plutonium concentration and activity from the analysis of a limited set of bones? To answer this question, a multiple linear regression was used in several studies. The key limitation of this approach is multicollinearity among independent variables since the activity concentrations from individual bones are highly correlated resulting in unstable and imprecise estimates of model coefficients. In addition, the number of individual bones allowed in a multiple linear regression model is limited, given a very small number of studied cases. Skeleton plutonium activity concentrations (Bq kg-1 of wet bone) for 19 whole-body tissue donors to the United States Transuranium and Uranium Registries (USTUR), were estimated based on post-mortem radiochemical analyses of the right side of the skeleton, where the total number of analyzed bones ranged from 72 to 89. At the USTUR, 87% of deceased Registrants are partial-body tissue donors with only 2 to 8 bones collected at autopsy. For these cases, the most commonly collected bones are rib, sternum, vertebral body, patella, clavicle, and femur middle shaft. This study applied principal components regression (PCR) by performing principal components analysis (PCA) on an analytical data set from 19 whole-body cases, followed by the selection of a set of 1 to 3 principal components as latent bones (independent variables) for a subsequent multiple linear regression modeling. Latent bone concentration (Clb) is not directly measured but is a linear combination of individually measured bone concentrations (Cbone). Latent bone concentrations, as independent variables in multiple linear regression, are uncorrelated with each other. For rib, sternum, and vertebral body, PCR analysis resulted in the first latent bone equation: Clb1 = 0.5759×Crib + 0.5755×Csternum + 0.5807×Cvert. In this case, the first latent bone alone explained 98.4% of total variance, and the skeleton plutonium concentration can be calculated as Cskel = (18.0±0.8)×Clb1 + 25.0±1.4. [USTUR-0580-21A]

Presentation Slides
Watch Video (YouTube)

Effect of osteoporosis on latent bone models to estimate plutonium activity concentration in human skeleton

Sergei Y. Tolmachev (USTUR), Maia Avtandilashvili (USTUR), Joey Y. Zhou (DOE)

The recently developed latent bone modeling (LBM) approach applies principal components regression (PCR) to estimate plutonium activity concentration in the human skeleton from measurements of a limited set of bone samples. The analytical bone dataset contains plutonium concentrations for 90 individual bone samples from 19 whole-body donors to the United States Transuranium and Uranium Registries. These samples were divided into 6 groups by bone type: skull (11 samples), long bone end (15), long bone shaft (14), cortical bone (29), trabecular bone (18) and other bones (3). Five of 19 studied individuals were diagnosed with osteoporosis. This study evaluated the effect of osteoporosis on LBMs for estimation of skeleton plutonium activity concentrations. For each bone group (except for the mixed bones), the PCR was performed with and without the 5 osteoporotic cases. The PCR models were fitted for 2 to 6 bones randomly sampled from each group, and 10,000 simulations were run for a given number of sampled bones. Regression residual standard error (RSE) for the PCR simulation was used to evaluate model performance. Excluding 5 osteoporotic cases from analyses significantly improved the PCR models in terms of relative RSE reduction compared to those obtained from the analyses of all 19 cases. The average RSEs for 2 to 6 bones were reduced by 60.2±0.4% for trabecular bone, 56.1±5.1% for long bone end, 53.2±1.8% for cortical bone, 48.4±2.4% for long bone shaft, and 22.4±1.9% for skull. Therefore, separate models should be used for non-osteoporotic and osteoporotic individuals when possible. The RSEs of PCR models for non-osteoporotic individuals were 1.9±0.4 for long bone end (epiphysis), 2.5±0.1 for trabecular bone, 2.8±0.1 for cortical bone, 2.8±0.2 for long bone shaft (diaphysis), and 4.2±0.1 for skull. The non-osteoporotic PCR model, accounting for all bone types, was developed by selecting 3 ‘best’ bones with the lowest RSE in each of 5 bone groups. When the analytical dataset for 14 non-osteoporotic cases was reduced from 90 to 18 bones (15 ‘best’ bones plus 3 others), a further improvement of the PCR model fit was achieved with RSE of 1.4±0.4. Due to the limited number of cases, the model to estimate plutonium concentration for osteoporotic individuals was not proposed. [USTUR-0581-21A]

Presentation Slides

Uncertainty evaluation of skeleton plutonium activity concentration estimated from a latent bone model

Joey Y. Zhou (DOE), Maia Avtandilashvili (USTUR), Sergei Y. Tolmachev (USTUR)

The recently proposed latent bone model (LBM) for non-osteoporotic individuals applies principal components regression (PCR) to estimate plutonium activity concentration in the human skeleton from measurements of a limited set of bone samples. This study developed a Monte Carlo method to evaluate uncertainty in LBM estimates of skeleton plutonium activity concentration (Cskel) using PCR analysis. For this study, the analytical bone dataset was prepared using plutonium concentrations in 18 preselected ‘best’ bone samples from 14 non-osteoporotic whole-body donors to the United States Transuranium and Uranium Registries. The tissue donors’ age ranged from 52 to 87 years and the Cskel ranged from 0.9 to 42.0 Bq kg-1 of wet weight. The bone set contained 3 samples from each of 6 bone types: skull, long bone end (epiphysis), long bone shaft (diaphysis), cortical bone, trabecular bone, and other bones. The PCRs were used to fit LBMs for 2 to 6 randomly sampled bones, and 10,000 simulations were run for a given number of bone samples. The simulation results indicated that the residuals of plutonium concentrations were normally distributed for each of 14 studied cases. The standard deviation of the residuals (SD) of normal distributions were used to determine the uncertainties associated with the estimated Cskel. Linear regression was used to derive a relationship between SD and Cskel for each number of sampled bones. The linear regression equations for 2 and 6 sampled bones were: SD = 0.061×Cskel + 0.846 (r2 = 0.573, p = 0.0011) and SD = 0.024×Cskel + 0.446 (r2 = 0.398, p = 0.0098), respectively. The higher uncertainties were associated with a lower Cskel and a smaller number of sampled bones. As Cskel increased, the estimated relative standard deviations (SD/Cskel) decreased from 100% to 8% for 2 bones and from 52% to 3% for 6 bones. [USTUR-0582-21A]

Presentation Slides
Watch Video (YouTube)

Latent bone modeling approach to select best combination of bones for estimating plutonium activity concentration in human skeleton

Sergei Y. Tolmachev (USTUR), Maia Avtandilashvili (USTUR), Joey Y. Zhou (DOE)

The United States Transuranium and Uranium Registries (USTUR) holds data and bone samples from 290 partial-body tissue donors. Bone samples collected at autopsy were radiochemically analyzed to estimate skeleton activity concentrations of plutonium, americium, and uranium. At the USTUR, the most commonly collected bone samples are rib, sternum, vertebral body, patella, clavicle, and femur middle shaft. Among these, patella, rib, and vertebral body are bones whose collection at autopsy is the easiest. This study applied the recently developed latent bone modeling (LBM) approach to select the best combination of sample bones for estimating the skeleton plutonium activity concentration (Cskel). The analytical bone dataset contained plutonium concentrations for the 6 most commonly collected bones from 14 non-osteoporotic USTUR whole-body tissue donors with known Cskel. The LBM models were built for all possible combinations from these 6 bones. The LMB model residual standard error (RSE) was used to determine the best combination of bones. For two bones (15 combinations), RSEs ranged from 1.096 to 4.888 with the best combination being patella and clavicle; for 3 bones (20 combinations), RSEs ranged from 0.853 to 2.557 with the best combination being patella, clavicle, and rib; for 4 bones (15 combinations), RSEs ranged from 0.792 to 2.073 with the best combination being patella, clavicle, rib, and femur middle shaft; for 5 bones (6 combinations), RSEs ranged from 0.970 to 1.382 with the best combination being patella, clavicle, rib, femur middle shaft, and sternum. The LBM RSE for the 3 easy-to-collect bones (patella, rib, and vertebral body) was 1.522. The LBM RSEs for the two bone combinations of these 3 bones were 1.366 for patella and rib, 2.018 for patella and vertebral body, 2.499 for rib and vertebral body. It is worth noting that the patella and rib combination had smaller RSE (1.366) than that of patella, rib and vertebral body (1.522), although, in general, a smaller RSE is associated with a larger number of bones in the LBM. [USTUR-0585-21A]

Presentation Slides