Internal dosimetry: towards harmonisation and coordination of research.

The CONRAD Project is a Coordinated Network for Radiation Dosimetry funded by the European Commission 6th Framework Programme. The activities developed within CONRAD Work Package 5 ('Coordination of Research on Internal Dosimetry') have contributed to improve the harmonisation and reliability in the assessment of internal doses. The tasks carried out included a study of uncertainties and the refinement of the IDEAS Guidelines associated with the evaluation of doses after intakes of radionuclides. The implementation and quality assurance of new biokinetic models for dose assessment and the first attempt to develop a generic dosimetric model for DTPA therapy are important WP5 achievements. Applications of voxel phantoms and Monte Carlo simulations for the assessment of intakes from in vivo measurements were also considered. A Nuclear Emergency Monitoring Network (EUREMON) has been established for the interpretation of monitoring data after accidental or deliberate releases of radionuclides. Finally, WP5 group has worked on the update of the existing IDEAS bibliographic, internal contamination and case evaluation databases. A summary of CONRAD WP5 objectives and results is presented here.

Considerations in assigning dose based on uncertainties from in vivo counting.

Lung counting is highly geometry dependent, especially at low photon energies. Monte Carlo simulations have been used to determine the magnitude of the errors obtained if it is assumed that the deposition is homogeneous, when in fact it is not. Simulation for a germanium lung counting system consisting of four, 70 mm x 30 mm diameter thick detectors have been performed for 70 deposition patterns. The detector efficiencies for 20, 40, 60, 120, 240, 660, and 1000 kiloelectron volts were calculated for a homogeneous deposition and these efficiencies were used to estimate the bias when the deposition was heterogeneous. A bias of 800% was not unusual. Whole-body counting is prone to errors that can arise from the activity distribution and/or size of the subject. The latter are geometry dependent and, for example, a bias result of 200% can be obtained when measuring children in a chair geometry using a calibration factor based on reference man. Thyroid counting is the least prone to error. If the measurement is done correctly, biases can usually be kept below 20%. However, if the measurement is made with a collimator or if the detector is in contact with the subject's neck, biases exceeding 200% can be obtained.

A study of thyroid radioiodine monitoring by Monte Carlo simulations: implications for equipment design.

Monte Carlo simulations have been performed to evaluate the design of collimated detectors used to measure 125I or 131I in the thyroid gland. Two detector sizes were simulated for each radioisotope: (i) for 125I monitoring 2.54 cm diameter and 7.62 cm diameter and 0.2 cm thickness and (ii) for 131I monitoring 2.54 cm diameter, 3.2 cm thickness and 7.62 cm diameter, 6.4 cm thickness. The virtual thyroid gland was 20 g. Activity was placed in both the gland and the remainder of the body in varying amounts to assess the efficacy of collimation. The results show that the detector should be sufficiently large so that its solid angle of acceptance when placed 15 cm anterior to the skin surface will include the whole of a moderately enlarged thyroid gland. Heavy collimation to reduce the contribution of extrathyroidal radioiodine within the subject's body is not normally required. It may be of more value as a positioning device and spacer ensuring an appropriate and constant neck to detector distance than in cutting down counts from extrathyroidal activity. In specifying a sensitive detector system for monitoring intrathyroidal radioiodine, a wide angle of acceptance and sufficient detector crystal thickness take precedence over collimation and shielding.

The Canadian whole body counting intercomparison program: a summary report for 1989-1993.

The Human Monitoring Laboratory, which is the Canadian National Calibration Reference Centre for In-Vivo Measurements, has been conducting an annual intercomparison program for whole body counting facilities since 1989. Each year a number of phantoms are prepared for participants to measure the accuracy, precision, size dependency, and identification capabilities of their whole body counters. Other tests are designed to measure the sensitivity of the whole body counter to geometry factors. Some phantoms are prepared with the same radionuclide from year to year so that time dependent data can be acquired, whereas other phantoms are prepared as unknowns. This article describes the tests that were performed, gives the results for Canadian participants, and compares the results to interim performance criteria. Measurement results from a human volunteer who ingested 137Cs contaminated caribou meat and was counted at a number of facilities in 1990 are also presented.

The optimization of the analysis of 63Ni in urine.

A method has been developed that separates 63Ni from urine. The subsequent estimation of activity is by liquid scintillation counting. The urine is wet-ashed by a new procedure which is much faster than conventional ashing methods. Interfering phosphates are removed before a precipitation of Ni as the dimethylglyoxime complex. The effects of the carrier weight, the pH of the phosphate removal step and the pH of the dimethylglyoxime precipitation have been investigated and optimized to give a mean recovery of 97 +/- 8% for 63Ni. The detection limit as defined by Currie (1968) is estimated to be 4.0 pCi (147 mBq) per sample.

Lung counting: a function to fit counting efficiency of a lung counting germanium detector array to muscle-equivalent-chest-wall-thickness and photon energy using a realistic torso phantom.

A function has been developed that will fit counting efficiency to chest wall thickness, adipose content of the chest wall, and photon energy. The adipose content of the chest wall is removed as a variable by use of the derived quantity: muscle equivalent chest wall thickness. The function has been tested on experimental data sets obtained from the LLNL and JAERI torso phantoms using a variety of germanium detector lung counting systems. The function is applicable over the range of existing phantoms and the range of equipment types tested. Subtle differences between the detector systems are shown by differences in the fit parameters. This analysis could be a useful addition to lung counting analysis software. The LLNL and JAERI phantoms have very similar counting characteristics as evidenced by the similarity of the function parameters when measured using a single detector system.

The Canadian National Calibration Reference Center for Bioassay and in-vivo Monitoring: a program summary.

The Canadian National Calibration Reference Center for Bioassay and in-vivo Monitoring is part of the Radiation Protection Bureau, Department of Health. The Reference Center operates a variety of different intercomparison programs that are designed to confirm that workplace monitoring results are accurate and provide the necessary external verification required by the Canadian regulators. The programs administered by the Reference Center currently include urinalysis intercomparisons for tritium, natural uranium, and 14C, and in-vivo programs for whole-body, thorax, and thyroid monitoring. The benefits of the intercomparison programs to the participants are discussed by example. Future programs that are planned include dual spiked urine sample which contain both tritium and 14C and the in-vivo measurement of 99mTc.

Examination of the effect of counting geometry on 125I monitoring using MCNP.

This paper theoretically investigates how the counting efficiencies of three selected detector (NaI) sizes are affected by neck-detector distance, detector misplacement, thickness of overlaying tissue, and size of the thyroid gland. The detector sizes were small, 2.54 cm diameter and 0.2 cm crystal thickness; medium, 7.62 cm diameter and 0.2 cm crystal thickness; large 30.48 cm diameter and 0.2 cm crystal thickness. The evaluations were determined using a Monte Carlo technique. The simulations show that the large detector minimizes the uncertainties for all problems except the thickness of overlaying tissue.

Chest wall thickness measurements: the alternative approach extended for 241Am.

The Human Monitoring Laboratory has extended the technique of determining the chest wall thickness of an individual using information from the spectrum produced by internally deposited radionuclides. The technique has been investigated both theoretically and practically using germanium detectors and the Lawrence Livermore Torso Phantom. The phantom was used with a lung set containing homogeneously distributed 241Am. Chest wall thicknesses were varied by using a series of muscle equivalent overlay plates that gave a range of 1.6 cm to 3.9 cm thickness. It was found that a 3-cm chest wall thickness can be estimated to within 18%. Using a spectral addition technique 1 kBq was estimated to be the "practical" lower limit of activity for this method.

The assessment of the effect of thyroid size and shape on the activity estimate using Monte Carlo simulation.

Monte Carlo simulations have been used to assess the uncertainty introduced into an activity estimate of radioiodine (125I and 131I) in the thyroid when the size and shape of the gland differs from that of the calibration phantom. The detector dimensions for the 125I simulations were small (diameter 2.54 cm, thickness 0.2 cm); medium (diameter 7.62 cm, thickness 0.2 cm); large (diameter 30.48 cm, thickness 0.2 cm). The detector dimensions for the 131I simulations were small (diameter 2.54 cm, thickness 3.2 cm); medium (diameter 7.62 cm, thickness 6.4 cm); large (diameter 30.48 cm, thickness 11.0 cm). Shapes simulated have included thyroid glands with a third (pyramidal) lobe, no isthmus, and rotated lobes. Sizes simulated have been 10 g, 20 g, and 40 g. The results show that the size of the uncertainty is dependent on the detector size, the neck-to-detector distance, and the type of radioiodine being measured (i.e., 131I or 125I). The worst case bias (on contact counting) for either 125I or 131I using the different sized detectors is as follows: small is between -40% and 40%; medium is between -33% and 17%; large is between -20% and 5%. If the detectors are placed at about 15 cm from the neck the bias values drop so that the uncertainty introduced if the subject has a smaller than standard thyroid becomes insignificant and becomes much improved if the thyroid is larger than the standard. The bias values for the detectors are as follows: small is between -20% and 5%; medium is between -23% and 2%; large is between -21% and 2%.

The effect of lung deposition patterns on the activity estimate obtained from a large area germanium detector lung counter.

Lung counters for in vivo detection of low energy photon emitters are typically calibrated using phantoms containing lung tissue equivalent material with the radioactivity homogeneously distributed throughout the material. If the activity in a measurement subject is heterogeneously distributed, the activity estimate for that subject will be uncertain due to the assumptions of distribution. The magnitude of the uncertainty for a four-detector germanium array, using the Lawrence Livermore National Laboratory torso phantom with a newly designed lung set that allows the activity to be localized in one or more of 16 areas, was estimated. The results show that detector arrays will reduce the uncertainties arising from the geometry of the lung deposition compared to single detectors. The estimated activity of an internal deposition that emits 17.5 keV photons can be overestimated by a factor of three, or underestimated by a factor of infinity (i.e., the activity is missed completely). As the photon energy rises to 59.5 keV the uncertainty in the activity decreases so that the maximum overestimate (underestimate) will be a factor of two (five). As the energy rises to 344.3 keV only the maximum underestimate changes: it becomes a factor of three.

Chest wall thickness measurements for enriched uranium: an alternative approach.

The Human Monitoring Laboratory has developed a technique to determine the chest wall thickness of an individual using information from the spectrum produced by internally deposited radionuclides. The technique has been investigated both theoretically and practically using phoswich detectors and the Lawrence Livermore Torso Phantom. The phantom was used with lung sets containing homogeneously distributed 93% enriched uranium, 20% enriched uranium, natural uranium, and 241Am. It was found that a 3-cm chest wall thickness can be estimated to within 9% when measuring 93% enriched uranium. The technique does not work for the latter two radionuclides because of an insufficient separation in the photon energies and poor resolution of the phoswich detectors. The technique is only of value for activity levels well above the detection limit.

An evaluation of germanium detectors employed for the measurement of radionuclides deposited in lungs using an experimental and Monte Carlo approach.

A study was undertaken to evaluate the performance of an advanced design broad energy germanium detector for the in vivo measurement of radionuclides in lungs. Relative counting efficiency, background, and sensitivity for lung counting arrays consisting of four, three, and two 80-mm-diameter by 20-mm-thick (80 x 20 mm) broad energy germanium detectors were simulated by collecting spectra with the single 80 x 20 mm broad energy germanium at each of four locations over a humanoid torso phantom. Regions of interest were evaluated for photon energies ranging from 17 to 1,500 keV. The 80 x 20 mm detector arrays were then benchmarked against a standard array of four 70-mm-diameter by 20-mm-thick (70 x 20 mm) broad energy germanium detectors. Since testing new equipment can be an expensive and time consuming process, an alternative approach, using Monte Carlo simulations instead of physical measurements, was also evaluated and compared to experimental data. With this approach, counting efficiency and minimum detectable amount were simulated for two sizes of germanium detectors (70 mm and 80 mm diameter) at four different crystal thicknesses (15, 20, 25, and 30 mm). For the experimental measurements, arrays consisting of three and four 80 x 20 mm broad energy germanium detectors resulted in an increase in counting efficiencies, relative to the standard array, at all photon energies. The greatest relative increase was observed for the four-detector array (24-35%). In contrast, counting efficiency decreased, relative to the standard array, by 24-28% with a two-detector array. Arrays consisting of two and three 80 x 20 mm broad energy germanium detectors resulted in decreased relative background at all photon energies, with the exception of the 946 keV photon for the three-detector array. The most significant decrease in background occurred with the two-detector array (28 to 40%), while background was increased by 18-43% for the four-detector array. Arrays consisting of three and four 80 x 20 mm broad energy germanium detectors resulted in increased relative sensitivity at all photon energies. The three-detector array provided the greatest sensitivity at photon energies below 344 keV. The four-detector array provided slightly better measurement sensitivity at photon energies greater than 344 keV. The two 80 x 20 mm detector array provided sensitivity unexpectedly comparable to the standard array. Monte Carlo predictions on how size affects counting efficiency and minimum detectable amount agreed well with the experimental results. From the Monte Carlo predictions, the effect of detector thickness on counting efficiency was unimportant at photon energies up to 60 keV and independent of detector diameter. At higher photon energies for both detector diameters, the counting efficiency decreased as the thickness decreased. The values of minimum detectable amount for the 70-mm and 80-mm diameter detectors did not differ by more than 15% at 17 keV or 20% at 60 keV when compared to detectors of equivalent thickness. Minimum detectable amount increased slightly at 17 keV and rose by approximately 52% at 660 keV, with decreases in thickness from 30 mm to 15 mm.

The BRMD BOMAB phantom family.

The Human Monitoring Laboratory (HML) has used the International Commission on Radiological Protection's Report on Reference Man and Canadian anthropomorphic data as guidance to design and construct a family of phantoms corresponding to Reference Man (PM), Reference Woman (PF), Reference Ten-Year-Old (P10), Reference Four-Year-Old (P4), Ninety-five Percentile Man (PM95), and Five Percentile Man (PM5). The PM series also has an accessory chest section (PMacc) to better simulate lung depositions. The phantoms are constructed from high-density polyethylene and fitted with end-recessed filling caps to minimize leakage problems. This paper describes the methodology of construction and presents data so that the phantoms can be reproduced. The phantoms have been used in Canada's National in-vivo Intercomparison Program, and results show that all Canadian in-vivo counting facilities have size-dependent calibrations. Selected data are presented to exemplify this dependence.

The uncertainty in the activity estimate from a lung count due to the variability in chest wall thickness profile.

Calibration of a lung counter requires the use of a realistic torso phantom. The depth profile of both torso phantoms' (LLNL and JAERI) chest plate covers is fixed and assumed to be equivalent to a person's chest wall; however, ultrasound measurements of humans have shown this to be an approximation. When the depth profile of a calibration phantom is different from that of a subject, then a systematic uncertainty will be introduced into the activity estimate. Monte Carlo simulation has shown that changes in the depth profile of the chest wall thickness affect the counting efficiency. Ultrasound measurements have suggested that the coefficient of variation in the depth profile of the chest wall thickness lies between 13% and 26% for male workers; therefore, the added uncertainty to an activity estimate will be an over or underestimate of about a factor of 1.07 resulting from the different depth profile. The factor will be somewhat higher for females, probably about 1.2 at the extreme. These additional uncertainties resulting from depth profile differences are small compared with other uncertainties commonly encountered in lung counting: detector positioning, deposition patterns of the activity, measurement of the chest wall thickness, etc.

Radiocesium in children from Belarus.

The body burdens of 137Cs were measured in 74 children from Belarus who had been exposed to fallout from the Chernobyl accident. The children were between the ages of 8 and 12 y old and were visiting the Ottawa area during the summers of 1991 and 1992. The body burdens varied from 0.04-2.25 kBq, which are less than or comparable to the amount of natural 40K in the children's bodies. There was no apparent dependence on age or sex. The body burdens were related, to some extent, with recorded fallout levels in the region of origin but did not appear to change significantly from 1991 to 1992. During their stay in Canada, radiocesium was being cleared from the children's bodies with a mean half-time of 33 d (range = 12-77 d). The highest intake rate of 137Cs was estimated to be 18 Bq d-1 and the highest dose rate was 0.13 mSv y-1.

Radiocesium body burdens in residents of northern Canada from 1963-1990.

Measurements of 137Cs body burdens in over 1,100 people from five northern Canadian communities were carried out with a portable whole body counting system during the winters of 1989 and 1990. These results are compared with over 3,000 similar measurements carried out during 1967-1969. Community mean body burdens and body concentrations had decreased by approximately a factor of 30 between the two survey periods. The dependence of body concentrations on the sex and age of the subjects has also changed significantly. This can be related to changes in the patterns of caribou consumption in the northern communities. Measurements of 137Cs in urine are also available for an earlier period (1963-1966) when world-wide fallout was at its highest level. A normalization procedure was developed to calculate the average radiocesium body concentration in each community from the concentrations in urine. From data spanning a period of nearly 30 y (1963-1990), lifetime radiation doses have been estimated for most communities in the Yukon and Northwest Territories. These cumulative doses vary from 0.3 to nearly 40 mSv, with an Arctic-wide average of about 12 mSv. No health effects would be expected at these levels.

Chest wall thickness measurements of the LLNL phantom for small area germanium detector counting.

The Lawrence Livermore National Laboratory (LLNL) phantom was developed to calibrate lung counting systems that are used to estimate plutonium and other low energy photon emitting radionuclides deposited in the lung. Originally, low energy photon counting systems consisted of sodium iodide or phoswich detectors, but they have been largely replaced by smaller germanium detector arrays. The average chest wall thicknesses of the LLNL phantom's torso plate and its overlay plates provided by the manufacturer refer to the regions covered by phoswich detectors; however, germanium detectors are of a different size and are placed in different locations on the phantom's torso plate. Previous work has shown that the manufacturer's data were not applicable for large area germanium detectors. The lung counting system at the Korea Atomic Energy Institute (KAERI) is a small area germanium detector array. Although the detectors are placed within the phoswich circles, only about 25% of the area is covered by the detectors. The LLNL phantom at KAERI has been examined to determine if the manufacturer's data are valid or if new chest wall thickness values must be determined. This paper presents chest wall thickness data for the LLNL phantom with and without its B-series overlay plates at 17 keV, 60 keV, 200 keV, and 1,500 keV and shows that these values are different from the manufacturer's values.

Comparison of the LLNL and JAERI torso phantoms using Ge detectors and phoswich detectors.

The Human Monitoring Laboratory has compared the LLNL and JAERI torso phantoms using its germanium detector lung counting system by measuring the counting efficiencies for radioactive materials in the phantoms at photon energies of 17.7 keV, 59.5 keV, 121.8 keV, and 344 keV to assess the similarity (or differences) in performance characteristics. The counting efficiencies obtained from the two phantoms were compared by converting the Chest Wall Thickness data and Adipose Mass Fractions of the phantoms to Muscle Equivalent Chest Wall Thicknesses. The counting efficiencies for the two phantoms were found to be within a factor of 1.44 of each other at 17.7 keV, 1.30 at 59.5 keV, 1.25 at 121.8 keV, and 1.17 at 344 keV when using a four detector array (JAERI efficiency divided by LLNL efficiency). However, individual detector responses show that the counting efficiencies from the two phantoms differ considerably in the region of the heart (up to a factor of 6 at 17 keV). Other areas above the lungs give counting efficiencies that are similar to each other. A routine intercomparison exercise with Cameco Corporation has shown that the counting efficiencies derived from the LLNL and JAERI phantoms were found to be within a factor of 1.18 (JAERI/LLNL) when a natural uranium lung set was used to calibrate a lung counter consisting of phoswich detectors. This work has also shown that over the energy range 63 keV-185 keV the LLNL phantom can be used to calibrate phoswich detector systems that are positioned on the back of the subject.

A joint HML-KAERI project--comparison of the LLNL and JAERI torso phantoms using four 50 mm Ge detectors.

The Health Physics Department of the Korea Atomic Energy Research Institute and the Human Monitoring Laboratory have collaborated to compare the LLNL and JAERI torso phantoms. The counting efficiencies of the phantoms at 17.7 keV, 59.5 keV, 121.8 keV, and 344 keV were measured with KAERI's germanium lung counting system. The data were made comparable by converting the chest wall thicknesses and adipose mass fractions of the phantoms to muscle equivalent chest wall thicknesses. The counting efficiencies of the two phantoms are within 12% to 17% of each other at 17.7 keV, 15% to 22% at 59.5 keV, 10% to 15% at 121.8 keV, and 7% to 10% at 344 keV. This joint study has shown that the LLNL and JAERI phantom are essentially equivalent for the purposes of calibrating a lung counting system that consists of two ACTII germanium detectors.