Skip to main content

Internal dosimetry - enhancements in application

Exploitable results

The Application of MIRD phantoms in whole body counting in the scanning geometry. To introduce computer supported calibration techniques the measurements of bottle phantoms filled with a Ho-166m and Co-60 solution on a whole body counter in scan geometry have been compared to computer simulations. For the simulation the Monte Carlo n-particle simulation code MCNP has been used to calculate the peak efficiency of the experiment. On the other hand side uncertainty factors of both, the measurement and the simulation, have been calculated to find out which method is more accurate. For the comparison 5 different bottle phantoms with different weight and height factors have been used. These phantoms have been simulated by using simple mathematical descriptions of the bottles. The scanning detector geometry has been approximated by calculating the peak efficiencies at 7 fixed points of the detectors and by calculating the average of these points. The differences between the results of the simulation and measurement were less than 8%. The most important factors influencing the standard deviation of the simulation are listed below. Here it should be explained how those values have been calculated. All uncertainties are stated as one standard deviation: -The most obvious factors are the statistical uncertainty factors of the measurement and the simulation due to a finite amount of events. -Due to the production process the calibration sources of the bottle phantoms also have a specific standard deviation of 3 %. -Different kinds of experiments have been performed to analyse the standard deviation due to geometrical reasons. The sum of all these geometrical factors results in an over all uncertainty of 2.0 %. -As mentioned above the integration over the efficiencies at every point is performed by averaging over 7 points the standard deviation amended to 2 %. -The simulation itself uses optimised models for the simulated physical effects, which can be summed to an uncertainty of 2 % and is called the simulation uncertainty. It can be seen that the calibration by measurement and calibration by simulation have comparable accuracies of 4.1% and 3.8% respectively. So the simulation can be used to calibrate the peak efficiency of a whole body counter. This gives the opportunity to use more anthropomorphic phantoms. In this study MIRD (Medical internal radiation dose) phantoms have been used to calculate uncertainty factors due to weight and height of the subject and due to inhomogeneous distributions in the human body. For whole body counters in the scan geometry the uncertainty due to height differences to the standard height of 170cm is less then 5%, but the uncertainties due to weight deviations to the standard weight of 70kg can be up to 25%. Inhomogeneous activity distributions influence the result by up to 25 % for whole body counters in the scanning geometries. These high influencing factors have been seen for low photo energies short time after intake or a long time after intake of Am-241 and Ra-226 respectively.
The work which was carried out within the work package "bioassay" shall provide commonly acceptable guidelines for optimum performance of inductively coupled plasma mass spectrometry (ICP-MS) measurements with main focus on urinary measurements of uranium, thorium and actinides. From the results of this work following recommendations can be made: Sampling of urine: 24 hours urine sampling should be collected to avoid large uncertainties in the quantitation of daily urinary excretion values. Wherever possible samples for three consecutive days should be collected and a number of aliquots should be drawn out of total sample for analysis. Care should be taken to avoid external contamination of samples by collecting samples in pre-cleaned sample polyethylene bottles. For storage, urine samples should be acidified and kept frozen before analysis. Sample preparation: Pre-dilution is of advantage where sensitivity of the instrument is adequate, in order to avoid obstruction of the cones of the transfer system from the ICP torch to the vacuum of the mass spectrometer. Salt removal from urine samples might be advantageous under certain circumstances but extreme care should be taken during handling and chemical processing to ensure that the reagent blank is extremely small. Once the salt removal chemical procedure with extremely low blank is established, only then it should be adopted. Otherwise the normal dilution procedure should be persisted with. In addition, several chemical blanks must be prepared simultaneously during the entire sample preparation procedure. At least 3 aliquots of each urine sample should be analysed to take into account any possible variations during the measurement by ICP-MS technique. Reagent blanks should be run with every sample, they should be quantified and correction due to these blanks should be applied. Measurement of total U in urine by ICP-MS at physiological levels (< 10ng/L) requires no sample preparation besides UV photolysis and/or dilution. At these concentrations levels, reliable measurements of 235U (or 235U/238U ratios) require pre-concentration steps such as destruction of organic material of the sample using microwave digestion and selective chemical purification using UTEVA resin. The advantage of using ICP-MS is that the measurement time is considerably shortened (10 minutes versus several days for alpha spectrometry). ICP-MS technique: ICP-MS measurements are simple, rapid and economical. New improved measuring techniques (HR-SF-ICP-MS) with detection limits in urine of 150pg/L for 238U, 30pg/L for 235U and 100pg/L for 232Th, respectively, are recommendable and offer much lower detection limits than alpha spectrometry. This method should become the routine technique monitoring of workers and of members of the general public for U isotopes incorporated in the human body. In addition to the technical questions of bioassay measurements, following substantial points have to be considered. The calculation of internal doses from incorporated radionuclides depends critically on the reliability of the biokinetic models employed. Preferably, these models should be based on sound physiological information and the parameters should be derived from experimental investigations on humans. However, even for radiologically important elements like U, Th, radium or lead, there are inconsistencies in the currently adopted biokinetic models. There is therefore a persisting need for further testing and improvement of these models. This can either be achieved through controlled experimental investigations or by testing the models against monitoring data. Thus, models and measurements may alternatively be used to validate each other. All of these aspects are addressed in this project. The model predictions for U and Th were calculated for various exposure scenarios and were compared with bioassay results, which were done in this project and autopsy data available in the literature. Individual monitoring and biokinetic modelling for a Finnish family due to natural U in drinking water was performed. Measured excretion was compared with model prediction. It was found that reasonable results are achieved using the ICRP biokinetic model for U against measured data. However, a discrepancy was found between the model and measurement for Th in excretion as well as in bone and whole body. It is proposed that a higher gastrointestinal absorption value might be separately given for ingested Th. Those recommendations might contribute to a more reliable monitoring of incorporated radionuclides in practice and further improve realistic dose coefficients of U and Th in internal dosimetry.
The main activities of our institution directed to compile the scientific results achieved by other project partners and summarize the outcomes from the point of view of their possible applications. The main goal of the project was to improve the in vivo and bioassay monitoring techniques to be able to provide improved methods for routine applications. Since the benefits and innovative features of the achieved results are detailed by the corresponding project partners, only the main summary conclusions are given here. In vivo monitoring: Though HPGe detectors have quite good characteristics for monitoring low energy photon emitters, in certain circumstances the need of detector cooling may limit their use in in-vivo applications. It seems very promising to apply properly manufactured Si detector arrangements (array, mosaic) for in-vivo measurement of 239Pu deposited in the lung since the lower limit of detection can be improved just because Si detectors are working in room temperature and the cooling system are not hindering the optimum detector positioning on the chest surface. In the higher photon energy range (above 100keV) the conventional large NaI(Tl) scintillation detector can advantageously be applied if the requirements for the spectral energy resolution is not too high. In this case another aspect can also be considered namely the installation and operational costs are much lower for scintillation detectors. However, especially in case of complex gamma spectra large NaI(Tl) detectors cannot compete with HPGe detectors. As for the calibration of in vivo monitoring systems introducing numerical calibration technique instead of using physical phantoms the main advantages can be summarised as follows. The method provides the possibility of: - source-detector geometry optimisation - simulation different activity distributions - individual specific calibration - time-dependent efficiency calibration - numerical reconstruction of any body part The difficulty in widespread application of the numerical calibration technique in the routine practice is that the method needs quite powered computer, special software, and properly skilled and experienced personals. Since all the conditions are available in certain, well-equipped and prepared institutions, those laboratories dealing with routine internal contamination monitoring can take the advantage to ask for this service. Bioassay monitoring: Mass spectrometry, and especially inductively coupled plasma mass spectrometry (ICP-MS), has evolved as an attractive alternative to alpha spectrometry and beta counting in bioassay monitoring. The work carried out in this project provided acceptable guidelines for optimum performance of ICP-MS measurements of uranium, thorium and certain actinides, including sampling procedure, operational parameters of the instruments, and interpretation of the measured data. This very sensitive technique allows for the detection of atoms present during a given counting time. More simple, rapid and cost effective method of mass-spectrometry was chosen to standardise bioassay monitoring procedure using high resolution sector field inductively coupled mass spectrometer (HR-SF-ICP-MS), by means of which it is possible to measure very low amount of thorium and uranium with fairly good precision. The detection limit for natural uranium isotopes is generally in the order of 20pg or about 0.5µBq. Summarising the findings on bioassay investigations the application of ICP-MS measurements are relatively simple, rapid and cost saving. The analytical capabilities of ICP-MS studied so far seem to provide sufficient proof that this method has potential to apply for the member of the public and to become as a routine technique to monitor workers for most of the radionuclides incorporated in the body.
The radiation safety of atomic industry personnel handling radionuclides, Fission and Activation Products (FAP), actinides, etc., inevitably involves radiation monitoring and internal exposure control. The specialized techniques of individual monitoring used at atomic enterprises include the application of dynamic air concentration assessment in the working room air, sampling of human excreta and in vivo monitoring. Even though, in vivo monitoring presents several advantages such as the possibilities of fast assessment of incorporated activity and high counting efficiency, it has several drawbacks as well. The most crucial one is the need for interpretation of the measurement, to convert the number of pulses in spectrometry channels into retained activity. This process, called calibration is based on the use of plastic phantoms (mannequins) containing a well-known activity. However, the standard reference phantoms do not take into account the individual anatomy of a given subject, the specific size, shape, weight of his (her) organs as well as their placement in the body. To make the things worse, the most radiotoxic incorporated radionuclides (such as 239Pu and 241Am) emit low-energy (13-60keV) -rays. As a result, determination of whole body (or organ) counting calibration coefficients is hampered due to intensive absorption and scattering of radiation in the patient's body (organs). To circumvent these obstacles, one could use mathematical simulation (particularly Monte Carlo Method (MCM) as almost the only valid calculation method for such a purpose) rather than measurement of reference phantoms. The multi-platform (MS Windows and UNIX) graphic user interface OEDIPE], French acronym corresponding to "tool for assessment of personalised internal dose" which has been developing at Institut de Radioprotection et de Surete Nucleaire (IRSN, France) is a very good example of what it could be done in this field. It has been specially designed for applications in internal dosimetry and whole body counting. To perform radiation transport calculations, the software OEDIPE creates automatically MCNP input files notably based on individual patient tomograms. The potential of OEDIPE in the frame of IDEA project has been studied through two applications. The first is related to the measurement of high-energy emitters in whole body counting and the second is related to the measurement of actinides in the lung. The application of voxel phantoms in whole body counting The study was performed with first, the creation of the voxel phantoms of the physical IGOR phantom family (6 sizes) dedicated to Fission and Activation Products measurement and second, the validation of experimental results with simulations. Agreement is rather good between measurement and simulation for all energies and any type of phantom since an average difference of less than 15% is observed showing the ability of Oedipe to be used in WBC for further investigations. The application in low energy monitoring The strategy used was practically the same as applied in case of whole-body phantoms. The utility was firstly tested for low and medium energy actinide emitters on anthropomorphic Livermore phantoms, the mannequins generally used for in vivo counting, in order to compare the results of simulation and measurement. From these results, two kind of experiments were done: (i) the demonstration of the utility's abilities for the simulation of real facilities showing a good agreement between simulation and measurement for energies higher than 20keV with a difference less than 10 and about 20% at 17.51keV and (ii) the study of geometry uncertainties, such as different anthropomorphic phantoms or different source geometries, on in vivo calibration was investigated. CONCLUSION AND OUTLOOK In the frame of the European project, the validity of the new software has been tested for low-energy -ray spectrometry of actinides and whole body counting when measuring plastic phantoms and its potentiality in various measurement situations. As a result, the interface proves to be useful for whole body counting calibration where the influence of a calibration phantom to the interpretation of the measured data has been studied. Moreover, the software enables taking into account individual patient's anatomy, which may result in up to two-fold and more correction factors for activity assessment. Consequently, as a result of its flexibility in accommodating complex geometry, the method developed not only represents a diagnostic tool for in vivo measurement, but also opens up new possibilities such as the optimisation of detection systems, the study of contamination with mixed actinides and any other simulation using MCNP where complex geometry is derived from a set of superimposed images.