Entwicklung eines Verfahrens zur numerischen Kalibrierung von Teilkörperzählern

Abstract

At the present state of the art of in vivo measurement techniques the accuracy of the assessment of the incorporated activity is governed by systematical errors due to the calibration of the counting systems rather than due to counting statistics. This applies especially for the in vivo measurement of low-energy photon emitters such as Am-241 and U-235 which are distributed inhomogeneously within the body. With conventional calibration procedures using physical phantoms these problems cannot be solved completely because those phantoms represent only some standard situations and it is extremely difficult to adopt the phantom calibration factors to the individual measuring situation. For this reason, in the Main Safety Department of the Research Centre Karlsruhe (FZK) a new procedure has been developed which allows for the calculation of individual calibration factors. The procedure is based on the simulation of the radiation transport from the contaminated organ or tissue within the body to the detector system using Monte-Carlo techniques (MCNP5 code). The development of the procedure was done in several steps. In the first step, simple simulations have been performed for point sources at well defmed reference points within the shielding room of the FZK partial body counter. These simulations have been performed for the two phoswich detectors as well as for the four HPGe sandwich detectors of the partial body counter. From these simulations both the energy and the spatial depence of the response of the detectors have been derived and compared to the respective measured functions. Based on this comparison some special parameters such as the density of the reflector material of the phoswich detectors or the effective volume of the planar HPGe crystals have been derived. After implemention of these parameters a very good agreement of the calculated and measured values has been achieved. After optimisation of the simulation a voxel-phantom was implemented into the MCNP5 code. For this purpose the MEET Man data set of the Institute of Biomedical Techniques of the University Karlsruhe has been applied. After implementation of the voxel-phantom the radiation transport was simulated from different source regions such as lungs, liver and skeleton taking into account different measuring geometries with the detectors being placed above the lungs, liver or knees, respectively. The derived calibration factors were compared to the respective values measured at the LLNL torso phantom and the bone phantoms of the New York University Medical Centers (NYUMC) and the U.S. Transuranium and Uranium Registry (USTUR). When considering the total efficiency of both phoswich detectors, the comparison revealed a very good agreement of calculated and measured calibration factors. There is, however, a discrepancy between the calculated and measured effiency ratios of the phoswich detectors placed over the left and right hand side of the thorax. The discrepancy has been shown to be due to some asymmetries in the chestwall of both the MEET Man and the LLNL torso phantom. The simulation of the HPGe detectors resulted also in a good agreement of calculated and measured total efficiencies. When considering the single detectors, however, there are significant discrepancies between the results of the simulation and the measurement.