The assessment of dose equivalents, Hp(0.07), to the skin of the hands during manipulation of radioactive sources for radiopharmaceutical preparations is a critical task in radiation protection. During these preparations, alpha-, beta- and gamma-emitting radionuclides in liquid solutions are contained in vials, syringes or other small receptacles, eventually inside plastic or metal shields, and hands are protected against contamination with gloves. Accidentally, some contamination of gloves or even of the naked skin can happen. Estimation of skin doses to fingertips can be carried out exploiting pre-calculated tables of dose rates in standardized geometries [1] , or using softwares based on analytical calculations such as Varskin [2] . We compared systematically, for 20 radionuclides commonly emlpoyed in nuclear medicine, Hp(0.07) in mSv/MBq for point-like and extended sources, either in contact or in presence of different thicknesses of interposed material (glass, plastic), as resulted from severeal releases of Varskin (mod2, 4, 5.3) with the output of Monte Carlo simulations in Geant4. Results indicate that the most recent version of Varskin gives the best agreement with Monte Carlo simulation, having implemented a more accurate treatment of both photons and electrons. A good agreement was found for all nuclides for a point source in contact with skin. When a material is interposed, however, the disagreement between analytical and Monte Carlo results increases with material thickness for some nuclides, particularly for pure beta emitters and low-energy electron emitters. These results can be explained by the deterministic approach in which electron energy deposition is treated in the analytical code and its influence in proximity of electron range, and to the neglection of bremsstrahlung emission. Another source of discrepancy, for some nuclides, is the difference in radionuclide emission spectra. When the accuracy of estimation of skin doses is critical, as after an anomalous exposition event with high activities in extended sources or with a thin layer of interposed material, Monte Carlo simulation is the most accurate calculation approach.
Read full abstract