Microdosimetric properties of ionizing electrons in water: a test of the PENELOPE code system.

The ability to simulate the tortuous path of very low-energy electrons in condensed matter is important for a variety of applications in radiobiology. Event-by-event Monte Carlo codes such as OREC, MOCA and PITS represent the preferred method of computing distributions of microdosimetric quantities. However, event-by-event Monte Carlo is computationally expensive, and the cross sections needed to transport simulations to this level of detail are usually only available for water. In the recently developed PENELOPE code system, 'hard' electron and positron interactions are simulated in a detailed way while soft' interactions are treated using multiple scattering theory. Using this mixed simulation algorithm, electrons and positrons can be transported down to energies as low as 100 eV. To our knowledge, PENELOPE is the first widely available, general purpose Monte Carlo code system capable of transporting electrons and positrons in arbitrary media down to such low energies. The ability to transport electrons and positrons to such low energies opens up the possibility of using a general purpose Monte Carlo code system for microdosimetry. This paper presents the results of a code intercomparison study designed to test the applicability of the PENELOPE code system for microdosimetry applications. For sites comparable in size to a mammalian cell or cell nucleus, single-event distributions, site-hit probabilities and the frequency-mean specific energy per event are in reasonable agreement with those predicted using event-by-event Monte Carlo. Site-hit probabilities and the mean specific energy per event can be estimated to within about 1-10% of those predicted using event-by-event Monte Carlo. However, for some combinations of site size and source-target geometry, site-hit probabilities and the mean specific energy per event may only agree to within 25-60%. The most problematic source-target geometry is one in which the emitted electrons are very close to the tally site (e.g., a point source on the surface of a cell). Although event-by-event Monte Carlo will continue to be the method of choice for microdosimetry, PENELOPE is a useful, computationally efficient tool for some classes of microdosimetry problem. PENELOPE may prove particularly useful for applications that involve radiation transport through materials other than water or for applications that are too computationally intensive for event-by-event Monte Carlo, such as in vivo microdosimetry of spatially complex distributions of radioisotopes inside the human body.