Comparison of methods for individualized astronaut organ dosimetry: Morphometry-based phantom library versus body contour autoscaling of a reference phantom

Abstract

One of the hazards faced by space crew members in low-Earth orbit or in deep space is exposure to ionizing radiation. It has been shown previously that while differences in organ-specific and whole-body risk estimates due to body size variations are small for highly-penetrating galactic cosmic rays, large differences in these quantities can result from exposure to shorter-range trapped proton or solar particle event radiations. For this reason, it is desirable to use morphometrically accurate computational phantoms representing each astronaut for a risk analysis, especially in the case of a solar particle event. An algorithm was developed to automatically sculpt and scale the UF adult male and adult female hybrid reference phantom to the individual outer body contour of a given astronaut. This process begins with the creation of a laser-measured polygon mesh model of the astronaut's body contour. Using the auto-scaling program and selecting several anatomical landmarks, the UF adult male or female phantom is adjusted to match the laser-measured outer body contour of the astronaut. A dosimetry comparison study was conducted to compare the organ dose accuracy of both the autoscaled phantom and that based upon a height–weight matched phantom from the UF/NCI Computational Phantom Library. Monte Carlo methods were used to simulate the environment of the August 1972 and February 1956 solar particle events. Using a series of individual-specific voxel phantoms as a local benchmark standard, autoscaled phantom organ dose estimates were shown to provide a 1% and 10% improvement in organ dose accuracy for a population of females and males, respectively, as compared to organ doses derived from height–weight matched phantoms from the UF/NCI Computational Phantom Library. In addition, this slight improvement in organ dose accuracy from the autoscaled phantoms is accompanied by reduced computer storage requirements and a more rapid method for individualized phantom generation when compared to the UF/NCI Computational Phantom Library.