Monte Carlo mixture model of lifetime cancer incidence risk from radiation exposure on shuttle and international space station

Abstract

Estimating uncertainty in lifetime cancer risk for human exposure to space radiation is a unique challenge. Conventional risk assessment with low-linear-energy-transfer (LET)-based risk from Japanese atomic bomb survivor studies may be inappropriate for relativistic protons and nuclei in space due to track structure effects. This paper develops a Monte Carlo mixture model (MCMM) for transferring additive, National Institutes of Health multiplicative, and multiplicative excess cancer incidence risks based on Japanese atomic bomb survivor data to determine excess incidence risk for various US astronaut exposure profiles. The MCMM serves as an anchor point for future risk projection methods involving biophysical models of DNA damage from space radiation. Lifetime incidence risks of radiation-induced cancer for the MCMM based on low-LET Japanese data for nonleukemia (all cancers except leukemia) were 2.77 (90% confidence limit, 0.75–11.34) for males exposed to 1 Sv at age 45 and 2.20 (90% confidence limit, 0.59–10.12) for males exposed at age 55. For females, mixture model risks for nonleukemia exposed separately to 1 Sv at ages of 45 and 55 were 2.98 (90% confidence limit, 0.90–11.70) and 2.44 (90% confidence limit, 0.70–10.30), respectively. Risks for high-LET 200 MeV protons (LET=0.45 keV/μm), 1 MeV α-particles (LET=100 keV/μm), and 600 MeV iron particles (LET=180 keV/μm) were scored on a per particle basis by determining the particle fluence required for an average of one particle per cell nucleus of area 100 μm 2. Lifetime risk per proton was 2.68×10 −2% (90% confidence limit, 0.79×10 −3%–0.514×10 −2%). For α-particles, lifetime risk was 14.2% (90% confidence limit, 2.5%–31.2%). Conversely, lifetime risk per iron particle was 23.7% (90% confidence limit, 4.5%–53.0%). Uncertainty in the DDREF for high-LET particles may be less than that for low-LET radiation because typically there is very little dose–rate dependence. Probability density functions for high-LET radiation quality and dose–rate may be preferable to conventional risk assessment approaches. Nuclear reactions and track structure effects in tissue may not be properly estimated by existing data using in vitro models for estimating RBEs. The method used here is being extended to estimate uncertainty in spacecraft shielding effectiveness in various space radiation environments.