Densely ionizing (high-LET) galactic cosmic rays (GCR) contribute a signifi
cant component of the radiation risk in free space. Over a period of a few
months-sufficient for the early stages of radiation carcinogenesis to occur
-a significant proportion of cell nuclei will not be traversed. There is co
nvincing evidence, at least in vitro, that irradiated cells can send out si
gnals that can result in damage to nearby unirradiated cells. This observat
ion can hold even when the unirradiated cells have been exposed to low dose
s of low-LET radiation. We discuss here a quantitative model based on the B
aD formalism, an approach that incorporates radiobiological damage both fro
m a bystander response to signals emitted by irradiated cells, and also fro
m direct traversal of high-LET radiations through cell nuclei. The model pr
oduces results that are consistent with those of a series of studies of the
bystander phenomenon using a high-LET microbeam, with the end point of in
vitro oncogenic transformation. According to this picture, for exposure to
high-LET particles such as galactic cosmic rays other than protons, the bys
tander effect is significant primarily at low fluences, i.e., exposures whe
re there are significant numbers of untraversed cells. If the mechanisms po
stulated here were applicable in vivo, using a linear extrapolation of risk
s derived from studies using intermediate doses of high-LET radiation (wher
e the contribution of the bystander effect may be negligible) to estimate r
isks at very low doses (where the bystander effect may be dominant) could u
nderestimate the true risk from low doses of high-LET radiation. It would b
e highly premature simply to abandon current risk projections for high-LET,
low-dose radiation; however, these considerations would suggest caution in
applying results derived from experiments using high-LET radiation at flue
nces above similar to1 particle per nucleus to risk estimation for a Mars m
ission. (C) 2001 by Radiation Research Society.