Projecting radiation-induced cancer risks across time and populations.

Various methods can be used to project the risks of radiation-induced cancer estimated in cohort studies beyond the period of follow-up and to other populations. The epidemiological evidence for the choice of risk projection model is reviewed based on data from studies such as those of the Japanese atomic bomb survivors and UK ankylosing spondylitis patients given x-ray therapy. The results of risk projections based on various approaches are presented, including those suggested by UNSCEAR and by the BEIR V Committee. It is emphasised that the continued follow-up of populations such as the Japanese atomic bomb survivors is of great importance in estimating lifetime risks, and that further parallel analyses are required to examine how risks vary across populations.