Janine Becker, Felix Heinzl, Maria Schnelzer, Peter Scholz-Kreisel, Munechika Misumi, Hiromi Sugiyama, Kyoji Furukawa, Markus Eidemüller
This paper updates radiation risk models for leukaemia and lymphoma incidence that can be applied to a variety of adult exposure scenarios and ages of onset of leukaemia and lymphoma, derived from the Life Span Study of atomic bomb survivors of Hiroshima and Nagasaki. Models were developed for four groups of haematological malignancies: acute lymphoblastic leukaemia (ALL), acute myeloblastic leukaemia (AML), chronic myeloid leukaemia (CML), and Hodgkin and non-Hodgkin lymphoma (LYM). Multi-model inference methodology was applied for all groups to construct flexible radiation risk models. The models revealed complex radiation risk patterns with strong dependencies on dose, age, and sex. The ALL group showed high radio-sensitivity with risk decreasing strongly with attained age. The AML group demonstrated a strong non-linear dose response requiring special considerations for fractionated or protracted exposures. The CML model showed a strong dependence on time since exposure. Excess relative risk for the lymphoma group was smaller than other groups and displayed significant sex differences. The developed models for leukaemia and lymphoma provide robust risk predictions for a large variety of adult exposure scenarios with ionising radiation. Multi-model inference proved to be an important tool in the construction of such flexible risk models. The models can be safely used in risk assessment tools or for lifetime risk evaluations.