Biologically-based risk estimation for radiation-induced chronic myeloid leukemia

Area of Science:

• Radiation biology
• Cancer epidemiology
• Biostatistics

Background:

• Radiation cancer risk assessment traditionally relies on simple statistical models of epidemiological data.
• Incorporating biological mechanisms into risk models can improve accuracy.
• Chronic myeloid leukemia (CML) is a significant radiation-induced cancer concern.

Purpose of the Study:

• To develop and present a biologically-based linear-quadratic-exponential (LQE) incidence rate model for radiation-induced CML.
• To incorporate biological details such as BCR-ABL induction, CML latency, and cell-killing effects.
• To estimate low-dose gamma-ray CML risk using Bayesian methods and Hiroshima survivor data.

Main Methods:

• Developed a linear-quadratic-exponential (LQE) model incorporating BCR-ABL induction, CML latency, and cell-killing.
• Defined Bayesian priors for nine LQE parameters using diverse biological and epidemiological data.
• Estimated model parameters using maximum likelihood and Bayesian posterior analysis with Hiroshima survivor data.

Main Results:

• The biologically-based LQE model yielded an upper 95% confidence bound for lifetime CML risk of 0.0042 Gy⁻¹ for low-dose gamma-rays.
• This estimate is lower than that from a likelihood-only LQE model (0.0049 Gy⁻¹) and substantially lower than a simple linear dose-response model (0.0158 Gy⁻¹).

Conclusions:

• Biologically-based models offer a more nuanced approach to radiation risk assessment for CML.
• The LQE model provides a refined estimate of radiation-induced CML risk, particularly at low doses.
• This biologically informed approach improves upon traditional statistical risk models.