Dose-response models and methods of risk prediction and causation estimation

Abstract

Dose-response models are mathematical expressions that describe the relationship between absorbed dose and radiogenic effects. The limited quality and quantity of human dose-response data make it necessary to use fairly simplistic models. Most current low-LET data support the linear-quadratic model in which radiogenic effects are linearly dependent at low doses and then become quadratically curved at higher doses. Some types of effects never exhibit a quadratic component, remaining linear over a wide range of absorbed dose. Future progress in developing more refined dose-response models is more likely to come from a better understanding of the fundamentals of radiation carcinogenesis rather than better data or better curve-fitting techniques. The risk of radiation injury is a prospective estimation of the probability that some harm will result in the future as a consequence of having been irradiated. Quantitative risk estimates for the carcinogenic, genetic, and fetal effects of low level radiation that have been determined by national and international organizations are of the order of magnitude of one chance fatality in 10,000/rem. Causation estimation is the retrospective analysis of the probability that cancer observed in an irradiated individual was caused by radiation as opposed to some other agent. Depending on the dose type of cancer, gender, age at time of irradiation, and time since irradiation, the probability of causation can range from 0% to 100%. Methods for calculation of the probability of causation for certain types of cancer and irradiation circumstances have been developed recently by the National Institutes of Health.