Quantitative models for lung cancer induced by cigarette smoke.

This discussion paper gives a limited history of work done at this Institute on quantitative modeling relating to lung cancer and cigarette smoking, a health hazard whose study has been given much encouragement by Norton Nelson. It first starts with the proposal that life shortening be considered as a measure of the impact of lung cancer using log normal and Weibull types of distributions of time to occurrence; second, it continues with an examination of the fits of the log normal and Weibull distributions to the Doll and Hill data on smoking and lung cancer in British physicians and a systematic review and development of mathematical models of carcinogenesis; and third, it reports on the current work that points out inconsistencies in the Armitage-Doll multistage model with the Doll and Hill data and suggests a two-stage clonal growth model that assumes promotion of clonal growth is restricted to cells initiated by the smoke. This proposal and related work support a current trend in risk assessment to adopt a two-stage clonal growth model that incorporates birth and death rates of cells and the transitional probabilities of the stages.