Age distribution and a multi-stage theory of carcinogenesis: 70 years on

Abstract

The Armitage Doll model demonstrates that the impact of age-dependent exposure to carcinogenic factors depends on whether the induced change occurs early, at the midpoint or late in carcinogenesis. 70 years on, updated modelling shows that their epidemiological observations still provide insight into clinical observations and their underlying molecular mechanisms.

Fig. 1. Change in incidence and mortality for a selection of cancers between Armitage and Doll publication and the modern era.

Change in incidence (2018) and mortality (2018 and 1950/51) with age for cancer of the prostate, bladder, colon and pancreas in men (m) and for cancer of the breast and lung in women (f) shown on a double logarithmic scale, that is, the logarithm of the event rate per million persons plotted against the logarithm of the mid-point of the age group. The data from the Armitage and Doll paper was extracted manually from their original figures including the straight line (displayed in red) which was drawn arbitrarily to give the best fit to the 1950/51 mortality data subject to the gradient being 6 to 1.