NEOPLASTIC DISEASE 



The most striking deviation from the regularity of the 

 relationship is with prostatic cancer. Instead of a straight 

 line with a slope approximately 6, this is close to a straight 

 line with a slope between 12 and 13, lying, of course, in the 

 old-age period only. This is undoubtedly related to hormonal 

 factors and is discussed later. 



60 70 80 



Fig. 1 1 . Similar graph to show different slope of the log-log relationship 

 for all cancers (male) and for prostatic carcinoma C and hypertrophy H. 

 To bring both on the same region the ordinates differ for the two sets. 



It has been pointed out by Armitage and Doll ( 1 954) that 

 one way of accounting for a log-log relationship of this sort is to 

 assume that malignancy results when a cell suffers a sequence 

 of somatic mutations. Let us assume a uniform population 

 of cells remaining constant in composition except for the 

 occurrence of mutation; the cells start in state o. They are 

 subject to mutation at a low constant rate per annum, taking 

 on state i. When a cell takes on state i, it becomes subject to 

 mutation to state 2, and similarly from 2 to 3 and so on. If 

 state n is followed by the development of fatal cancer, then 

 the age incidence of death from cancer should be a straight 

 line with a slope of tz — i when both specific age incidence 

 and age are plotted logarithmically. 



190 



