A population perspective on multistage carcinogenesis.

The concept of multistage carcinogenesis is one of the central dogmas of cancer research today. Recent laboratory work suggests that many specific genetic changes are associated with carcinogenesis. At the same time, there is increasing evidence that cell proliferation kinetics are important in carcinogenesis. In 1971, Knudson proposed his now celebrated two-mutation model for embryonal tumors. In fact, a two-mutation model that explicitly considers cell kinetics is the most parsimonious model consistent with the incidence of both childhood and adult cancer in human populations. This model has been known to be consistent, as well, with other epidemiologic and experimental data. But can such a model be reconciled with the laboratory evidence suggesting that many genetic alterations are required for malignant transformation? In this paper, I examine multistage carcinogenesis from the population perspective. Can cancer incidence data in populations provide some insight into the number and nature of stages involved in malignant transformation? I focus attention on carcinoma of the colon because of the considerable amount of information that is now available on the genetics of these tumors. Also, as is the case with retinoblastoma, colon cancer occurs in sporadic and dominantly inherited forms. A comparison of the incidence of these two forms of colon cancer suggests that mutation at the familial adenomatous polyposis (FAP) locus is not necessary for colon cancer, and that inheritance of the gene for FAP is not equivalent to inheriting one of the essential steps on the pathway to colon cancer. Thus, colon cancer does not follow the retinoblastoma paradigm. The incidence of colon cancer in the general population and among polyposis subjects is consistent with two or three mutations on the pathway to malignancy. A three-mutation model is more consistent with the genetic alterations observed in colon carcinogenesis. I present and discuss a working model for colon cancer.