Evolution of cancer epidemiology.

As the 20th century passes into history, it seems timely to reflect upon current directions and progress in epidemiologic studies of cancer. The most important legacy of cancer epidemiology to date is the general recognition that an array of exogenous exposures are responsible for most cancer occurrence (1). Evidence supporting this conclusion includes: 1) the notable variation in cancer incidence internationally, 2) migrants' and/or their descendants' frequent development of cancer rates characteristic of the new area of residence, and 3) etiologic studies demonstrating a substantial fraction of cancer arising from exogenous exposures. An important watershed was a 1981 article by Doll and Peto estimating the proportion of US cancer deaths due to major categories of exposures (1). The two most important categories were nutritional factors (estimated to be responsible for approximately 35 percent of cancer occurrence) and tobacco use (30 percent); others included reproductive factors/sexual behavior (7 percent); occupation (4 percent); alcohol drinking (3 percent); geophysical factors, including ionizing radiation and ultraviolet radiation from sunlight (3 percent); pollution (2 percent); iatrogenic exposures (1 percent); food additives (<1 percent); industrial products (<1 percent); and other and unknown factors. In an ongoing debate, some epidemiologists have questioned the entire premise of these estimates; others proclaim their continuing accuracy, while a few suggest rigorous reevaluation at periodic intervals to direct cancer epidemiologic research priorities. Differences as well as similarities between cancer epidemiology and other areas of epidemiology have changed over time. Prior to the second half of the 20th