Host‐tumor Antagonism. Xxxii. The Epidemiology of Cancer*

Abstract

Epidemiology is the science which deals with the pattern of disease occurrence in such a way that when the data are examined by statistical methods and logic, they may point to a working hypothesis regarding etiology. For many years epidemiology was concerned chiefly with infectious diseases, but in more recent times the techniques have been applied in many other fields, often with gratifying results. For example, the statistical association of certain cancers of the skin with occupations involving exposure to tar and crude oils led to preventive measures. Another example is the association of bladder cancer with work in the analine dye industries.Wynder's (1) criteria for identifying an environmental agent as a cause of cancer are:1. The relative risk of cancer should rise in proportion to the degree of exposure.2. The incidence of cancer among specific population groups should be consistent with the distribution of the agent.3. Withdrawal, reduction or modification of the agent acting on a population group followed by a decrease in the incidence of a given cancer, after a suitable latent period.4. The agent should be shown to be carcinogenic to some animal species. Various other methods have been used in the epidemiologic study of cancer, e.g., investigations of the social histories of hospital patients with cancer of various organs in comparison with control groups, statistical studies of correlations between cancer mortality in local populations and factors in the environment in which they live, and "prospective" statistical studies of large groups of people with one or more known habits to determine their eventual causes of death (2).Steiner (3) introduced another method, in which he studied at autopsy a large population with cancer of various types, and compared the frequencies of these cancers with those in the countries from which the deceased patients' ancestors had emigrated in the recent or distant past. Although this brilliant work will be reviewed in some detail later, his findings are briefly summarized here. Steiner reported a racial and geographic study of 6,072 cancer autopsies from a total of 35,293 autopsies in the Los Angeles County Hospital. His conclusions were that the etiologic factors in cancer development are unknomm except in some cases of * Parts I-XXXI of this study have been published in the Jozonal of the American Geriatt Attending Physician, Swedish Hospital of Brooklyn.