Comments on the epidemiology of large bowel cancer.

Available evidence strongly suggests that there are at least two epidemiological subentities of large bowel cancer. Tumors of the intermediate portions display a socioeconomic gradient not observed for cancer of the cecum or lower rectum. Dietary factors are the most plausible etiological candidates. High-fat intake may be a prominent factor, but other dietary constituents could be relevant and may be studied in populations in which stomach and large bowel cancer are infrequent. The eqidemiological usefulness of precursor lesions is emphasized.