Diet and cancer.

Most of the common cancers in Canada have been associated in some way with diet. More than half of all newly diagnosed cancers and cancer-related deaths are limited to three sites in each sex: lung, prostate, and colorectum in men and breast, colorectum, and lung in women. The earliest and strongest epidemiologic evidence associating diet and cancer has come from descriptive correlation studies. The findings of subsequent, more powerful, analytic studies have been inconsistent. The author reviews the importance of diet for common cancers in Canada and considers the methodologic limitations of various study designs.