The present study was designed to examine the epidemiological association of food consumption with cancer mortality in Japan, if any, in comparison with similar examinations in developed countries including Japan, in a search for dietary factors in cancer etiology. Whereas food consumption patterns in Japan was believed to be specific among developed countries, this study revealed that the correlation matrix among major foods both for international and Japanese data made it possible to categorize these foods into two groups with some similar patterns, since there should be observed a positive or negative correlation for each pair of foods. The patterns for correlation of food consumption to cancer mortality in Japan were found to be considerably different from those in developed countries, and this suggested that the epidemiological association between the two in Japan, if any, could be specifically evaluated from that in develop countries. The urban-type region for food consumption in Japan, however, indicated a possible westernization for the relation of food consumption to cancer mortality as well as for food consumption patterns, whereas the traditional Japanese region for food consumption did not indicate any similarity with westernization for the relationship between food consumption and cancer mortality, as well as for food consumption patterns. The above findings also suggested that there might be a possible significance for the regionalization of Japan in such studies, and that the urbanization of food consumption patterns in Japan would show that this country has a similar pattern of association between food consumption and cancer mortality as observed among developed countries in Europe and/or North America.