The Spatial Organization of Cancer Mortality.

The examination of human cancer occurrence from a spatial perspective can yield useful information that aids in the search for factors that influence cancer risk. The spatial implications of theoretical models of carcinogenesis suggest an approach for empirically testing theories relating to this problem. The spatial analysis of empirical cancer data can yield clues leading to the development and testing of causal hypotheses. Different models of the relation- ship between cancer risk and environmental factors yield different expected spatial patterns for two forms of skin cancer, malignant melanoma and other skin cancers. The empirical patterns are in basic agreement with those expected. Both lung and skin cancers exhibit complex pat- terns of spatial variation among U.S. counties. One-dimensional spatial analysis is used in this study to decompose these patterns by spatial scale in order to assess the nature of possible environmental risk-affecting agents. Large-area variation (or regional trend) is important in both lung and skin cancer patterns, but small-area patterning (or neighborhood effects) are important only for lung cancer. Significant differences between the patterns for males and females exist; for example, female lung cancer rates seem to be influenced by nearby urban places whereas male rates are strongly influenced only by larger urban places. [ABSTRACT FROM AUTHOR]