This chapter examines spatio-temporal changes in breast cancer clustering in the Northeastern United States to assess the statistical significance of clusters using GIS-based kernel methods. It first describes higher-than-average breast cancer mortality rates in the Northeast and introduces statistical methods for detecting geographic clusters of disease. A GIS-based kernel method based upon the theory of Gaussian random fields is applied to the breast cancer mortality data taken from the National Center for Health Statistics’ Compressed Mortality File. The method makes use of a map of rates, smoothed using a Gaussian kernel. The maximum smoothed value is compared with the statistic’s critical value to identify significant clusters. Results from the analyses show changes in spatio-temporal clustering patterns in the Northeast during the period 1968-1998. The results reveal not only the existence of statistically significant breast cancer clusters, but also the changing patterns of those clusters over time. Since environmental risk factors may play an important role in explaining the unknown etiology of breast cancer, analyses of spatio-temporal changes of breast cancer clustering may provide important clues to the study of breast cancer and environment relationships.