Aggregation of colon cancer in family data.

Although only a small proportion of common cancers show familial aggregation, studying such families can elucidate the roles of shared environment and genes in the development of neoplasia. We report an analysis of nine colon cancer pedigrees using new nonparametric objective methods to measure familial aggregation as a means of determining the existence of heterogeneity in the data. Each family was selected through a proband with nonpolyposis colon cancer who had a first-degree relative with documented colon cancer. To assess the aggregation of different cancers in these families we employ a method which evaluates both excess number of cases as well as distribution by risk in family members. We find that eight of the nine families exhibit significant aggregation of colon cancer: endometrial cancer aggregates in three families, breast in none, kidney in one, and all sites in eight. In this way, we show that two families fit the criteria for Cancer Family Syndrome, and that one is not a high-risk cancer family.