Epidemiologic and genetic follow-up study of 544 Minnesota breast cancer families: design and methods

In 1944, a case-control family study was initiated at the Dight Institute for Human Genetics at the University of Minnesota to study the influences of childbearing breastfeeding, and hereditary susceptibility on the occurrence and age-of-onset of breast cancer. Index cases (probands) were women ascertained at the Tumor Clinic of the University of Minnesota Hospital. Medical history and life style information were obtained on probands and relatives, and all cancers were histologically verified. A total of 544 families were studied, with probands diagnosed between 1931 and 1952. All of the records and pathology slides have been maintained from the original study; for most probands this includes the original tissue blocks. We are conducting a historical cohort study of selected of selected first- and second-degree female relatives (sisters, daughters, nieces, granddaughters) of the probands and a group of control women identified as the spouses of all male first- and second-degree relatives (brothers, sons, grandsons, and nephews). The subsequent development of breast cancer is being determined to quantify the absolute risk associated with a positive family history. Current disease status is ascertained with mammography, and stromal density is measured using digital imaging. Segregation analysis will be applied to examine how non-genetic factors such as diet, exogenous hormone use, and body fat distribution influence risk in women at high risk because of family history. A subset of families are being selected for molecular analysis of the BRCA1 gene or for linkage analyses to identify putative susceptibility loci other than BRCA1. Documented cancer histories were known for at least three generations, and the current study extends the pedigrees up to four or five generations for every family, allowing a detailed description of familial risk. This cohort study of breast cancer families is likely to be important in both quantity and quality of data and will serve as a major genetic epidemiologic resource, being free of selection bias and having relevant non-genetic exposure determined in at least four generations.