Lessons Learned from Genetic and Epidemiologic Studies of Cancer

Epidemiologists have developed tools to assess risk for future breast cancer in people without apparent disease.  The most widely-used tool is the Breast Cancer Risk Assessment Tool (BCRAT, or informally, the Gail model), which is based on family history of breast cancer, age at menarche, year of first birth and results of previous breast biopsies.  Researchers can now weigh the value of lessons from genome-wide association technology, which scans the genome and establishes the relationship between common inherited genetic variation and risk of disease, in making decisions about patient care.  The criteria for use of tests involving individual genetic data should be the same as for any clinical or preventive test: Does the additional information improve stratification on the basis of future risk of preventable disease?  Stratification should allow women at greatest risk, and, therefore, with the most potential for benefit, to receive a more aggressive intervention than the standard intervention; correspondingly, women at lowest risk can safely receive a less aggressive intervention than standard.  An effective program based any test should give better overall results than a simpler program that assigns standard intervention to everyone. These results should incorporate disease averted or mitigated; all costs, including financial and from adverse events; and positive or negative effects on other diseases.  Risk stratification that does not meaningfully change clinical management has no utility.