Application of risk prediction to breast cancer prevention


Session type:

Douglas Easton1
1University of Cambridge, Cambridge, UK


There is growing interest in the idea that prevention strategies for breast cancer might be targeted on women at high risk. Until very recently, the main targeting was identification of carriers of BRCA1 and BRCA2 mutations in women with a strong family history, and use of prophylactic surgery and/or MRI screening. However, NICE guidelines for management of familial breast cancer have recently been updated to include prevention by tamoxifen/raloxifene in high risk women. In addition, there is increasing recognition that mammographic screening for breast cancer is only moderately effective, and might be more effective if targeted at women at increased risk.

In parallel, the ability to discriminate risk has improved markedly. Genome-wide SNP studies have identified close to 100 common genetic variants associated with breast cancer risk. While the risks associated with the common low-penetrance alleles are modest, these risks combine multiplicatively with each other and with BRCA1 and BRCA2, and the combined effects of the currently identified loci are sufficient to identify individuals at substantially higher and lower cancer risk. For example, 1% of the population have a risk that is more than 3-fold that of the general population, while 1% of the population have a risk that is less than one-third of the population risk. In addition, next generation sequencing has provided the potential for ‘panel' testing of a larger set of breast cancer associated genes.

An effective targeted prevention programme will depend critically on being able to model combined effects of known genetic markers with mammographic density, family history and other important risk factors. We are developing our online risk prediction tool BOADICEA to provide such predictions. Equally, it will depend on the acceptability of risk-based management to individual women and to the population at large.