Breast cancer is now the most common cancer in the UK (see CRUK website). However, thanks to extensive research and improved management, survival rates for breast cancer have been improving for thirty years. Ongoing research is aiming to identify ways of reducing the incidence of breast cancer by improved prevention, as well as to understand the complex biological disease pathways in order to develop new treatments. Sporadic breast cancer is a complex disease associated with both genetic and environmental risk factors. There are now about 20 genetic variants known to be linked to disease susceptibility and more are being identified. Whilst the risk conferred by each is typically small, in combination there may be potential to use these markers to improve risk prediction.
A new study evaluates risk associated with 14 breast cancer risk variants (SNPs), alone or in combination, for half a dozen cancer subtypes [Reeves et al. (2010) JAMA 304(4):426-434], to look at how individual variants and polygenetic risk models correspond to breast cancer risk and subtype. This large prospective study used more than ten thousand women with breast cancer and about as many healthy controls without breast cancer. The scientists used meta-analysis of results from the study as well as of other studies. The analyses of the results suggested that SNPs in the FGFR1 and TNRC9 genes, as well as a third SNP on chromosome 2, were most closely tied to overall breast cancer risk, and that risk prediction was most reliable for oestrogen receptor (ER) positive cancers and lower grade tumours.
The researchers also developed polygenic risk models using combined data on four, seven, or 10 of the SNPs that were most strongly associated with breast cancer. They concluded that women under 70 years of age with the highest polygenic risk scores have an estimated breast cancer risk of 8.8% compared with a risk of 4.4% in women with the lowest polygenic scores. They also found that the polygenic risk score was substantially more predictive for ER-positive cancers (ranging from a high of 7.4% to a low of 3.4%) than ER-negative breast cancers (a range of just 1.4% to 1.0%).
This is an important study which evaluates the predictive value of selected genetic markers identified by different studies and finds that it varies for different tumour subtypes. However, the authors caution that whilst their findings are potentially useful for understanding disease mechanisms, they would not be useful for individual breast cancer risk prediction or population stratification as known risk factors for breast cancer such as family history are more predictive. Nevertheless, this is an interesting line of enquiry and it may be that ultimately the combination of established and novel environmental and genetic risk factors could refine risk prediction for targeted population screening.