The research focussed on just seven common breast-cancer susceptibility alleles that have been well validated through genome-wide association studies
(excluding rare genes that are responsible for familial breast cancer risk, such as BRCA1/2
). Although the risk conferred by each individual locus
is small (relative
risk of less than 1.3), by modelling the combined effect of multiple risk alleles using a simple multiplicative model, and converting that information into an absolute
risk over a specified time period, a woman’s overall risk of breast cancer can vary approximately sixfold.
The results suggest that risk profiling based on genetic susceptibility is not clinically useful at an individual level, as it does not provide sufficient discrimination to warrant personalised prevention. However, it may be valuable for stratifying the population in order to target screening programmes more effectively; the authors state that “the efficiency of population-based preventative programs such as mammography could be improved by targeting women who are at the greatest risk for breast cancer according to genotype”. For example, the NHS currently offers screening to all women at the age of 50, when their 10-year risk of breast cancer is around 2.3%. If genetic screening were used to stratify this population, around 20% of women would be classified as low risk and never reach this threshold value (due to the effects of competing mortality), and therefore should not be compelled to go for screening, whilst the top 5% of women at highest risk would reach the threshold value at only 41 years of age.
The paper is accompanied by an editorial [Hunter DJ, et al. (2008) NEJM 358(22): 2760-2863
], which provides an overview of the process of finding new gene-disease associations through systematic genome-wide searches. It highlights the difficulties and limitations of the technique, and the complexities of understanding the results in a clinical situation, suggesting that “patients should be wary of companies that seek to sell such information”
prior to further elaboration, interpretation and validation.
Comment: This paper is the first of its kind to try to apply individual genetic risk calculations to population screening, and is particularly unusual in its consideration of the implications for both individualised disease prevention and public health policy. As such, it provides a major step forward in thinking as regards how to use information about genetic risk factors in standard health practice. Nonetheless, as the editorial highlights, the greatest impact of these discoveries may still lie in an improved understanding of the pathology of disease, and the ultimate translational challenge remains to develop this plethora of data into detailed mechanistic understanding, clinically useful risk prediction and novel therapeutic strategies.