NEW YORK (GenomeWeb News) – In a study appearing online today in the Journal of the American Medical Association, an international research team outlined their efforts to assess relationships between more than a dozen low penetrance SNPs and breast cancer risk — looking at how individual variants and polygenic risk models corresponded to breast cancer risk and sub-type.
Using SNP data on more than 10,3000 women with breast cancer and about as many healthy controls, researchers from the University of Oxford and the French National Genotyping Center evaluated risk associated with 14 breast cancer risk variants — alone or in combination — for half a dozen breast cancer sub-types.
Their findings suggest polygenic risk scores are somewhat more predictive for estrogen receptor positive breast cancers, though several individual SNPs also corresponded to greater risk of developing specific breast cancer sub-types.
"The polygenic risk score was substantially more predictive of [estrogen receptor]-positive than of [estrogen receptor]-negative breast cancer, particularly for absolute risk," senior author Gillian Reeves, a cancer epidemiology researcher at the University of Oxford, and her co-authors wrote.
For the project, collaborators at the French National Genotyping Center used the Applied Biosystems Taqman array to genotype DNA from 10,306 women with breast cancer and 10,393 women who do not have the disease.
Participating individuals had been enrolled through a prospective British and Scottish research project called the Million Women Study, and researchers selected the 14 SNPs based on information gleaned from published GWAS, reviews, and meta-analyses.
Overall, the team found that half of the SNPs tested were significantly and independently linked to breast cancer risk.
After tossing out two of the three SNPs in linkage disequilibrium in the gene TNRC9, the researchers then tested the dozen remaining SNPs for associations with tumor characteristics such as invasiveness, estrogen receptor status, histology, and more.
That analysis suggested 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, though the information provided by each also varied by cancer type.
For instance, their results indicate that the per allele odds ratios for variants in FGFR2 and TNRC9 were higher for both estrogen receptor positive cancers and lower grade tumors.
On the other hand, the per allele odds ratio chromosome 2 variant, known as rs13387042, was highest for breast cancer cases in which both breasts were affected with lobular rather than ductal breast tumors. Three other SNPs were more tenuously linked to other tumor traits.
After verifying these patterns using a meta-analysis of published study data, 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.
From these models, they concluded that polygenic risk scores could offer some insights into breast cancer risk, but were most informative for estimating risk of estrogen receptor positive cancers.
Women under 70 years old with the highest polygenic risk scores have an estimated breast cancer risk of 8.8 percent compared with a risk of 4.4 percent in women with the lowest polygenic scores, the researchers reported.
When the same models are applied to estrogen receptor positive cancers, women with the highest risk score have an estimated risk of 7.4 percent of estrogen receptor positive cancer, compared to 3.4 percent risk in women in the lowest polygenic risk score group.
Even so, the team noted, the predictive power of these models does not yet seem to be powerful enough to apply them for breast cancer screening — consistent with past studies suggesting genetic information adds relatively little risk information to existing breast cancer risk models.
"[A]s others have suggested, sub-dividing women on the basis of such polygenic risk scores is not at this stage a useful tool for advising women about risk of for population-based breast cancer screening programs," the researchers concluded, "but may ultimately be useful for understanding disease mechanisms."