Risk Predictor Works for ER Positive Breast CA

(MedPage Today) -- A risk score based on a number of uncommon genetic variants appears to predict risk of estrogen receptor-positive breast cancer but with questionable clinical utility, researchers said.

Women in the top 20% for the risk score had double the overall cumulative incidence of breast cancer to age 70 as did those in the bottom 20% at an estimated 8.8% versus 4.4%, according to Gillian K. Reeves, PhD, of the University of Oxford, England, and colleagues.

The predictive power appeared limited to estrogen receptor-positive tumors, with a cumulative incidence estimated at 7.4% versus 3.4% for the top and bottom quintiles, they reported in the July 28 issue of the Journal of the American Medical Association.

For estrogen receptor-negative breast cancer, the genes predicted little difference in cumulative incidence between the highest and lowest risk score quintiles (1.4% versus 1.0%, respectively).

A high genetic risk score appeared as predictive as having a first-degree relative with breast cancer for women in the developed world, the researchers noted.

But the expense of genetic testing may not be warranted clinically, Reeves' group implied.

Other established risk factors have greater effects on breast cancer incidence than the differences seen in the study, they noted. Nor were there interactions between those risk factors and the genes in their analyses.

"Subdividing women on the basis of such polygenic risk scores is not at this stage a useful tool for advising women about risk or for population-based breast cancer screening programs but may ultimately be useful for understanding disease mechanisms," the researchers cautioned in the paper.

These findings follow on the heels of questions raised about genetic prediction models for breast cancer in one of the most promising clinical settings -- predicting breast cancer response to neoadjuvant chemotherapy.

Journal editors issued an "expression of concern" to warn about potential problems with the validity of a 2007 Lancet Oncology paper reporting positive results with a gene signature to predict neoadjuvant chemotherapy response that is now being used in three ongoing clinical trials.

Genome-wide association studies have detected a number of gene variants of interest in breast cancer with either definite or probable links to risk.

Reeves' group gathered 14 of these single-nucleotide polymorphisms (SNPs) for analysis in the prospective Million Women Study.

Genotyping of blood samples from 10,306 breast cancer cases and 10,393 controls in the British study revealed the highest odds of breast cancer with two SNPs, FGFR2-rs2981582 and TNRC9-rs3803662.

Both had their strongest predictive power with regard to estrogen receptor-positive disease in the case-control analysis and when meta-analyzed with all published data.

For estrogen receptor-positive cancer, the pooled odds ratio was 1.30 (95% confidence interval 1.26 to 1.33) for each additional FGFR2 allele and and 1.24 (95% CI 1.21 to 1.28) for each additional TNRC9 allele.

For estrogen receptor-negative cancer, the pooled odds ratios were lower at 1.05 (95% CI 1.01 to 1.10) and 1.12 (95% CI 1.07 to 1.17), respectively. Both showed an interaction by cancer type (P<0.001).

The next strongest association was for 2q-rs13387042, which appeared to predict bilateral breast cancer (per-allele OR 1.39 versus 1.15 for unilateral, P=0.008 for interaction).

It also more strongly predicted lobular breast cancer compared with ductal tumors (per-allele OR 1.35 versus 1.10, P<0.001 for interaction).

Three other SNPs showed at least some evidence for an interaction with at least one tumor characteristic examined in the study.

Overall breast cancer risk appeared strongly and significantly associated with seven SNPs, which were combined into the risk prediction score. But the number of SNPs included in this risk model didn't appear to materially alter the findings.

The researchers warned that their study had limited power to detect breast cancer risk associated with relatively rare high-risk alleles according to some of the tumor characteristics despite the large study size.

They also urged caution in interpretation associations that weren't highly statistically significant given the large number of tests.

The Million Women Study is supported by the U.K. Medical Research Council and Cancer Research U.K. Genotyping of blood samples was supported by the Institut National du Cancer, France.

The researchers reported that they had no financial conflicts of interest to disclose.

Primary source: Journal of the American Medical Association Source reference: Reeves GK, et al "Incidence of breast cancer and its subtypes in relation to individual and multiple low-penetrance genetic susceptibility loci" JAMA 2010; 304: 426-434.