Refining Personalized Breast Cancer Risk Assessment

Polygenic risk scores (PRS), which can help determine breast cancer risk in patients, have primarily been developed and validated for populations of European descent. To make a PRS more accurate, researchers developed and validated a novel global PRS that utilizes individual ancestral genetic composition. The data were presented at the virtual American Society of Clinical Oncology annual meeting.

In this exclusive MedPage Today video, study co-author Holly Pederson, MD, director of medical breast services at Cleveland Clinic, explains the study and its implications on reducing racial disparities in breast cancer risk estimation.

Following is a transcript of her remarks:

In the way of background, multigene panel testing, which came about in the fall of 2013, uses next-generation sequencing to rapidly sequence multiple genes simultaneously at a lower cost. And so we’re able to sequence highly penetrant gene mutations like BRCA1, BRCA2, P10, P53, PALB2, and CDH1, and more moderately penetrant genes like CHECK2 and ATM. And that’s what’s tested for when you have multigene panel testing for breast cancer — hereditary susceptibility.

But there are over 300 SNPs, or single nucleotide polymorphisms, that are very common variations that are seen in more than 1% of the population that, in and of themselves, don’t influence risks significantly, but in aggregate may actually have a significant effect on risk, both in women who carry genetic mutations and those who don’t.

And so it’s felt in 2021 that we’ve probably identified most, if not all, of the highly penetrant and moderately penetrant genes, and the remaining mystery of heritability is probably largely explainable by this genomic combination of single nucleotide polymorphisms.

So the challenges that we faced were that current polygenic risk scores are very accurate for women of European descent, but when you look at women of non-European descent, the calibration was such that there was an overestimation of risk for all non-European women essentially. And in fact, in Black women there was almost a two-fold overestimation of risk, despite the fact that rates of breast cancer in Black women are similar to, if not slightly lower than, white women. But the polygenic risk score was not calibrated appropriately because it was developed in European women.

And so what we actually were able to develop is a framework for a polygenic risk score that delivers the personalized genomic breast cancer assessment to any and all interested U.S. women by looking at their genomics. So how was this created? We really had to create an entirely new methodology in terms of assessing the statistical risk of women. And the ancestry was not determined by self-reported ancestry, but actually by genomic or genetic ancestry.

And the SNPs that were ancestry specific helped to stratify women according to the percentage of ancestral weight that they inherited from each of three primary continents — Europe, Southeast Asia, and Africa. And it’s felt that American women can essentially be stratified according to their inheritance from those three continents.

So it’s very exciting because at the end of it all, the polygenic risk scores for all ancestries were precisely calibrated, such that it was accurate for non-European populations. Clinically, the polygenic risk score has been in its infancy and in part due to concerns over the lack of applicability to non-European populations. In fact, current NCCN [National Comprehensive Cancer Network] guidelines advise against the use of current polygenic risk scores outside of the context of a clinical trial because of significant limitations in interpretation.

But it’s felt that knowledge from the PRS in a given woman may help to explain up to 20% of breast cancers that are unexplainable by the highly penetrant and moderately penetrant genes and may help newly diagnosed breast cancer patients make decisions about contralateral mastectomy because it may help predict the risk on the other side. And for the first time we may be able to identify low-risk women. Up until now it’s either [been] average risk or high risk. And lower-risk women may want to know that, but may also opt for different risk-based screening based on their risk level as may high-risk women.

And so, you know, where are we headed next? The next steps will be to combine this polygenic risk score with traditional risk estimation models that incorporate clinical and personal and family history to really sub-stratify, sort of a mathematical model genomically, to make it more relevant to an individual woman.

We will refine this model as more data becomes available in non-European populations. The ability to discriminate between high- and low-risk women of course was greatest in European women because we have the most data, but we have great data points that were well-validated to use in other populations, and the discrimination or the ability to distinguish high risk from low risk was seen across the board.

So this is really an important step in personalized medicine and personalized risk stratification and will help women to make more informed decisions about their risk and risk management — whether to opt for MRI screening, for instance, or whether to take preventive medication. And it’s a step in the right direction to reducing racial disparity in breast cancer risk estimation.