The Polygenic Risk Score in Breast Cancer Risk Assessment

Polygenic Risk Scores

Polygenic risk scores (PRS) are an estimate of an individual’s genetic liability to a complex disease such as breast cancer. The score describes the combined risk of many risk-inducing single nucleotide polymorphisms (SNPs) that make an individual susceptible to breast cancer. SNPs are the most common type of genetic variation, most of which have do not affect an individual’s health, however, numerous low-penetrance SNPs have been found to be associated with complex disease such as breast cancer development. The higher the gene penetrance, the more likely the mutation will result in the development of the disease. Occurring alone, certain SNPs are associated with a slight increase in breast cancer risk, however in combination, they can increase breast cancer risk significantly. Mutations in more well-known high and moderate penetrance breast cancer-risk genes, such as BRCA1/2, TP53, PALB2, ATM, and CHEK2 have been identified in 3-5.7% of breast cancer cases among the general population. SNPs are more common than these, explaining approximately 30% of familial breast cancer risk.

The key aspect when calculating PRS is that it is calibrated to the risk within the assessment population. PRS’ on their own give a snapshot of genetic risk in that population only; to enhance breast cancer risk prediction, this PRS score can be combined with known risk factors (such as mammographic density, family history and age) and incorporated into existing risk models. For calculating the PRS, please see this publication by Professor Evans, Eleanor Roberts, and Sacha Howell from the MCRC team.

Poignant clinical trials

Numerous large-scale studies are investigating the viability of breast cancer PRS’ and how their clinical implementation in breast cancer screening could be facilitated, with one large-scale, pivotal trial already completed.

These include:

• PROCAS 1 & 2, and BC-Predict
• MyPeBS
• WISDOM
• BCAN-RAY

PROCAS 1 & 2, and BC-Predict

The Predicting Risk of Cancer at Screening (PROCAS 1) study was conducted from 2009–2014 for women in the National Health Service National Breast Screening Programme (NHSBSP) and was the largest recruiting research study in the UK, involving 57,900 women over six years. This study provided the women ten-year risk estimates of developing breast cancer and found that risk‑stratified screening utilising the PRS within risk prediction models is likely to identify a larger proportion of women who are high-risk. Importantly, it also reported this risk assessment method as acceptable to patients. Of the 10,000 women whose DNA was used to generate PRS scores, 94% wished to know their score, highlighting the need for a more tailored breast screening programme.

To further understand the effects of a breast screening programme such as that used in PROCAS 1 on the NHS workforce, this programme needed to be implemented in a practical setting to assess its feasibility. This would show if this screening method could be accomplished in real-time and whether risk feedback could be provided in a meaningful way within a critical time window of approximately 8 weeks.

“We identified ~20% of the population who are at moderate- or high risk of developing breast cancer. The previous risk model only picked up ~3.7% of those at risk” – Professor Gareth Evans

Subsequently, in 2017, the PROCAS 2 study commenced, aiming to assess the clinical feasibility of offering a breast cancer screening programme similar to that from PROCAS 1 which was inclusive of calculating a PRS at a woman’s first breast screening appointment. This study aimed to determine whether this screening method was practical in the NHSBSP and show whether risk could be effectively communicated to the recipient within the specified crucial timeframe.

An automated system called BC-Predict was developed which collected standard risk factor information from self-reported questionnaires and mammographic density (MD) calculated from mammograms; and a PRS from a saliva sample. BC-Predict produced real-time (within 6–8 weeks) feedback letters to invite high- or moderate‑risk women to follow up appointments to discuss prevention and early detection. This demonstrated the feasibility of real-time breast cancer risk feedback and showed that women who were identified as high-risk were amenable and willing to take preventive medication with prescriptions written for 77.5% of women at high‑risk.

MyPeBS

My Personalised Breast Screening (MyPeBS) study is currently in the recruitment phase evaluating risk-stratified screening programmes where women are undergoing either the country’s current national breast cancer screening programme or a personalised strategy, which screens women at high or moderate risk of breast cancer more often (yearly or biennially, respectively), and women at a lower risk of breast cancer less often (every four years). The study aims to assess whether a personalised approach is equally or more acceptable than the current country-specific standard mammography programme. This study is similar in size to PROCAS, spanning six countries: Belgium, France, Israel, Italy, United Kingdom and Spain, with the DNA of every participant being collected.

Researchers hope to gain insight into whether personalising screening programmes based on individual risk is better at detecting stage II or higher breast cancers. The study will also evaluate whether four-yearly screening is acceptable in those who are designated as low risk based on their PRS, mammographic density and all the standard risk factors, resulting in economic benefits and fewer unnecessary screenings.

BCAN-RAY

The currently recruiting Breast Cancer Risk Assessment in Younger women (BCAN-RAY) study offers risk assessment consisting of a breast cancer risk factor questionnaire, low-dose mammogram and DNA for PRS to women between 30–39 and who do not have a family history of breast cancer as this population would not otherwise be eligible for screening. Professor Evans noted that if a woman has a 3% risk of developing breast cancer by the time the woman turns 40 years of age, they will receive annual screening throughout the trial. Therefore, rather than ~2% of the population receiving extra screening due to family history, this could increase to 20% of the population between 40–50 years old. One of the main aims of this study is to work out how strong a risk factor breast density is in younger women as it is currently unknown for this age group.

Benefits of PRS for women and healthcare systems

The implementation of PRS’ into an official risk-stratified breast cancer screening programme will allow women to receive a more nuanced, intelligent screening based on their individual risk. High-risk women have the choice to receive preventative treatments and regular screening. Preventative treatment will not only significantly decrease breast cancer risk but decrease the proportion of women who develop late-stage breast cancer that requires more intensive treatment, improving quality of life and healthcare burden. It is also hoped that low-risk women will be able to delay screening and receive less frequent screening, further economically benefitting healthcare systems.

Health inequalities and polygenic risk scores

The greatest limitation with the current use of the PRS is its suitability across different ethnicities. The Genome Wide Association Studies (GWAS) used to inform PRS’ have been conducted on primarily White, European women. The genetic risks identified in these populations cannot be transferred to other ethnicities and doing so increases the risk of overprediction. A version of the PRS for Asian women was recently published, highlighting advances in this area. However, the risk of overprediction is largest amongst Black women, with studies estimating the average breast cancer overprediction in Black Asian and minority ethnic (BAME) populations to be 40%, increasing to as high as 90% for Black women specifically.

“An almost entirely new polygenic risk score is needed for the Black community” – Professor Gareth Evans

The biggest challenge in implementing a recalculated PRS for the Black community is recruiting a large enough cohort from various populations across the world.

Currently, a validated risk prediction model called the Black Women’s Health Study (BWHS) model using data from the BWHS, an ongoing US-based prospective cohort study of 59,000 Black women, was developed with discriminatory accuracy that is on par with measures of discriminatory accuracy for similar risk factor-based models in predominantly White populations. The creators of this model hope that when it becomes feasible to incorporate a PRS into routine healthcare, the improvements to risk prediction may mitigate inequities in breast cancer outcomes for Black women.

Not only do future GWAS need to be more inclusive of BAME groups, but the accurate recalculation of the PRS to curate ethnicity-specific PRS’ is paramount in avoiding disenfranchising a small proportion of the population who already experience health inequalities and furthering the disparity in breast cancer risk assessment. A larger focus on BAME populations, particularly Black women of different origins will broaden the use of the already promising PRS, enabling wider accessibility when implemented into healthcare systems.

This article was written by MSc Science Communication Student Lily Hamilton as part of a project placement at Manchester Cancer Research Centre.