Breast Cancer Polygenic Risk: What a Common-Variant Score Can and Cannot Tell You

Breast Cancer Polygenic Risk: What a Common-Variant Score Can and Cannot Tell You

This article is for educational purposes only. It is not a substitute for advice from a licensed physician, board-certified medical geneticist, or board-certified genetic counselor. For any decisions about testing, treatment, or care, consult a qualified clinician. In emergencies, call 911.

Bottom line: Harmful BRCA1 and BRCA2 changes are rare, found in roughly 1 in 400 people. Most breast cancer happens in people without one. In them, the genetic part of risk is spread thinly across thousands of common DNA spots. A 2007 study in Nature found the first of these spots, near the genes FGFR2 and TOX3. A 2019 study packed 313 of them into one number called PRS313. That number shifts your odds. It does not decide your future, and no clinical guideline yet tells doctors to change your care because of it.

What you’ll learn

• What a polygenic risk score is. It is a single number that adds up thousands of tiny genetic effects. It is not a test for one broken gene.
• What “higher risk” means in plain numbers. Every figure here is anchored to about 13 out of 100 US women.
• Why the score is less reliable when your ancestry is not mostly European.
• What changes your mammogram schedule today, and what does not.

Polygenic Means Many Small Nudges, Not One Broken Gene

Most readers arrive with a single gene picture in mind. One gene breaks, risk jumps. That is the wrong picture here.

The variants behind a polygenic score are common. Large fractions of the population carry them. Each one nudges risk by a small amount. None of them is “the breast cancer gene.”

Think of a credit score. No single late payment defines it. Hundreds of small entries add up to one number. A polygenic risk score works the same way.

Where the story started

A 2007 genome wide study in Nature ran in two stages. Stage one compared 4,398 breast cancer cases with 4,316 controls. A later stage tested 30 DNA spots in 21,860 cases and 22,578 controls from 22 studies.

Five new regions showed a strong and steady link. Four of them sit next to likely genes: FGFR2, TNRC9 (now called TOX3), MAP3K1 and LSP1. In standard genome map labels, FGFR2 sits at 10q26 and TOX3 sits at 16q12.

The authors framed the gap plainly. Known susceptibility genes, they wrote, account for less than 25% of the familial risk of breast cancer.

The list got much longer

A 2017 analysis by the Breast Cancer Association Consortium was far larger. It compared 122,977 cases with 105,974 controls of European ancestry. It added 14,068 cases and 13,104 controls of East Asian ancestry.

That single paper reported 65 new risk regions. It also explained where the risk variants sit. Most of them fall in distal regulatory elements, the authors reported. In plain terms, they sit in switch DNA near a gene. They usually do not break the protein itself.

• FGFR2 makes a growth-signal receptor on the cell surface.
• TOX3 makes a protein that binds DNA and helps control other genes.
• The risk variants near them mostly change how loudly those genes are switched on.

What this pattern is not

This is a risk pattern, not a family disease.

• It is not autosomal dominant inheritance, where one copy of a single faulty gene from either parent is enough to raise risk sharply.
• There is no 50% coin flip for each child.
• There is no carrier status. A carrier is a person who carries one specific disease-linked variant. Common variants are carried by huge slices of the population, so the word does no useful work here.

A polygenic result is not a diagnosis of a hereditary cancer syndrome. Only a clinician can judge whether your family pattern calls for a different kind of test. Bring the question to your primary care doctor.

Section recap: Polygenic risk comes from thousands of common variants, each with a tiny effect, mostly sitting in regulatory DNA near genes such as FGFR2 and TOX3. It does not travel through a family like a single broken gene.

Why This Is Not BRCA1 or BRCA2

This difference changes what you should do next. The two kinds of genetic finding are not two grades of one thing. They are different objects.

Key numbers from the National Cancer Institute:

• Harmful BRCA1 or BRCA2 changes appear in about 0.2% to 0.3% of people.
• That is roughly 1 in 400.
• More than 60% of women who inherit one develop breast cancer in their lifetime. That is more than 60 out of every 100.
• About 13% of women in the general population develop breast cancer. That is about 13 out of every 100.

That third figure is a statement about penetrance. Penetrance means the share of people with a variant who go on to develop the condition. High penetrance variants are rare and strong. Common variants are frequent and weak.

	Rare BRCA1/BRCA2 change	Common polygenic variants
How common	About 1 in 400 people	Carried by large fractions of the population
Effect of one variant	Large	Very small
Penetrance	More than 60 out of 100 by end of life	A modest shift around the 13-in-100 baseline
Clinical path	A set high risk care plan	No path driven by score alone

Why a BRCA-negative relative changes nothing here

Say your mother tested negative on an oncology-ordered BRCA panel. That result is real and useful. It does not close the question.

The two tests look for different things. A panel hunts for rare, damaging variants in named genes. A polygenic score sums thousands of common ones. A negative panel does not rule out a polygenic effect.

The reverse is also true. A polygenic score is not a search for hereditary cancer syndrome variants. The main US hereditary cancer framework, NCCN Version 2.2026, is built around pathogenic and likely pathogenic variants in named genes, not around scores. An earlier version of that framework named BRCA1, BRCA2, CDH1, PALB2, PTEN and TP53 as genes in scope.

Other genes also sit outside any polygenic score. The BOADICEA risk model treats truncating variants in BRCA1, BRCA2, PALB2, CHEK2 and ATM as separate inputs from the 313-variant score. Panel testing for those genes is a talk to have with a clinician. A consumer score does not answer it.

If the single gene path is what you came for, this site covers it separately. Read BRCA1 and BRCA2: breast and ovarian cancer risk instead.

Section recap: A rare BRCA variant sits in about 1 in 400 people and carries more than 60-in-100 lifetime risk. A polygenic score is a different object built from common variants, and it cannot rule a BRCA variant in or out.

How Much Risk? Turning the Numbers Into People Out of 100

Every number below is anchored to one baseline. The NCI SEER program reports that about 13.0% of women are diagnosed with breast cancer at some point in life, based on 2021-2023 data. That is about 13 out of 100 US women, often stated as about 1 in 8.

More from the same source:

• About 321,910 new cases and 42,140 deaths in the United States in 2026.
• Five-year relative survival is 91.9%, based on 2016-2022 data.

Two words the reports use

Relative risk compares one group’s chance with another group’s chance. It is a ratio, so it tells you nothing on its own. Twice a tiny number is still a tiny number.

An odds ratio is a closely related ratio used in studies that compare people with and without a disease. Both need a baseline before they mean anything to you.

What PRS313 measured

The 2019 study built the score from 94,075 cases and 75,017 controls of European ancestry across 69 studies. It checked the score in 11,428 cases and 18,323 controls from 10 forward looking studies, plus 190,040 women from UK Biobank.

The headline figure is an odds ratio of 1.61 per standard deviation (95% CI 1.57-1.65). A standard deviation is one step along the spread of scores in the population. The area under the curve was 0.630. That figure shows how well a score sorts cases from non-cases.

The extremes are more useful to picture:

• Lifetime risk in the top 1% of the score was 32.6%. That is about 33 out of 100 women. Turn it around: about 67 out of 100 are never diagnosed.
• Now compare the top 1% with women in the middle fifth of the score. The top 1% had 4.37 times the risk of estrogen receptor positive disease. For estrogen receptor negative disease the figure was 2.78 times. Estrogen receptor status describes whether tumor cells grow in response to the hormone estrogen.
• Those multipliers are measured against the middle fifth. They are not measured against the population average. Quoting them any other way overstates them.
• In the lowest 1%, the same comparisons were 0.16 and 0.27. Low scores move the number down as well as up.

What the score leaves out

A polygenic score is one input. The BOADICEA model exists because genetics alone is not enough. That model combines the 313-variant score with rare variants, family history, hormonal and reproductive history, lifestyle factors and breast density.

Two figures from that paper are worth holding onto:

• The 313-variant score explains 20% of the polygenic variance in breast cancer risk.
• With all factors combined in a UK population, predicted lifetime risk ran from 2.8% at the 1st percentile to 30.6% at the 99th. That is about 3 out of 100 up to about 31 out of 100.

Those UK figures come from the full model, not from a score alone, and they are not US numbers. Age, breast density, and family history all sit outside the score itself.

The NHGRI puts the limit bluntly. A polygenic risk score can only explain relative risk, it says, and it does not provide a baseline or a timeframe for how a disease might progress.

One more warning. Different score versions and different reference groups produce different percentiles for the same person. A percentile quoted without its model cannot be read. Ask your clinician to look at the report itself, not at a remembered number.

Section recap: Against a baseline of about 13 out of 100, the top 1% of PRS313 carried about 33 out of 100 lifetime risk. Most people with a high score are never diagnosed.

Where the Score Came From: Ancestry and Portability

This is the caveat most explainers skip, and it is the one that changes how much weight to give your own result.

Almost every published breast cancer score was built in cohorts that are overwhelmingly of European ancestry. The 2007 discovery panel was designed around common variants in Europeans from the start. The 2019 PRS313 development and validation cohorts were European ancestry throughout.

A 2024 review in Nature Reviews Genetics states the consequence directly. Scores developed in populations of predominantly European genetic ancestries can increase health disparities, because they predict poorly in people of diverse and complex ancestries. The review also covers the methods being built to fix this, so the problem is recognized rather than ignored.

The NHGRI says the same thing in plainer words. Most genomic studies so far have looked at people of European ancestry. There may not be enough data from other populations to calculate a score well.

What a 2025 study in South Africa exposed

A 2025 paper in Nature Communications is the first reported breast cancer genome wide study in a resident African population. It compared 2,485 cases with 1,101 population matched controls.

Findings worth knowing:

• Two risk regions reached the significance bar. One sits between UNC13C and RAB27A on chromosome 15 (rs7181788). The other sits inside USP22 on chromosome 17 (rs899342).
• A European ancestry-derived polygenic model explained only 0.79% of the variance in that dataset.
• The new signals did not show up again in West African ancestry data. The authors read that as genetic difference between African populations.
• The authors state their own findings need further checking, and call for larger pan-African studies.

So these are new signals, not settled risk factors. The wider point still lands. Real risk regions exist that current score models do not contain.

The 23andMe page for clinicians lists relevant ethnicities for each polygenic report, so the product itself is scoped by ancestry group.

The practical inference

If your ancestry is not well covered in the reference data, your percentile is shakier than the report claims. That is true for a high result. It is just as true for a low one. A reassuring number is not a reason to skip screening you are otherwise due for.

Ask this at your next appointment. How much should my ancestry change the way you read this? A primary care clinician or a genetic counselor can put the answer in context.

Section recap: Score weights come mostly from European ancestry cohorts, and a European derived model explained only 0.79% of variance in South African Black women. Outside the discovery group, high and low results are both less certain.

Reading the Report: A Risk Estimate, Not a Diagnosis

Most of the worry around these reports comes from mixing up three very different papers.

Result type	What it is	What it supports
Polygenic risk estimate	A position in a population distribution	Context and questions, not action
Pathogenic single gene result	A specific harmful variant found by a clinical lab	Established management pathways
VUS	A variant of uncertain significance, meaning the lab cannot yet say whether it matters	No change in care; may be reclassified later

• Polygenic scores do not produce VUS findings. If you hold both a panel report and a polygenic report, you are reading two different kinds of document.
• Consumer arrays read a fixed set of DNA spots. That is not the sequencing used in clinical panels.
• A score is population-relative. A pathogenic variant finding is about you.

What the companies and the agencies say

23andMe’s own clinician page states that its polygenic reports “are not intended to diagnose any disease”. The same page says the reports describe whether a person has variants associated with higher risk, but do not describe a person’s overall risk. That is company product material, not a scientific source. It is quoted here only to describe the product.

The NHGRI statement is the clearest one available.

Polygenic risk scores are not yet routinely used by health professionals because there are no guidelines for practice and researchers are still improving how these scores are generated.

The professional bodies are engaged with the question. In 2023 the American College of Medical Genetics and Genomics issued a points to consider statement on the clinical use of polygenic risk scores. The citation is Abu-El-Haija A et al., Genetics in Medicine 2023;25(5):100803, PMID 36920474. The full text sits behind a paywall. So this article cites its existence only, and draws no wording from it.

What to do with the PDF

The National Society of Genetic Counselors revised its position statement on consumer genetic testing in November 2025. It notes that such tests vary widely. The differences it lists are:

• analytical and clinical validity.
• scope of results.
• variant interpretation and reporting.
• regulatory oversight.
• how much a health professional is involved.

The same statement recommends genetic counseling before or after a test, for people unsure whether it meets their needs. Genetic counseling is a structured talk with a trained professional. It covers what a genetic result does and does not mean for you and your family.

Bring the actual report file, not a remembered sentence. Board-certified genetic counselors can be found through the public directory at NSGC.org.

Section recap: A polygenic report is a risk estimate. It is not a diagnosis and not a VUS. No practice guidelines yet govern its use. Take the file itself to a genetic counselor or your primary care clinician.

What Actually Changes Your Screening

There is a clean line here. On one side sit dated advice from named bodies. On the other sits what a polygenic score has not yet earned.

The current US position

The US Preventive Services Task Force issued its breast cancer screening statement on 30 April 2024. The grade B advice is biennial screening mammography for women aged 40 to 74 years. It applies to women at average risk. It does not depend on any polygenic score.

Two honest gaps sit in the same statement:

• For women aged 75 and older, there is not enough evidence to weigh benefits against harms.
• For women with dense breasts on an otherwise clear mammogram, there is not enough evidence to judge extra ultrasound or MRI.

A second US guideline, and an older one

The American Cancer Society published an average risk screening update in 2015. It advises regular screening mammograms from age 45. Screening is yearly from 45 to 54. From 55 it moves to every 2 years, or stays yearly by choice. Women aged 40 to 44 should have the chance to start yearly screening. The ACS does not advise a clinical breast exam for screening at any age.

The extra MRI threshold comes from an older ACS paper. The 2007 ACS guideline advises screening MRI for women with a lifetime risk of about 20-25% or more. That threshold belongs to the ACS 2007 statement. Both ACS papers predate 2021, so weigh them next to the 2024 task force statement.

That lifetime risk figure is worked out with tested risk models and family history. It is not read off a consumer report.

Canada is set province by province

Screening in Canada is set up province by province, so the same score gives different practical options across the border. Ontario is one example:

• Routine screening for average risk: a mammogram every 2 years, ages 50 to 74.
• People aged 40 to 49 can self-refer without a provider referral, and the program suggests discussing your own risk with a provider.
• A high risk stream for ages 30 to 69 offers a yearly mammogram with MRI or ultrasound.
• Entry to that high risk stream requires a family or medical history assessment by a health care provider or genetic counselor. It is not triggered by a consumer score. The page does not publish a lifetime-risk percentage threshold.

Do not apply Ontario’s rules to other provinces.

Prevention levers, and one term worth knowing

Some risk factors sit outside your DNA. The BOADICEA model treats lifestyle, hormonal and reproductive factors and breast density as separate inputs. Those are worth raising with your own clinician.

Chemoprevention means taking medicine to lower cancer risk before any cancer appears. It is a clinician-led choice for people with clearly raised risk. None of the guidelines cited in this article base that decision on a polygenic score.

The trials that would settle this have not reported

Two large randomized trials were designed to test whether tailoring screening to risk improves outcomes:

• WISDOM (NCT02620852) compares tailored screening with yearly screening. Target enrollment is 100,000. It started on 31 August 2016 and is still recruiting, with completion due 30 September 2029.
• MyPeBS (NCT03672331) compares tailored, risk based screening with standard screening in women aged 40 to 70. Actual enrollment is 53,142. It started on 18 July 2019 and is active but no longer recruiting, with primary completion due 31 August 2027.

Neither trial has reported primary outcome results. So the idea that a high score should mean extra scans is a hypothesis under test. It is not a guideline. Ask your primary care clinician what your own schedule should be.

Section recap: The 2024 task force advice is a mammogram every 2 years from age 40 to 74, whatever your score says. The 20-25% lifetime risk MRI threshold belongs to the ACS 2007 guideline. It is worked out from risk models and family history. Screening trials based on risk have not yet reported.

If a Diagnosis Happens: What Recent Approvals Changed

Keep this section separate in your mind. A polygenic score does not predict treatment. A high score does not decide what therapy anyone receives.

Treatment choice is driven by tumor biology. That means hormone receptor status, HER2 status, and mutations found in the tumor itself, such as PIK3CA, AKT1 or PTEN changes. Those are properties of the cancer, not of the DNA you inherited.

Two recent approvals show how that works in practice. Both share three features:

• Each is for advanced or metastatic disease, not for prevention.
• Each is chosen from tumor features, not from inherited DNA.
• Each has a different status in the US and Canada.

Drug	US FDA	Health Canada
Datopotamab deruxtecan-dlnk (Datroway)	Approved 17 January 2025	Not authorized as of 25 September 2026
Capivasertib (Truqap) with fulvestrant	Approved 16 November 2023	Authorized 26 January 2024

Datroway

The FDA approved datopotamab deruxtecan-dlnk on 17 January 2025. It is for adults with unresectable or metastatic HR-positive, HER2-negative breast cancer who already received endocrine-based therapy and chemotherapy for advanced disease.

The supporting trial, TROPION-Breast01 (NCT05104866), enrolled 732 patients. Median progression free survival was 6.9 months, against 4.9 months with chemotherapy. Overall survival was not statistically significant: 18.6 months against 18.3 months.

In Canada the picture is different. A brand-name search of the Health Canada Drug Product Database for Datroway returned zero records on 25 September 2026. An active-ingredient search for datopotamab also returned zero records. The only Canadian submission found for the drug covered a lung cancer indication and was cancelled by the sponsor on 21 March 2025.

Truqap

The FDA approved capivasertib with fulvestrant on 16 November 2023. It is for adults with HR-positive, HER2-negative locally advanced or metastatic breast cancer carrying one or more PIK3CA, AKT1 or PTEN alterations, detected by an FDA-approved test. FoundationOne CDx was approved as the companion diagnostic.

In the CAPItello-291 trial (NCT04305496), 708 patients took part, of whom 289 had those alterations. In that altered group, median progression free survival was 7.3 months against 3.1 months with placebo plus fulvestrant.

Health Canada authorized Truqap on 26 January 2024. The decision summary records a Notice of Compliance, with sponsor AstraZeneca Canada Inc.. Two Canadian drug identification numbers, 02544733 and 02544741, are listed as marketed.

Any question about treatment belongs with a treating oncologist, not with a risk report.

Section recap: Treatment is selected by tumor biology, not by inherited common variants. Truqap is FDA approved and Health Canada authorized; Datroway is FDA approved but not authorized in Canada.

Your Daughters, Your Siblings, and Your Insurance Paperwork

For a polygenic result, family and paperwork are one topic. Both hinge on the same fact. A chance is not a finding.

What children inherit

Children inherit half of each parent’s variants. They do not inherit the parent’s score. Which half they get is close to a lottery across thousands of positions.

So a parent in the top decile can have a child near the middle. Sibling scores also spread widely. This is not a case where one result predicts the rest of the family.

Cascade testing means testing relatives one at a time after a specific harmful variant is found in a family. It was designed for single gene pathogenic variants. It does not apply to a score, because there is no single variant to look for.

What does travel through a family

The family history does. Write it down and take it with you:

• Which relatives had cancer, on both sides of the family.
• Which cancer each of them had.
• How old each of them was at diagnosis.
• Three generations if you can manage it.

That written history is what a referral decision rests on. The NCCN hereditary cancer framework, Version 2.2026, exists to help clinicians identify people who may benefit from cancer risk assessment, genetic counseling and testing. That framework focuses on pathogenic and likely pathogenic variants, and referral is driven by the pattern of cancers and ages at diagnosis in relatives. A consumer score is not the trigger. A clinician applies the criteria.

The US legal picture, stated precisely

The Genetic Information Nondiscrimination Act of 2008 is known as GINA. Public Law 110-223 covers two areas:

• Health insurance (Title I): private insurers, Medicare, Medicaid, federal employee plans and the Veterans Health Administration.
• Employment (Title II): hiring, firing, promotions, pay and job assignments.

Now the gap, stated plainly. GINA’s health insurance protections do not cover long-term care insurance, life insurance, or disability insurance. Some states offer extra protections. Two more limits are worth knowing: GINA does not apply to employers with fewer than 15 employees, and its employment protections do not apply to the US Military.

If you are mid-application for a life insurance policy, that gap is the part to read twice. Check your own policy terms and your state’s rules.

Canada has a separate federal law

Canada’s Genetic Non-Discrimination Act, S.C. 2017, c. 3, was assented to on 4 May 2017. It bans requiring a person to take a genetic test, or to disclose test results. That ban applies as a condition of supplying goods or services, entering a contract, or offering particular contract terms.

Penalties are steep. On indictment they reach a fine of up to $1,000,000 and up to 5 years in prison; on summary conviction, up to $300,000 and 12 months. Exceptions exist for health care practitioners and medical researchers. Insurance terms still vary by province and by insurer, so check yours directly.

Telling your relatives

No clinical duty comes with a score of this kind. Sharing it is a personal choice. What is worth sharing is the family cancer history. That can affect whether a relative meets referral criteria.

And name the worry honestly. A percentile is not a prognosis. If the number is keeping you up, book time with a genetic counselor or your primary care doctor.

Section recap: Children inherit half of a parent’s variants but not the score, so cascade testing does not apply. GINA covers health insurance and employment but not life, disability or long-term care cover.

Frequently Asked Questions

Q1. Does an increased-likelihood result mean I will get breast cancer?

No. It is a position on a scale, not a verdict. In the 2019 study, lifetime risk in the top 1% of the score was 32.6%. That is about 33 out of 100 women.

Read the other side of it. About 67 out of 100 were never diagnosed. The baseline for US women is about 13 out of 100. A score moves the odds. It does not settle the outcome. Discuss your own report with a clinician.

Q2. Will my daughters inherit my high score?

They inherit half of your variants, not your number. Scores vary widely between parent and child and between siblings. Cascade testing, which means testing relatives after a specific harmful variant is found, was built for single gene results and does not apply here. The more useful thing to pass on is a written family cancer history with ages at diagnosis. Take it to their doctor when they are ready.

Q3. Will this affect my health or life insurance?

In the US, GINA restricts the use of genetic information in health insurance underwriting. It does not cover life, disability, or long-term care insurance. That gap is real and worth knowing before you submit a life insurance application. In Canada, a federal law prohibits requiring a genetic test or its disclosure as a contract condition. Provincial and insurer-specific terms still differ, so read your own paperwork and ask the insurer directly.

Q4. Can my employer find out?

GINA’s Title II restricts the use of genetic information in hiring, firing, promotions, pay and job assignments. Two exceptions matter. It does not apply to employers with fewer than 15 employees, and its employment protections do not apply to the US Military. A worry about your own workplace is a legal question, not a medical one. An employment lawyer is the right person to ask.

Q5. Should I get a second opinion, or is a consumer report the same as a medical test?

They are not the same. Consumer arrays read a fixed set of DNA spots, not the sequencing used in clinical panels. 23andMe states that its polygenic reports are not intended to diagnose any disease. The NHGRI notes that polygenic scores are not yet routinely used by health professionals, because no practice guidelines exist. The NSGC advises counseling before or after such a test for people who are unsure. Take the report file to a board-certified genetic counselor or your primary care clinician.

Summary: What to Do With the Number

A polygenic breast cancer score is real science. It is not yet a clinical tool. The useful output of reading one is a better-prepared appointment, not a change you make on your own.

What the evidence supports today:

• Breast cancer has a large common variant component, traced from the 2007 FGFR2 and TOX3 signals through the 2017 consortium study.
• PRS313 shifts risk measurably: an odds ratio of 1.61 per standard deviation, and 32.6% lifetime risk in the top 1%. The baseline is about 13 out of 100.
• Accuracy falls outside the discovery group. A European derived model explained only 0.79% of variance in South African Black women.
• No practice guidelines yet govern clinical use of these scores.
• Screening recommendations today come from dated guideline statements, not from scores.

Three things to bring to your next appointment:

1. A written three generation family cancer history, with ages at diagnosis.
2. The actual report PDF, not a summary sentence.
3. One question: does my family pattern meet criteria for a genetics referral?

That third question is the one a clinician can act on. Board-certified genetic counselors are searchable through the National Society of Genetic Counselors, and your primary care clinician can start the referral.

This article is for educational purposes only. It is not a substitute for advice from a licensed physician, board-certified medical geneticist, or board-certified genetic counselor. For any decisions about testing, treatment, or care, consult a qualified clinician. In emergencies, call 911.

References

1. Easton DF et al. (2007) Genome-wide association study identifies novel breast cancer susceptibility loci. Nature 447(7148):1087-1093. https://pubmed.ncbi.nlm.nih.gov/17529967/
2. Michailidou K et al. (2017) Association analysis identifies 65 new breast cancer risk loci. Nature 551(7678):92-94. https://pubmed.ncbi.nlm.nih.gov/29059683/
3. Mavaddat N et al. (2019) Polygenic Risk Scores for Prediction of Breast Cancer and Breast Cancer Subtypes. American Journal of Human Genetics 104(1):21-34. https://pubmed.ncbi.nlm.nih.gov/30554720/
4. Lee A et al. (2019) BOADICEA: a comprehensive breast cancer risk prediction model incorporating genetic and nongenetic risk factors. Genetics in Medicine 21(8):1708-1718. https://pubmed.ncbi.nlm.nih.gov/30643217/
5. National Cancer Institute (NIH), SEER Program. SEER Cancer Stat Facts: Female Breast Cancer. https://seer.cancer.gov/statfacts/html/breast.html
6. National Cancer Institute (NIH). BRCA Gene Changes: Cancer Risk and Genetic Testing (reviewed 2024-07-19). https://www.cancer.gov/about-cancer/causes-prevention/genetics/brca-fact-sheet
7. Hayat M et al. (2025) Genome-wide association study identifies common variants associated with breast cancer in South African Black women. Nature Communications 16:3542. https://pubmed.ncbi.nlm.nih.gov/40229280/
8. Kachuri L et al. (2024) Principles and methods for transferring polygenic risk scores across global populations. Nature Reviews Genetics 25(1):8-25. https://pubmed.ncbi.nlm.nih.gov/37620596/
9. National Human Genome Research Institute (NIH). Polygenic Risk Scores (last updated 2020-08-11). https://www.genome.gov/Health/Genomics-and-Medicine/Polygenic-risk-scores
10. National Human Genome Research Institute (NIH). Genetic Discrimination and the Genetic Information Nondiscrimination Act of 2008. https://www.genome.gov/about-genomics/policy-issues/Genetic-Discrimination — and Government of Canada, Genetic Non-Discrimination Act (S.C. 2017, c. 3). https://laws-lois.justice.gc.ca/eng/acts/G-2.5/page-1.html
11. US Food and Drug Administration. FDA approves datopotamab deruxtecan-dlnk for unresectable or metastatic HR-positive, HER2-negative breast cancer (2025-01-17). https://www.fda.gov/drugs/resources-information-approved-drugs/fda-approves-datopotamab-deruxtecan-dlnk-unresectable-or-metastatic-hr-positive-her2-negative-breast — and FDA approves capivasertib with fulvestrant for breast cancer (2023-11-16). https://www.fda.gov/drugs/resources-information-approved-drugs/fda-approves-capivasertib-fulvestrant-breast-cancer
12. Health Canada, Drug Product Database and Drug and Health Products Portal (queried 2026-09-25). https://health-products.canada.ca/dpd-bdpp/ — Regulatory Decision Summary, Truqap, control number 274859. https://dhpp.hpfb-dgpsa.ca/review-documents/resource/RDS1718733982360
13. ClinicalTrials.gov. WISDOM (NCT02620852). https://clinicaltrials.gov/study/NCT02620852 — and MyPeBS (NCT03672331). https://clinicaltrials.gov/study/NCT03672331
14. Government of Ontario. Ontario Breast Screening Program (page updated 2026-07-02). https://www.ontario.ca/page/ontario-breast-screening-program
15. Nicholson WK, Silverstein M, Wong JB, et al.; US Preventive Services Task Force (2024) Screening for Breast Cancer: US Preventive Services Task Force Recommendation Statement. JAMA 331(22):1918-1930. https://www.uspreventiveservicestaskforce.org/uspstf/recommendation/breast-cancer-screening
16. Abu-El-Haija A et al.; ACMG Professional Practice and Guidelines Committee (2023) The clinical application of polygenic risk scores: A points to consider statement of the American College of Medical Genetics and Genomics. Genetics in Medicine 25(5):100803. https://pubmed.ncbi.nlm.nih.gov/36920474/
17. Cheng HH et al. (2026) NCCN Guidelines Insights: Genetic/Familial High-Risk Assessment: Breast, Ovarian, Pancreatic, and Prostate, Version 2.2026. JNCCN 24(1):2-10. https://pubmed.ncbi.nlm.nih.gov/41671423/
18. Oeffinger KC et al. (2015) Breast Cancer Screening for Women at Average Risk. JAMA 314(15):1599-1614. https://pubmed.ncbi.nlm.nih.gov/26501536/ — and Saslow D et al. (2007) American Cancer Society Guidelines for Breast Screening with MRI as an Adjunct to Mammography. CA: A Cancer Journal for Clinicians 57(2):75-89. https://pubmed.ncbi.nlm.nih.gov/17392385/
19. National Society of Genetic Counselors. Consumer-Initiated Genetic Testing Position Statement (approved June 2019, revised November 2025). https://www.nsgc.org/Policy-Research-and-Publications/NSGC-Position-Statements/Position-Statements/Post/at-home-genetic-testing-position-statement
20. 23andMe. Health Predisposition reports for clinicians (accessed 2026-09-25). https://www.23andme.org/clinicians/genetic-reports/health-predispositions/

---

Last updated: 2026-09-25

Author: genelumen editorial team. This article aggregates 20 sources from peer-reviewed medical literature, public health agencies and drug regulators (tier 1=14 / tier 2=5 / tier 4=1). Sources include the National Institutes of Health (NCI/SEER and NHGRI), the US Food and Drug Administration, Health Canada and the Government of Ontario. They also include the US Preventive Services Task Force, the American Cancer Society, ClinicalTrials.gov and PubMed-indexed publications.

This article is for educational purposes only and is not a substitute for medical advice from a licensed physician, board-certified medical geneticist, or board-certified genetic counselor. In emergencies, call 911.

Related: Hereditary Cancer articles

🌐 日本語版: 乳がんの遺伝リスクは BRCA だけではない｜GWAS 研究が示すポリジェニックスコアと日本の検診