Breast cancer (female)

Breast cancer has a well-studied common-variant component that is entirely separate from BRCA1 and BRCA2. This score combines 11 of those common variants, in a version developed for Korean women, and applies to people with breast tissue.

What this breast cancer score reads

This is PGS003396, an 11-variant score developed by Jee and colleagues in 2022 specifically for Korean women, using GWAS data from 95,283 East Asian participants. It applies to people with breast tissue, and it was built and validated in one population.

Eleven variants is very few for this trait. The point of the paper was to test what a compact, East Asian-derived score could do, not to build the most powerful predictor available.

How well it predicts breast cancer

In 41,031 Korean women, each standard deviation higher score carried a hazard ratio of 1.35, with an area under the curve of 0.58. An area under the curve of 0.58 is close to the low end of useful: 0.5 would be a coin flip.

For scale, the widely used 313-variant breast cancer score developed in European-ancestry women reports an odds ratio of 1.61 per standard deviation and an area under the curve of 0.630, with women in the top 1% reaching a lifetime risk near 33%. This score is a weaker instrument than that one, and neither is a substitute for a risk assessment that includes family history and breast density.

What this score cannot see: BRCA1, BRCA2 and PALB2

The American College of Medical Genetics states the point directly: a breast cancer polygenic score does not assess monogenic risk such as BRCA1-associated breast cancer. Those genes carry rare variants with far larger effects, they change screening and prevention decisions, and they require clinical sequencing that a consumer genotyping chip does not perform.

If your family history includes breast cancer under 50, ovarian cancer, male breast cancer, or several affected relatives, that pattern is a reason to ask about genetic counselling regardless of any polygenic percentile, high or low.

What else it leaves out

Mammographic breast density, age at first period and at menopause, pregnancies and breastfeeding, hormone therapy, alcohol intake and body weight after menopause all shape breast cancer risk, and clinical risk models use them. None are in a DNA file.

Ancestry and accuracy

This score was derived and validated entirely in Korean women. For people of East Asian ancestry that is a genuine advantage, since most breast cancer scores were built in European-ancestry cohorts and transfer poorly. For everyone else, its measured performance simply has not been established.

How much of this score your file covers

All 11 variants in the published score are present in a standard consumer DNA file, so this reconstructs the published score in full.

Common questions

Does a low score mean I do not need mammograms? No. Screening recommendations are based on age and clinical risk factors, and a low polygenic score does not rule out significant risk, which is one of the explicit cautions in the American College of Medical Genetics statement on these scores. This score in particular has an area under the curve of 0.58, which is far too weak to change screening.

How does this compare with a BRCA test? They answer different questions. A BRCA test looks for rare damaging variants with large effects on risk, sequenced in a clinical laboratory. This score sums 11 common variants with small effects. A polygenic result, high or low, says nothing about your BRCA status.

Related

Alzheimer's disease · Atrial fibrillation · Inflammatory bowel disease · Prostate cancer (male) · Systolic blood pressure · Type 2 diabetes protection (PPARG Pro12Ala) · Venous thromboembolism (clots)