The Myriad Model, also known as the Frank model, was one of the first models to estimate the risk of having a BRCA1 or BRCA2 mutation based on family history.  It is the only breast cancer risk model that directly relates to actual genetic test results, and it is constantly being updated as additional results become available.

The current tables are available at:

https://www.myriadpro.com/hereditary-cancer-testing/hereditary-breast-and-ovarian-cancer-hboc-syndrome/prevalence-tables/

While this was a major advance in its time, and while the data remain fascinating, the strengths and weakness of this model must be appreciated in order to avoid misusing its output.

The model is based on patients who have had genetic testing ordered from Myriad genetics, using the parameters:

  • The patient’s personal history of cancer
  • The patient’s family history submitted on the requisition for up to 2 family members
    • Any ovarian cancer
    • Breast cancer ,age 50
    • No provision for multiple primary cancers
  • The results of the genetic test
  • Whether the patient is Ashkenazi Jewish or not

The models is weak in that it overestimates risk when there is no ovarian cancer and no breast cancer under age 50 in the family, and it is weak when there is a strong family history.  The source of these weaknesses is easily explained.

No ovarian cancer and no breast cancer under age 50 in the family

According to the original Tables published in 2002, a non-Ashkenazi woman who does not have breast cancer herself and has with No ovarian cancer and no breast cancer under age 50 in the family had a risk of mutation of 3.9%.  In today’s updated tables, her risk has dropped to 1.5%.  However, the actual risk of mutation for such a patient is likely to be about 1/350 or about 0.2%.

The high risk of mutation can be explained by the effect that the population was limited to women for whom genetic testing was ordered, hence there was something in their family history that suggested that this test should be done.  On the other hand, the family history used in the calculation was that provided on the requisition.  Putting these 2 facts together would make it likely that the family history was not accurately added to the requisition or not added at all.

It is also of interest that the risk of mutation has dropped since 2002 for all categories of non-Ashkenazi women, to the point that no one who does not have cancer herself has more than a 7.2% risk of mutation.  This is likely because more women are being tested now, even with a lower risk of mutation.  In addition, the highest risk women are being subsumed in groups of lower risk women.  For example, a relative with breast cancer at 25 adds no more risk than a relative with breast cancer at 49.  By the same token, women with more than 2 relatives with cancer or relatives with bilateral breast cancer or breast plus ovarian cancer are all hidden in categories of much lower risk.

Overall, use caution when using the Myriad tables, as they may be misleading at the extremes of family history.

Share This