What's your Gail risk assessment?

Options
2»

Comments

  • otter
    otter Member Posts: 6,099
    edited June 2011

    I've been looking at this thread for almost a month now, and it's been nagging at me.

    FWIW, if I had used the GAIL risk calculator four years ago, which would have been 7 months before my BC diagnosis:

    ... my 5-year risk of BC would have been just 2.0% ("average" for a woman that age being 1.5%); and my lifetime risk of BC would have been 13.3%, "average" lifetime risk being 10.2%.

    The main reasons why my risk would have been that high were because I've never had children (never been pregnant, even) and I'd had two cysts aspirated.  An aspiration of fluid from a cyst counts as a "biopsy" in the GAIL formula, just as if it was a calcified nodule being sampled with a core needle or surgical scalpel.  I thought that sucked, so I asked for an explanation.  I was told any time there was something in the breast that warranted taking fluid or tissue to be tested for abnormalities meant the woman had an increased risk of other abnormalities in that breast, including cancer.  Oh, really?

    Now, of course, having been diagnosed with invasive BC in early 2008, I cannot use the GAIL risk calculator to estimate my risk of a new primary tumor.  I wrote to the National Cancer Institute (http://www.cancer.gov/global/contact) to ask if there were any risk calculators available for women who had already been diagnosed with BC, so we could find out our risk of developing a second primary BC... and they said "sorry, no".

    otter

    · Share on FacebookShare on Twitter
  • leaf
    leaf Member Posts: 8,188
    edited June 2011

    Since the Gail model is barely better at predicting breast cancer (for any individual woman) as total chance, I'm not at all surprised at your results, otter.   

    To get the concordance statistic, they take one random woman from each of 2 groups: one from the group that ended up getting breast cancer, and one from the group that ended up NOT getting breast cancer.  They compare their Gail scores (before any breast cancer occurred.)  About 55 or 60% of the women who are in the breast cancer group will have gotten a higher score than the no breast cancer group.  Unfortunately, that also means that about 40 or 45% of the women who ended up getting breast cancer had a LOWER Gail model score than the random woman who never got breast cancer.  

    In the one study with atypia, the model was even worse:  in half of the pairs, the atypia patient had a higher score than the normal patient, and in the other half of the pairs, the normal patient had a higher Gail score than the atypia patient.  That means the Gail score was totally worthless to predict whether or not the patient got atypia (in this particular study.)

    This page in Halls' calculator looks at how the Gail model is calculated.  It looks like it adds up relative risks, then sticks that value into an equation to get absolute risk. http://www.halls.md/breast/gailmods.htm  The Gail model actually works quite well for larger GROUPS of women.  It will pretty accurately predict how many women in a larger group will get breast cancer.  However, it is really terrible at predicting WHICH of those particular women in the group will get breast cancer.

    · Share on FacebookShare on Twitter
  • rn4babies
    rn4babies Member Posts: 409
    edited June 2011

    5 Year Risk 11.9....Lifetime Risk 45.9

    · Share on FacebookShare on Twitter

Categories