New MAYO Cancer Risk Assessment Tool for Benign Breast Disease

Lea7777

https://www.mayoclinic.org/breast-cancer-risk-pred... (when I copy and paste the link, it always gets cut off). Here it is again, below.

www.mayoclinic.org/breast-cancer-risk-prediction/i...

Still getting cut off

put in the w w w and then dot mayoclinic.org/breast-cancer-risk-prediction/itt-20150095

I think I tricked it.

Development of the model was published in April 2018 here

https://www.ncbi.nlm.nih.gov/pubmed/29676945

Model for Predicting Breast Cancer Risk in Women With Atypical Hyperplasia.

Mayo felt the Tyrer Cuzick IBIS model over estimated cancer risk for women with atypia, so they devised this model. For the highest accuracy, if you have ALH you need to know the # of foci you have.

When I tested it, my 30 year risk with the Mayo model was about 5% less than the Tyrer Cuzick IBIS model, inputting my # of foci of ALH, which I had requested in the past from pathology.

I don't like the results of either model, which is way more than the 25%-30% lifetime risk that is the general guidance.

If this was already posted, maybe a year ago when it came out, apologies for the duplicate.

momoschki

Wow, this is scary. I was dx’ed with ADH 8 years ago at the age of 53. This model predicts a 30 year risk of 35%, which is higher than any doctor has predicted. Hard to take in, especially since I have another dreaded 6 month check coming up in less than 2 weeks..

Lea7777

I know this is scary and much more like the Tyrer-Cuzick IBIS model that every doctor I know has discounted as way over-estimating risk and inaccurate. But Mayo, the source of this new model, wouldn't screw it up! Would they? Predictions compared to actual results are explained in this study

New Model Can Assess Breast Cancer Risk in Women With Atypical Hyperplasia

by Dave Levitan

April 30, 2018

in the CancerNewtwork site

https://www.cancernetwork.com/breast-cancer/new-mo...

W w w then a dot then cancernetwork.com/breast-cancer/new-model-can-assess-breast-cancer-risk-women-atypical-hyperplasia

"The ratio of predicted cancers to observed cancers at 5, 10, and 20 years was 0.96, 0.87, and 0.92, respectively." So this is about 90% accurate over 20 years with real observations. Interesting that the 20 year accuracy figure exceeded the 10 year, but ALL are 87% accurate or more.

Unless there are other factors that I do not understand or am not aware of, this should revolutionize what we are being told about our risk. Results from Mayo's Breast Cancer Risk Assessment Tool (BCRAT) were first published in the March 10, 2015, issue of the Journal of Clinical Oncology, from what I can tell. So that is almost exactly 4 years for experts digest and review these studies and for Mayo to refine the model if needed. It is also 4 years to raise any flaws or qualifications that may impact the accuracy of the results. I could find nothing that rebuts these results. Finally it is 4 years where this sort of research should be filtering down to the doctors we see so they can pass it on--again, if it is valid and accurate. But Mayo wouldn't screw it up? Would they?

@Momoschki, I see you mentioned you were diagnosed with ADH. The biggest risk in this model is ALH rather than ADH. Just want to be sure you have made that distinction, which is in your favor.

momoschki

Lea, thanks, I somehow missed the ADH/ALH distinction. I’ll need to go back and reread more carefully. The risk calculator itself does not make the distinction, which is how I ended up with the 35% number.

I find that even after 8 years, I’m still rooting around looking for something that will conclusively nail down my risk. The uncertainty of all of this is the hardest part for me. Doctors have previously estimated my risk to be in the neighborhood of 20-25%, which I guess I had come to terms wuth

Barbara1120

@Lea

Well, this is distressing, isn't it? I met with a team at Mayo in July 2018, and not one of them mentioned this study. In fact, all team members, rather confidently, encouraged me to continue on Anastrozole and discouraged a discussion of any other preventative measures (except life style). Once I read your post, I shot an email to my MD at Mayo about this study and haven't received any follow-up, as of the moment.

I also had an appointment with Tari King in September 2018 - she's the author of the 29 year study on LCIS out of MSK - and she also never mentioned this Mayo model and I definitely discussed with her the awful numbers of the Tyrer Cuzick. She did not mention that the risk assessment modeled by Tyrer Cuzick was essentially confirmed by Mayo. She, too, discouraged any preventative measures other than Anastrozole and lifestyle.

I have been told by my local doctors in Chicago that my risk is much less than either the Tyrer or, now, the Mayo model indicate. It's higher than I'd like but less than either of those two models.

So, what gives? Frankly, these models make me think that taking these AIs and introducing lifestyle management/changes are a waste of time. At the high end of the spectrum, for someone with 3 or more foci of LCIS, the risk is over 60% for one's lifetime. From what I understand, if LCIS shows up on a biopsy in one place, you can almost be guaranteed it's lurking somewhere else (multi-focal). Even if one can tolerate an AI for 5/7/10 years, you can at best, hope to reduce it to a 30% risk. That's not very comforting.

light1candle

Lea777, thanks for posting this. I am glad to see a new calculator designed for women with atypia, but I find this to be a very "stripped down" assessment. When I went into the tool I only had three questions: age, biopsy findings (choose one: non-proliferative disease, proliferative disease w/o atypia, or atypical hyperplasia [AH]), plus the number of foci (one, two, three or more, unknown). Did I miss something? I didn't see any questions that distinguished between ADH and the various levels of lobular neoplasia, including nothing about pleomorphic or florid variants. This seems like important information for a predictive tool that is "intended to be used by health care providers" who presumably have that sort of information.

Also nowhere in either of my two pathology reports does it offer information on the number of foci, so I had to guess about that, which makes a big difference in the results the calculator provides. Does one biopsy site equal one focus? Or can you have multiple foci with a stereotactic core biopsy and one rather large excisional biopsy, with 28 slides?

Beesie

Same issues here as noted by light1candle, and adding to her post, the model doesn't ask about family history or breast density, which are significant factors in risk assessment. Something doesn't seem right.

Lea7777

On the Mayo calculator I found that IF you have ALH, no other questions besides age at diagnosis are asked. That could account for the streamlining of questions. That tells me according to this model, ALH is destiny, regardless of when I had my first period or if I am overweight.

As for pleomorphic or florid variants, I would think that would make a difference.

For info on # of foci, I asked specifically that pathology provide that info. In fact, it was Mayo that I asked, in a second opinion. (Did not know about this assessment tool when I was at Mayo and Mayo did not mention it). For me, Mayo counted foci not as areas of the breasts with atypia or how many slides of atypia, but affected cells. So I have mucho foci. One radial scar had 4 foci of ALH. After looking at the printed records again, those 4 foci alone give me a 2%-2.5% annual risk. It's in writing in my Mayo records. That does not even include the ADH foci or LCIS.

Here's one factor I could think of that might make the risk outcomes appear so dismal. The study on which this assessment tool was based was first published in 2015, I believe. It looked at data from at least 20 years back in order to compare the 20 year risk with actual 20 year observations. Twenty years ago, or even ten years ago, the methods to detect atypia were not as advanced and fine tuned as today. For atypia to be diagnosed, the foci may have had to be more evident and easily identifiable. So perhaps the more advanced needle biopsies and MRIs of today (and that diagnosed us) see every little abnormality. And that really racks up the foci count. I don't know, just trying to rationalize. But there should be disclaimers then in the Mayo study about new and old technology, I would think.

When I was at Mayo about 6 months ago, I did ask Mayo about this chart below, which is based on Mayo research and validated by Vanderbilt University. Authors are shown here, and names show Mayo connections. I think it originally came out at the beginning for 2014.

Amy C. Degnim, M.D.,¹William D. Dupont, Ph.D.,²Derek C. Radisky, Ph.D.,³Robert A. Vierkant,⁴Ryan D. Frank,⁵Marlene H. Frost, Ph.D.,⁶Stacey J. Winham, Ph.D.,⁴Melinda E. Sanders, M.D.,⁷Jeffrey R. Smith, M.D., Ph.D.,⁸David L. Page, M.D.,⁹Tanya L. Hoskin, M.S.,⁴Celine M. Vachon, Ph.D.,¹⁰Karthik Ghosh, M.D.,¹¹Tina J. Hieken, M.D.,¹Lori A. Denison,¹²Jodi M. Carter, M.D., Ph.D.,¹³Lynn C. Hartmann, M.D.,⁶ and Daniel W. Visscher, M.D.¹³

I was told to wait and see by Mayo in response to this chart. They were not adopting this research as the standard at that time. But apparently Mayo did feel confident enough about this foci-stratified research to put out a risk assessment tool and make it available for online.

Source:

Published in final edited form as:

Cancer. 2016 Oct; 122(19): 2971–2978.

Published online 2016 Jun 28. doi: [10.1002/cncr.30153]

PMCID: PMC5030128

NIHMSID: NIHMS787891

PMID: 27352219

Extent of Atypical Hyperplasia Stratifies Breast Cancer Risk in Two Independent Cohorts of Women

farmerlucy

I had ALH. My genetic counselor gave me 50% lifetime risk. This one calculated 60%.

Lea7777

Both of those #s are in the ballpark, @Farmerlucy. But they are both higher than just about every reputable cancer source out there for ALH. Even LCIS, which has a higher risk than ALH, is shown as about a 20%, maybe up to 35% lifetime risk.

light1candle

I found a 2016 article, 'Extent of Atypical Hyperplasia Stratifies Breast Cancer Risk in 2 Independent Cohorts of Women'. This article uses the data from the Mayo and Nashville studies to talk about how multiple foci of Atypical Hyperplasia (both ADH & ALH) increases breast cancer risk.

https://onlinelibrary.wiley.com/doi/pdf/10.1002/cncr.30153

The article answered my question about foci: yes, a pathologist can count multiple foci in one biopsy event. But apparently the practice of counting and reporting the number of foci of AH is not standard practice for many pathologists.

In the Mayo and Nashville studies: "For each sample with AH, the number of separate foci of ADH and ALH was recorded, and the number of slides was evaluated per subject. Separate foci of AH were defined as involving separate terminal duct lobular units with clear separation by mammary stroma between lobules."

The article also addresses some of the different findings for ADH and ALH. "In both cohorts (Mayo and Nashville), women with ALH were more likely to have 3 foci in comparison with women with ADH."

About age at time of biopsy: "The number of foci of AH was not associated with the age at biopsy in either cohort."

Findings: " [F]igures summarizing the cumulative incidence of BC in women based on the extent of AH provided important estimates of absolute risk and showed a 25-year risk of 30% for the group as a whole but a significantly higher risk (47%) for the minority of women with multiple foci of AH. Furthermore, the risk was fairly constant over time, with averaged annual risks of approximately 1%, 1.5%, and 2% with an increasing number of foci of AH."

Lea7777

Thanks for that additional research, Light1Candle. That would make sense with the higher risk of cancer with ALH vs ADH.

momoschki

So then here’s another question: if one’s pathology report does not specify number of foci, is there any way to make this determination ? Can 2 biopsy samples contain one focus? And can 2 (or more) positive samples contain only one focus?

pegasus68

I'm not sure about this. When it asks for 'atypia', does that include LCIS? LCIS is known to have higher risk than ALH. It doesn't seem to request that distinction (i.e. like the Gail model states that it cannot be used for women with LCIS). I'm also uncomfortable it seems to completely ignore all other risk factors.

My last Tyrer-Cuzick performed by a genetics counselor came up at a ridiculous 76% (which we know is an overestimate). Even if I cheat and 'demote' my LCIS to ALH to be able to perform the Gail model, my risk comes up at 42% (which reportedly underestimates for AH.) This Mayo model, entering an unknown number of foci but ignoring all other risk factors, comes up at 35% for me. (Wasn't reported, although a number of lobules were indicated to be involved.) Not sure what number to believe.

Lea7777

Here's my dilemma, @pegasus68. I think those 60% and 70% numbers are high. But Mayo put it in writing that my risk, based on the slides from my left excisional biopsy, is 2%-2.5% per year. Assuming I live another 25 years, which is very reasonable. 25 years x (average of 2% and 2.5% per year) = 25 years x 2.25%/yr = 56.25%.

MsRukia

it doesn’t seem as though my risk is incredibly high. However, I wonder how accurate it is? It didn’t ask too many questions. It was interesting though

Lea7777

I wonder how accurate it is? -- that is the key question MsRukia. It is not like this is some silly online game or survey where it does not really matter what the outcome is. Glad your risk #s did not come out high.

light1candle

Just some observations: I fooled around with this tool a little bit, varying only the age. The tool tells you to input your age at biopsy/diagnosis. The first time I tried the tool I mistakenly entered my current age (65) instead of my age when the biopsy/diagnosis occurred (barely 64) and found that one year made a quite a bit of difference in the risk percentages : 20 yr. risk at 41.98 for age 64, compared to 20 yr. risk of 34.92 at age 65. (I kept the answers to other two questions constant as having Atypical Hyperplasia and 3 or more foci of AH.)

After messing around with the calculator, I see that it divides age into 5 yr. increments: ie everyone from 60 thru 64 gets the same scores, assuming the answers to the other two questions stay the same. The risk numbers change for women 65-69, and again for 70-74, etc. I also noticed that if age and AH are constant, you will get the same score for an unknown number of foci as with just one foci.

I charted out the numbers for ages 30, 35, 40, 45, 50, 55, 60, 65, 70, using AH and 3 or more foci (as is typical with ALH/LCIS). The highest risks for all projections (5 yr, 10 yr, 15, yr, 20 yr. 25 yr, 30 yr) are for women between 50-59 years old when diagnosed with AH. The risk percentages are somewhat less for younger ages and for older ages. Interestingly, the risk numbers are *almost* identical for women from 45-49 yrs old and for women from 60-64 yrs. old.

Lea7777

My goodness, LIght1candle, you get a PhD in the Mayo assessment tool. Perhaps you should join their staff! Interesting that the characteristics you found for highest risk pertains to me.

Your findings on risk is the sort of work I'd like the researchers to do and then tell us--who or what type of breast cells are most at risk for invasive cancer? Maybe those general numbers we see that hover around 30% are right, on average. But what are the characteristics of the group that makes up that 30%? Within that specific group, the risk may be 90%.

Somehow I am a bit hopeful that this kind of info will emerge. But can we hang on (to them) that long?

Thanks for your work and for sharing with us, Light1candle. When I go back to Mayo, which probably will not be for about 6 months, I'll ask about this tool and share what I find.

JaneyK12

Just came upon this great thread. Many thanks to all you great mathematicians!

I have ALH. It was discovered from a mammo as microcalcifications 3 years ago. I had the excisional biopsy in 2016 and have been on tamoxifen (which led to a hysteroscopy) so am now on Arimidex. When I asked my BS if he got it all--he said he couldn't--which to me means it was diffused through my breast and that translates to me as many foci. And now, after reading this thread, it seems that my lifetime risk is at least 20% if not more. And according to all I've read 20% means you should have yearly MRIs. After reading all these estimates of future BC, that leads me to this issue.

Here's my question: my BS says he's leaving it up to me if I want my yearly MRI. Of course I want it! Why would I not want to catch it before it arrives! And I have dense breasts. I asked if the dye affects my kidneys and is that his reason to not have it. He said no. I think he said it may lead to false positives. I personally can deal with that if it means I can prevent BC.

I decided I needed to be seen by an MO a few years ago so I have another good mind to guide me and she is compassionate. She says yes, have the MRI because of the ALH and the density.

What would you do? And just for my own peace of mind, please let me know if any of your surgeons left the decision up to you.

Please PM me, if possible.

Thanks.