Tyrer-Cuzick IBCIS risk assessment questions

Lea7777

We don't have a lot of studies that detail other potential risk factors LCIS women may have.

Right, no data, no conclusions.

The LCIS picture may also be more complicated: if you go on to get invasive breast cancer where you had a spot of LCIS, you may have different pathology than if you get invasive breast cancer at a place that looked totally normal in the past.

True. We are a complex lot. All the more reason to have the models and data that do pertain to LCIS be complete and accurate.

LL99LL

Thank you Lea7777, Mammbear, leaf, Barbara1120 and others here for your research and thoughts on this topic.

light1candle, in response to your question - "did any of you find that data (the ratio indicating volume of disease) in your pathology report":

This info was not included in my pathology report. I asked my surgical oncologist to find out. She seems to be very current with LCIS research and it turned out she had already reached out to the pathologist to ask this question before I asked her to do so. She let me know the total # of slides and the # of slides showing LCIS a few days later. Given the results of the 29-year study, I would hope that pathologists would start automatically including this info when LCIS is found- LCIS diagnosed 10/2017; ALH and sclerosing adenosis diagnosed 1/2018.

Lea7777

LL99LL, of all the advanced testing that is done, counting slides should be the easy part of the pathology report and should be included, you are right.

Barbara1120

LL99LL,

In an earlier post discussing your results from the Tryer-Cuzick test, you reminded us to check the “competing mortality rate box when inputting our profile assessment. I’ve looked at the version of the test I’ve completed (linked below) and I don’t see that option anywhere on the test. Am I overlooking that option or perhaps using a different version of the test? Thanks for your help.

http://ibis.ikonopedia.com/

cattledoglv

I think some people are misunderstanding the 12% chance of getting BC. This is a Lifetime risk which is not the same thing as being say, 45 and having a 4% risk. At my age, 39, with no other risk factors, I would have a 1.5% chance of getting breast cancer, but my risk goes up a little each year to become around 12% in my 80’s. . However, I was just dx with ADH, a few other things not mentioned on this model and I have extremely dense breasts. It says I have a 6% chance within 10 years, but a 43% chance before 85. That is how you can have a 10%, or lower than 12% chance in these models.

Lea7777

Cattledoglv, would you care to share how you are dealing with your increased risk? Give your cattle dog(s) a hug for me.

light1candle

Again, slightly off-topic, but here's something else I have wondered about LCIS risk. Most of the research I have read says that the risk of LCIS progression to DCIS or invasive BC is about 1-2% per year from the time of diagnosis. It's that "from the time of diagnosis" which bothers me.

Since most LCIS is not usually found by imaging but is found incidentally on biopsy for something else that did show up on imaging, might we assume the possibility that the LCIS has been there silently for some time (maybe many years)? Why should we just assume that the clock starts ticking only when it is found? If this is the case, then presumably younger women diagnosed with LCIS would have a *much* increased risk of progression than someone diagnosed in their 60's like me.

In my case, I was diagnosed in 2017 at the age of 64 after my regular mammogram showed some suspicious microcalcifications, which were followed up with biopsies. The stereotactic core needle biopsy found ALH & intraductal papilloma where the calcifications were. That was followed by an excisional biopsy which added LCIS (both classic A & B types), radial scar, & columnar cell changes -- a real soup of benign but high risk proliferative stuff. Here's the thing: I have been having call-backs for diagnostic mammos on various calcifications off & on since at least 1994 when I was 41. The 2017 mammo was the first digital 3-D mammo and it led to a recommendation for biopsy, where they found the LCIS and set the clock ticking. Perhaps this was because the rad saw enough change in the calcs compared to previous mammos to recommend biopsy, or perhaps the new technology revealed details that were hidden before.

But who is to say that those abnormal cells have not been around for years before the 3D mammo & biopsy? At a risk of 1-2% per year for (possibly) 23 years, is that risk cumulative? I am currently doing 6 mo. imaging but am considering PBMx. It might change (or hasten) my decision-making on mastectomy if I thought that the clock really started to tick 23 years ago...

Lea7777

Good question, light1candle, and one I just posed yesterday to the Johns Hopkins Ask an Expert View a Question forum on breast cancer.

http://www.hopkinsbreastcenter.org/services/ask_ex...

QUESTION 4/26/2018

How can breast cancer risk for LCIS be calculated from age of diagnosis, when the LCIS could have been present but undetected for many years previously? The 2% cumulative risk per year could have started a decade ago, so the patient would already be at 20% lifetime risk upon diagnosis. Each year of life going forward would add to the risk.

ANSWER 4/27/2018

There is no way of knowing how long the LCIS was there; risk can only be determined moving forward from biopsy diagnosis.

Just one more mystery about LCIS. The question you pose, light1candle, is crucial and basic as a starting point for all risk related numbers that follow. Without knowing how long your LCIS has been present, the accuracy of all projections going forward is reduced. Furthermore, all of the studies to produce the statistics we now use never incorporated how long LCIS had been present in the variouis subjects.

When so much is unknown and breast cancer risk outcomes are so disparate, I think the tendency is to do one two extremes (1) Discount the projections, studies, and models as too unreliable at this point and proceed as we have always done--regular screeinings, no drugs, no surgery. (2) Seize control of what we can control and remove the culprit, our breasts with abnormal cell structure.

light1candle

Lea7777, thank you for posting about your research and your well-reasoned questions. I find this thread very helpful as I sort through the research myself -- kind of a like an LCISer's book discussion group. ;-) .I haven't really had anyone in my personal sphere to bounce around these kinds of questions or concerns.

It disturbs me that I can't seem to get satisfactory answers from my interactions with my doctors & other medical professionals. For instance, why don't pathologists count the ratio of slides with LCIS to the total number of slides? Also, one of my first questions on this site was did other women get information on hormone receptor status in their pathology reports, because I did not. Yes, I do understand that by far most (like 95%) cases of LCIS are ER+/PR+ , but I have also read that the cases that hormone receptor negative are much more likely to progress to cancer, and perhaps high grade cancer at that. Why can't we get this information to help give us more data points in our decision making process? I suspect that if a women knew she had hormone receptor negative LCIS, she wouldn't bother considering the anti-hormonal drugs (duh!), and would move right on to BPMx. I'm sure it has to do with cost effectiveness, but I still think this information would be helpful.

There just doesn't seem like there is a lot of consensus about LCIS risk factors and I will admit to sometimes feeling like we are the "poor cousins" to people with other high risk conditions or precursor conditions.

Lea7777

I too asked about my hormone receptor status and was told it is not something determined for LCIS, light1candle. For the other medical issues I have had in my life, which fortunately have not been many nor too serious, I have found specific, accurate, and agreed upon information that I could put to use in my specific situation. Not so in this case.

This is the Poor Cousins Book Club.

light1candle

LOL! The Poor Cousins Book Club! Love it!!!

Lea7777

Something must be amiss with this model. When many family members have both breast and ovarian cancer, the risk is less than if those family members were cancer free. Those findings are shown here:

How can the results of Case 1 and Case 2 possibly be valid? Interestingly, Case 7 has the same five female relatives with breast and ovarian cancer as Case 2 and the only difference is the lack of LCIS in Case 7. But the results of Case 2 with LCIS and Case 7 without LCIS are the same, 55%. In Case 3 with the Hormone Replacement Therapy, I chose current user. At least 6 years had been completed and at least 6 more were planned for the future. Twelve years of ongoing HRT makes no difference in breast cancer risk?

I inputted these figures numerous times, clearing out the results in between.

I did some other pre-menopausal trials, with and without LCIS. Again when female relatives had breast cancer the risk was lower than all healthy female relatives.

Maybe the browser produces these strange results? Maybe some of you can try these #s and see if your risk results are different. I don't want to think this model is flawed, but it is not looking too accurate right now.

Lea7777

I finally heard from the genetic counselor regarding these bizarre IBIS results. Again, to recap just what these bizarre results are--Women with LCIS have a lower lifetime breast cancer risk if many female relatives have breast and ovarian cancer than if those relatives were cancer-free per the Tyrer-Cuzick IBIS version 8 risk assessment model.

The model is linked here and this is the same model used with me by my genetic counselor. She inputted my data at her office. I inputted my data on my computer at this link at home. We came up with identical outputs.

http://ibis.ikonopedia.com/

The genetic counselor said she did not have time to recreate any of my inputs and outputs to verify my findings--that a lack of family breast cancer history for women with LCIS increases breast cancer risk--which I repeatedly stated were incomprehensible and indicated a flaw within the underlying equations or the software of this model.

She also said that the Tyrer-Cuzick IBIS model is used mainly to see if the patient has reached the lifetime breast cancer risk level of 20% or more, which is the level that indicates MRIs would be useful. She emphasized the model is not to be used for decisions on surgery.

The statement printed at the top of the page for this model also never mentions its usefulness for decisions regarding surgery and only mentions genetic counseling and testing. "This tool estimates the likelihood of a woman developing breast cancer specifically within 10 years of her current age and over the course of her lifetime. The tool is utilized to inform women and help support the decision making process for genetic counseling and testing."

One more set of 3 identical women with LCIS, except for family history. Again, the woman with no family history of cancer has the highest breast cancer risk. Cancer in the family LOWERS the breast cancer risk for women with LCIS per Tyrer-Cuzick IBIS, which cannot be correct.

poppyseed67

Thanks to Lea for posting the link to the Tyrer-Cruzick test (reposting for reference: ibis.ikonopedia.com )

I did a few variations on my own situation just to see what had the greatest influence on risk. It seems as if it's mostly LCIS that makes my risk skyrocket to 60+% lifetime risk. I also received a diagnosis of Atypical Ductal Hyperplasia, which boosts risk but not as much as when I include LCIS. Next I added in breast density (heterogeneously dense) and that raised it just a few percentage points, so it seemed negligible. Then I included my mother's breast cancer, which she believes was environmentally caused by taking hormone replacement....so, I don't consider it a significant factor in my case. Yet, even that didn't change the final risk number, which was 68%. That suggests that family history is less important than other factors when LCIS is present, or the model is faulty, or both.And, 68% is still a high number, although within the next 10 years the risk is much lower than that.

Lea7777

In the last week I had a geneticist and an oncologist (who work at the same place, so their outlook might naturally be similar) tell me that the Tyrer-Cuzik IBIS risk assessment model is used by them to see if the patient meets the 20% lifetime risk threshhold. At that point MRIs are suggested and I think insurance needs the 20% before they will pay. After that they don't pay much attention to the model.

poppyseed67

Thanks for sharing that context.

Lea7777

Spoke with a breast surgeon this week about my high lifetime risk with Tyrer-Cuzik IBIS model. She asked me, "Which verision did you use?" I replied, "8, the most recent." She said, "That one especially overstates risk for LCIS and hyperplasia. It is flawed." She used the f-word: Flawed. I can think of some other f-words I'd use to describe this model that delivers terribly inaccurate, inflated results for those of us who find ourselves with a diagnosis of proliferative breast disease.

If they can't even get their grammar straight (using then instead of than for questions on HRT), maybe it should be obvious that it is flawed. And it is not just a little off, which is to be expected, but grossly flawed to render the risk outputs meaningless for those with LCIS and hyperplasia.

What is also disappointing is that when I google words like "Tyrer Cuzik IBIS breast risk assessment overestimate overstate inaccurate too high" I get one study from way back in 2010 (if the recent V.8 is especially bad, that means improvements have not been made) and this very post. I think there should be more info out there on the inaccuracy of this model, which is the only one that can be used for those with LCIS. Numbers that are off by about 50% show a tremendous disrespect for the women who use this model, some of whom are paying a lot of money to the professional genetic counselors who employ it in their practice.

MelissaDallas

Predictors fo women with LCIS are essentially no more accurate than a coin toss

leaf

Its not just the women with LCIS. The modified Gail model with or without other risk factors like breast density was described as "Thus, for any given woman, the two models were better at prediction than a coin toss—but not by much." https://academic.oup.com/jnci/article/98/23/1673/2...

I think its important that medical providers give you some sort of idea about how good our models are - or, in the case of breast cancer prediction, how terrible they are. I don't think that is commonly done - probably most people are too scared, numb, or not intelligent enough to understand, at least intially.

Luckily, LCIS is not something like glioblastoma (the most common brain cancer) or pancreatic cancer.

Lea7777

I've heard you should always call heads, based on the construction of most coins, MelissaDallas.

I have very little experience with the Gail Model, but with Tyrer-Cuzik IBIS, there should be a disclaimer that has stronger language than "no responsibility for clinical decisions."

http://ibis.ikonopedia.com/ This is the online model for general public use and there is no disclaimer.

The feedback from the medical professionals* I have paid to see regarding IBIS has been:

--IBIS overstates. But when I insisted on a factor, such as by 10% or +/- X points, I was told the overstatement was anecdotal and could not be quantified.

--I don't have time to review your concerns with IBIS but at a conference we had in 2012 we learned about its shortcomings. IBIS model is used primarily to see if the threshhold of 20% is reached.

--IBIS accuracy was not addressed, but many of my other concerns were.

--IBIS lack of accuracy was acknowledged, but then it was suggested I multiply 2% times my estimated remaining years of life and the result will be close to what IBIS produced.

--IBIS accuracy or inaccuracy was not addressed directly, but the questions I asked produced answers indicating my risk was way less than the 63% from IBIS. IBIS model is used primarily to see if the threshhold of 20% is reached.

--Did not know about IBIS, but this was not a breast or a cancer specialist, so I would not expect familiarity.

--IBIS is FLAWED!

It may appear I am flogging this IBIS model to death and just piling on out of frustration. That is 10% true, but what I really want is to have something "out there" on the net, so when women with LCIS or breast neoplasia Google with questions on the accuracy of the IBIS model, they can find something specific.

You are right Leaf, that luckily LCIS is not like brain cancer! Maybe the experts cannot come up with an accurate model. That's ok. But don't pretend a coin flip is accurate 'cause that's a sham.

*I'm not dissing any of these professionals, except the one who did not have time. On the contrary, they have gone above and beyond in spending time with me and offering their advice. Excellent bedside manners, one and all, and I am lucky to be able to see them. The one exception was not a doctor.

Ok, done.

leaf

Its OK, Lea 7777. Its a shock to learn the data and/or models on which potentially life-altering decisions are made. Especially when its YOU.

You have EVERY RIGHT to be frustrated and angry.

Lea7777

Updating this feedback list

The feedback from the medical professionals I have paid to see regarding IBIS has been:

--IBIS overstates. But when I insisted on a factor, such as by 10% or +/- X points, I was told the overstatement was anecdotal and could not be quantified. (Nurse Practitioner from prestigious cancer center)

--I don't have time to review your concerns with IBIS but at a conference we had in 2012 we learned about its shortcomings. IBIS model is used primarily to see if the threshhold of 20% is reached. (Geneticist from prestigious cancer center)

--IBIS accuracy was not addressed, but many of my other concerns were. (oncologist from nearby hospital)

--IBIS lack of accuracy was acknowledged, but then it was suggested I multiply 2% times my estimated remaining years of life and the result will be close to what IBIS produced. (breast surgeon from prestigious cancer center)

--IBIS accuracy or inaccuracy was not addressed directly, but the questions I asked produced answers indicating my risk was way less than the 63% from IBIS. IBIS model is used primarily to see if the threshold of 20% is reached. (oncologist from prestigious cancer center)

--Did not know about IBIS, but this was not a breast or a cancer specialist, so I would not expect familiarity. (family doctor and nurse)

--IBIS is FLAWED! (breast surgeon who did my excisional biopsy, which receives compliments wherever it is seen in various examining rooms)

--IBIS is used only for insurance purposes and we always tell women this is not their risk (Nurse Practicioner at a high risk breast clinic)

LL99LL

Barbara1120 - I'm sorry I didn't reply to your question sooner. I've been away from this forum for awhile. I went to look for the "competing mortality" check box you asked about, and found that they've updated the risk calculator since I last used it. It looks like that question is no longer asked.

Lea7777 and leaf - thanks so much for your research and insights. It is so incredibly frustrating to not have anywhere near accurate info when making decisions about preventive mastectomy and hormonal therapy.

light1candle

A recent Australian clinical study of breast cancer risk in LCIS women, "Validation of the IBIS Breast Risk Evaluator for Women with Lobular Carcinoma In-Situ" finds the IBIS risk calculator flawed, with a tendency to overestimate risk. The study was published June 21, 2018.

https://www.nature.com/articles/s41416-018-0120-z

"Management advice for women with lobular carcinoma in situ (LCIS) is hampered by the lack of accurate personalised risk estimates for subsequent invasive breast cancer (BC). Prospective validation of the only tool that estimates individual BC risk for a woman with LCIS, the International Breast Cancer Intervention Study Risk Evaluation Tool (IBIS-RET), is lacking."

"...Informed decision-making would be facilitated by accurate personalised risk estimates for future invasive BC."

"Results

The mean observed 10-year risk of invasive BC was 14.1% (95% CI:11.3%-17.5%). IBIS-RET overestimated invasive BC risk (p = 0.0003) and demonstrated poor discriminatory accuracy (AUC 0.54, 95% CI: 0.48 – 0.62)."

"Conclusions

Clinicians should understand that IBIS-RET Version 7.2 may overestimate 10-year invasive BC risk for Australian women with LCIS. The newer IBIS-RET Version 8.0, released September 2017, includes mammographic density and may perform better, but validation is needed."

Lea7777

Thanks, Light1Candle, for posting. I'm glad the Aussies are getting after it.

""Management advice for women with lobular carcinoma in situ (LCIS) is hampered" Hampered? That's an understatement. 8.0 is what I used, and was told this version was especially flawed.

Mezza70

this has been one of the best reads for me so far .... recently diagnosed with LCIS with very high lymph node involvement and have been not able to research mine as it’s possibly a miracle at stage 111C to be isolated unbelievable really ... a pets scan will determine my longevity moving forward as we will determine whether it’s spread or not ... I won’t be scared into chemo unnecessary with no knowm benefits on any website worl

MelissaDallas

Mezza, you have invasive cancer, which is a completely different situation than what is being discussed about LCIS. If you have LCIS along with your invasive cancer, the invasive cancer "trumps" any precancerous or high risk conditions

Lea7777

Mezza, I see you are new to this forum. Lots of confusing and scary stuff at first. You may wish to post on the ILC or the Stage III Forum for information and responses more specific to your situation. Good luck to you.

Anonymous

I've been told repeatedly over the years, that risk with LCIS is about 20-30%.

Lea7777

Here's an additional piece of anecdotal evidence, acquired recently, from a breast doctor at a top medical facility for research and treatment. This doctor even knew the gentlemen who created the model on a first name basis. Conclusion: Tyrer-Cuzick is inaccurate, but the mid-60% lifetime risk the model produced will allow that MRI.