Luminal A vs. Luminal B stage 1 BC

mammalou · August 2016

I'm beginning to wonder if you can trust any of these tests.

Optimist52 · August 2016

This is kind of worrying to me because I feel the decision for me not to have chemo was based very much on my Oncotype result and I didn't have any other tests. Reading through this thread, I see that the mitosis rate is important and thankfully mine was 1. Does anyone know, is this the most important aspect of the histology of the tumour?

DaraB · August 2016

Optimist, my mitotic rate was also 1. But my PR+ was on 20% and Her2 was originally neutral with 2+, and it was the FSH that came in as negative. Plus, my surgical path report on the DCIS found with the tumor was grade 3. My MO said that sometimes one can have tiny cells that are actually pretty aggressive. The Prosigna is similar to the MammoPrint, but does not seem to be used at a lot of centers. I'm at City of Hope which is a research facility and my MO said they are finding that the test really targets early stage bc found to be intermediate with onco. There are so many new cutting edge practices like the new immunotherapies being tried, but I still think sometimes it's like we are still in the Dark Ages with a lot of medicine practices. As many have said, we just need to make the best decision we can and then not look back or question our decisions.

Lisey · August 2016

I look at additional genetic testing as a 'second opinion' for us intermediates (grays). The TailorX study has shown that low low oncotype scores under 11 can safely go without chemo. which makes me believe the mammaprint was pulling from the higher lows (11-18) when saying they were high risk. However, mammaprint said I was a low score and my mitosis on both biopsy and tumor path was a 1. I'm almost comforted that mammaprint found me lowrisk, when it clearly pulls more lower risk oncotypes to highrisk, so I tend to trust that result more. The only reason mammaprint isn't used more widely is because the technology for formalin tissues wasn't developed until just a few years ago and it doesn't give a prognositc value for how much chemo would help like the oncotype does. I think if it did, it would be the better test than the oncotype.

Having said all this, I HIGHLY recommend anyone who has a score higher than 11 and lower than 30 to get a second opinion genetics test during their decision making window. I'm not saying this for people who've already decided, but for those who may be viewing this thread. Clearly, the tests conflict a lot, and since mammaprint pulled quite a few lower RS oncotypes into the high risk category, (18 out of 52 low RS's were switched to high risk via mammaprint/blueprint).. and it divided up the intermediates rather evenly between low and high risk (14 out of 25 designated low risk), it would be good to know where you fall. Dara and I were initially told the same thing, now our paths diverge thanks to these 2nd opinion tests.

DaraB · August 2016

I agree Lisey. I think these second tests or opinions really do only apply to the intermediates. My MO voiced the same thing.

BarredOwl · August 2016

Hi:

As I mentioned earlier in this thread, it is important to be cautious in drawing conclusions about treatment decisions or testing from a single report. The concerns may be even greater with an abstract, since many abstracts are not peer-reviewed, can be preliminary in nature, and do not include much in the way of background or discussion that may provide useful context, for example pointing out potential biases or deficiencies in study design, mentioning conflicting studies, or discussing important limitations or caveats about results.

The Maroun (2015) abstract describes a study that included 86 patients. A very small study at a single center may have significant limitations due to the small sample size and study design, so some of the findings might not be borne out in a larger or more rigorously designed study. The Maroun abstract also has obvious typographical errors, stating in the results that there were "50 ODX low RS cases", while the table says there were 18 + 33, which would be 51, while the "Grand Total" is listed as 52. This type of quality issue exemplifies the impact of a lack of peer review and preliminary analysis.

Maroun (2015) is not the only head-to-head study available. Here is an earlier and larger one, also an abstract for what it's worth, in a group of 437 patients, with some different findings:

Dabbs (2014): http://meetinglibrary.asco.org/content/128573-144

(There is an accompanying poster presentation, but I cannot access it.)

For example, if I am reading it correctly, in this 437 patient study, among 136 "High Risk" MammaPrint patients, only 16 are said to have Low Risk Recurrence Scores by Oncotype (exact scores not stated). Long-term follow-up would be required to figure out which test was more predictive (or neither) in those particular patients. More information and much larger, long-term studies would be needed to identify factors that might guide test selection in the hormone-receptor positive, HER2-negative setting.

Some of the commentary about this abstract noted: "Oncotype DX was validated in patients treated with tamoxifen, while MammaPrint is intended to be used to gauge risk of recurrence for patients ahead of systemic treatment." Perhaps it is not realistic to expect total concordance between such tests. Indeed, a more recent comparative prospective study of such multiparameter tests also found differences.

Unfortunately, none of these multiparameter tests has been shown to be a perfect predictor of individual outcome. However, the available clinical validation studies suggest some "prognostic" advantage of these tests over the sole use of traditional clinico-pathologic methods of estimating average distant recurrence risk, particularly in node-negative ("N0") groups. The types of tests have different designs, different outputs, and varying degrees of validated clinical significance in particular test groups. They are what we have today and can be used to inform recurrence risk and risk/benefit analyses.

BarredOwl

Lisey · August 2016

Barred Owl, you need to be paid to educate us on this site. You find some amazing abstracts.. I was looking for a larger study, but couldn't find a recent one that invoived non-fresh samples for Mammaprint.

Here is my report... I guess if there's an intermediate level on the mammaprint I have it... ha! See how close I am to 0? Not very reassuring, but it is what it is. Anyway, if you look at the variables from the sample sheet on Agendia's webpage it does say low risk averages 10 year: 10% (95% CI: 4% - 15%) - which does align to the oncotype's score of 10% for me if I take Tamox for 10 years.

KBeee · August 2016

I agree with Barred Owl's statement that no test is 100%. These tests make predictions based on the genetics of the tumor and other factors, but no test can take into account environment and how it interacts with the genetics of the tumor and the genetics of the individual. It is one piece of information (an important one), but in making the decision for chemo, (in my unmedical opinion) one needs to take into account these tests, family history, pre-existing medical conditions, environmental factors, etc. Example... I had an oncotype of 16, but my MO and I decided that I had no pre-existing health conditions to put me at risk, I preferred to be aggressive, I have a horrible family history of people having cancers that have not behaved, I was exposed to a lot of x rays as a child (had every bone in my body x rayed a few times), am in a high risk profession (firefighter), and may have been exposed to unknown factors (of the 5 females who grew up on my block, 4 have had breast cancer in their 30s or 40s). Perhaps because of bad luck or perhaps because of one of the other factors, despite clear margins with surgery, being node negative, doing chemo and tamoxifen, exercising daily, eating a vegetable based diet, and being thin, I recurred within 18 months in 2 different places. It just happens. I do not think the tests were wrong or invalid, I simply think more things come into play than we can account for with any one test. I would not change a thing with my treatment plan if I had to do it all over again. That's why everyone needs to work with your MO and take into account as much as you can. And once you make a decision, stick to it with no regrets, no matter what happens. We can what-if things forever, but at the end of the day, none of us has a crystal ball, and we just have to make the best decision we can.

Hopeful82014 · August 2016

Well said, Karen. Thank you. (And many thanks to Barred Owl as well. Your posts are always well-reasoned and illuminating.)

DaraB · August 2016

I think all of us would certainly prefer to have a crystal ball that can definitively say what we should each do, but unfortunately, I haven't come across one yet. You're all correct that no test is 100%, in fact no several tests are perfect. We all gamble and weigh the odds of results as to whether we will be in one group of recurrence or another. I so agree that we just have to listen to our team, make a choice, and then not look back or second guess. I, too, just on the thank you wagon Barred Owl for the great information, articles, and perspective you offer. Thanks all...

BarredOwl · August 2016

Hi Lisey:

Re: " . . on Agendia's webpage it does say low risk averages 10 year: 10% (95% CI: 4% - 15%) - which does align to the oncotype's score of 10% for me if I take Tamox for 10 years."

It is generally consistent, but does not exactly align. While it does not affect your decision, I think these may be different types of recurrence risk estimates (i.e., one with no treatments and one with 5-yrs tamoxifen). Here is some text from the professional pages on Agendia's web site that suggests that the 10% (95% CI, 4% - 15%) 10-yr recurrence risk for the "low risk" cohort is without any adjuvant treatment (either hormonal therapy or chemotherapy):

http://www.agendia.com/healthcare-professionals/breast-cancer/mammaprint/

(To locate this text, scroll down to the orange "CLINICAL UTILITY" panel, click (+) to expand, then scroll to the bottom of the section)

"TRANSBIG STUDY (2)

Published in 2006 by Buyse et al., the independent study conducted on behalf of the TRANSBIG consortium which consisted of 302 patients, found that MammaPrint can stratify patients into a binary risk classification of either Low Risk or High Risk, with a statistically significant difference in the probability of metastasis free survival at 10 years.

Results

Low Risk patients have a ∼10% (95%, CI 4-15) chance of cancer recurrence within 10 years without any adjuvant treatment (either hormonal therapy or chemotherapy)
High Risk patients have a ∼29% (95%, CI 22-35) chance of recurrence within 10 years without any adjuvant treatment (either hormonal therapy or chemotherapy)"

This is the Buyse (2006) reference cited, a study in "untreated patients with node-negative breast cancer":

Buyse (2006): http://jnci.oxfordjournals.org/content/98/17/1183.full.pdf

MammaPrint reports may include other types of risk estimates as well.

Meanwhile, the estimated 10-yr risk by Oncotype (from the first section of the node-negative report, Graph 1, and printed to the left of Graph 1) is a "10-yr risk of distant recurrence after 5 year of tamoxifen", based on a study of patient from the NSABP B-14 study.

Paik (2004): http://www.nejm.org/doi/pdf/10.1056/NEJMoa041588

Patients should be cognizant of possible differences in the types of recurrence risk estimates provided to them (e.g., 5-yr? 10-yr? Risk of distant recurrence? Other? Recurrence risk with 5-yrs tamoxifen? Recurrence risk with no adjuvant chemo and no hormonal therapy? Other?)

BarredOwl

Lisey · August 2016

Adding this - pasted from another thread:

Hi Everyone. I called and spoke to the chief Pathologist at Agendia. She told me they test all samples TWICE to confirm the status. (including mine which was only 0.009 outside of the borderline) If there is a discrepancy, they test it 2 more times after that. She confirmed to me I'm in the low zone (twice) and while she agreed that logically, my numbers are 'more risky' than someone not so close to the zero, but she said the numbers they give are AVERAGES for the entire spectrum of Low Risk with the variances given. She also mentioned the numbers are for no chemo or hormonal therapy, so there's that.

So in a way, our assumption that those of us closer to zero are right to think we have more risk (in the Low Risk group) than someone farther from zero. It would be inverse for the High Risk group, is accurate.. but they don't have a way of quantifying that difference so they just use averages. I think it said 1.3% (0-3.1%) for the first 5 years for Low Risk. FFPE samples are only for 5 years at this point, since the technology is more recent. She also said that Agendia has phased out their Targetprint as well because they are doing the 'mammoprint/blueprint' suite for everyone instead.

I am glad I was tested more than once and both times was in low risk. But my risk is probably more in the 3.1% rather than 1.3%.. but I"m good with no chemo.

BarredOwl · August 2016

Hi Lisey:

It is an important point that the recurrence risk estimates are based on risks determined for a group and are expressed as averages with confidence intervals ("CI").

Below are links to selected references regarding the MammaPrint test, which may be of interest to those receiving the test.

From the other thread:

The MammaPrint Index ("MPI") number appears to be generated from the patient's microarray data by a software program which uses an algorithm or "classifier". The program takes the mRNA expression data from 70 genes, determines the "sample expression profile", and assesses its degree of similarity to a standard low risk expression profile (the standard or low risk "template" profile is a "mean expression profile of 44 tumors with a known good clinical outcome"). The degree of similarity to the standard low risk profile is expressed as the MPI value between -1.0 and +1.0, with +1.0 being the highest degree of correlation. After this correlation step, the calculated MPI value is used to classify one as either "high risk" or "low risk" (relative to the "clinical classification threshold" of zero), within the accuracy of the test.

The MPI value reflects the degree of "correlation" between the sample profile and the low risk standard or template profile. The output of the test is the final risk classification of high risk orlow risk. Various clinical validation studies were performed to assess the "prognostic" ability of the test. They determined various types of average recurrence risk among a group of patients classified as "high risk" and among a groups of patients classified as "low risk". Thus, the average recurrence risks provided are based on risks determined for the high risk or low risk group as a whole.

It is my understanding that the clinical validation studies did not look at how recurrence risk might vary within the low risk group according to MPI value. They did not determine how risk might vary within the high risk group according to MPI value. Thus, whether positive "low risk" MPI values closer to the clinical classification threshold of 0 are or are not associated with greater risk compared with those closer to +1.0, or whether negative "high risk" MPI values closer to 0 are or are not associated with lower risk compared to those closer to -1.0, is not known. One should be cautious about drawing conclusions in the absence of data. As the Agendia pathologist stated, "they don't have a way of quantifying that difference so they just use averages." Again, they don't have a way of quantifying the difference, because it has never been determined in a clinical study. It may not behave intuitively, or as a continuous variable.

They designed a binary test (high or low). They have not looked at the recurrence risks associated with individual MPI values, and therefore, they have not established differences in recurrence risk (or the size of such differences, if any, or the statistical significance of such differences, if any) between specific MPI values within the same risk category (e.g., Low risk, MPI +0.3 versus Low risk MPI +0.9). In my layperson's understanding, hypothesizing statistically significant differences in recurrence risk between different MPI values within the same risk category where none has been measured in a clinical study would be speculative.

BarredOwl

Selected References re MammaPrint:

van't Veer (2002): http://www.nature.com/nature/journal/v415/n6871/pdf/415530a.pdf

van de Vijver (2002): http://www.nejm.org/doi/pdf/10.1056/NEJMoa021967

Buyse (2006), TRANSBIG: http://jnci.oxfordjournals.org/content/98/17/1183.full.pdf

Knauer (2010): http://dev.agendia.com.previewdns.com/wp-content/uploads/2013/05/Knauer_2010_Breast_Cancer_Res_Treat.pdf

Drukker (2013), RASTER: http://onlinelibrary.wiley.com/doi/10.1002/ijc.28082/epdf

Piccart (2016), MINDACT: This EORTC news feature includes the text of the AACR 2016 abstract re the primary analysis from the MINDACT trial of MammaPrint:

http://www.eortc.org/news/mindact-mammaprint-genetic-test-can-reduce-use-of-post-surgery-chemotherapy-among-early-stage-breast-cancer-patients/

KNardo88 · August 2016

Confirmed just yesterday with my MO that I am in fact Luminal B... 20% ER+, PR-, HEr2-, BRCA1+, Stage 1A IDC, 0.9cm, Grade 3, Node Negative, LVI Absent. Initial biopsy taken by clinic that does report ki67 showed 80-90%, however, 2nd opinion (Dana Farber) does not report ki67 at all due to unreliability. Underwent BMX and 3/4 of the way through chemo (ACx4, standard dose) and then I will be going on Tamoxifen for 10 years. My MO told me not to worry about Luminal B, as that is what the chemo is for, and all my other characteristics and pathological findings give him confidence that I'll never see this beast again. He did say that tamoxifen can be less effective in luminal B cases, but certainly not ineffective. Hope this helps put some minds at ease?

Lisey · June 2018

I"m bumping this thread because I feel it's full of great info and really important. FYI: 2 years after posting this thread, I'm doing great and still on Tamoxifen. No issues so far. Oh and Agendia has never sent me a bill... still waiting.

Luminal A vs. Luminal B stage 1 BC

Comments

"TRANSBIG STUDY (2)

Categories