Lani - Your Opinion on a Paper at SABCS

[Becky]

Dear Lani

Did you see the presentation on ER+/PR- molecular assay paper given on Friday? Naturally, Jean and I attended as there is not much information on this pathology.

Unfortunately, no Her2 marker (or any other marker besides ER and PR) were tested but I found the work fascinating in that when they looked at assays of ER+PR+ (++) versus ER+PR_ (+-) and ER-PR- (--) they first determined what the standard assay looked like molecularly for ++ and --. Then they put the samples in their correct column via what the tumors stained for and found that some ++ actually molecularly looked like +- or even (rarely) -- and vice versa. Then they moved the +- out of the +- column into the ++ or -- column as to what they looked like. Then it left the true +- women (which really are their own class).

It was interesting work. Of course they mentioned that +- women have the worst pathology (even over --) but when questioned that Her2 status wasn't looked out, the presenter admitted that most of the +- women were Her2+ (therefore the worst outcome for when this work was done a few years ago). And why there would be tamoxifen resistance in these women.

Jean and I later discussed this in light of women on this board who are ++ but do not respond to endocrine therapy and the -- who do not respond well to chemo (and that there cancer may actually look like and behave like the opposite).

Did you see this and what did you think?

[Hopeful]

Becky,

I read the abstract for this several times, and thank you very much for explaining the talk that went along with it, because it is difficult to "tease" such information out of a few paragraphs (i.e., the concluding sentence, "breast cancer subtypes by IHC may not fully overlap with those defined by gene expression profiling" does not convey nearly the depth of information contained in your post.)

I also saw this abstract: http://www.abstracts2view.com/sabcs/view.php?nu=SABCS07L_999930 about "Biomarking the estrogen responsiveness of bc," wherein it is stated, "a global index of dependence on estrogen (GIDE) which integrates the transcriptional changes that occur with estrogen deprivation varies more than 30-fold among ER+ tumors . . .Despite compelling preclinical data, markers such as PgR and HER2 do not appear to define patients with a particularly high relative benefit of an aromatase inhibitor over tamoxifen. Further study of the molecular changes that occur with estrogen deprivation and association of these with clinical outcome may be expected to identify the hallmarks of breast cancer that are supported by estrogen stimulation."

This seems to piggyback on what you are asking Lani to comment on, I think - i.e., that having a positive or negative hormone receptor is not necessarily an indication of response to endocrine therapy, even though the therapy does what it is supposed to do (in the case of AI's, almost total ER deprivation) because the bc itself "looks and acts" like its receptor-opposite.

These seem to me to be profound findings that are not getting much airplay.

I will be interested in Lani's response, as well.

Hopeful

[alw]

Thanks for posting this, Becky.

I'm interested in this as well, and read the submission. My sister's original pathology was ER-/PR-. After neoadjuvant chemo, the mass removed at mastectomy was ER+(75%)/PR-.

[Lani]

Please note that the study you are interested is by the group including CKent Osbourne and Rachel Schiff that I believe to be doing some of the most interesting work on the understanding of the pathways involved in her2+ breast cancer. They have written a lot on how it is influenced by/influences the ER pathway--they along with Graciela Arpino published the original work on how/why the ER+PR- subset of tumors might benefit more from AIs than tamoxifen.

ER status is notoriously inexact unfortunately. And it looks like the results one gets with other methods one looking at ER and PR content differ markedly from that obtained using the traditional Allred method.

This paper shows that they are trying to dissect out the cherries vs the peaches vs the grapes in the fruit=salad that is called breast cancer, which so many other researchers lump together when trying to figure out cause, prognosis and best treatment.

I believe it is only by dividing and then subdividing out these types that we will find the cause (could be different for each), prognosis (we know these vary widely by subtype) and best treatment for (which 3 or 4 pathways are most important to the survival of any one breast cancer tumor, but some subgroups may share these).

That said, it is probably the molecular pathways which drive the tumor, rather than how they look on microscopy or how the inexact "art" of staining (and then reading the results) which can be modified by multiple factors (%ER+ can vary with stage of the menstrual cycle, %ER+ can be obscured if methylene blue was used at the time of the sentinel node biopsy, etc).

Why tumors do or do not respond to AIS, tamoxifen and/or chemo appears to be multifactorial and there is de novo (immediate) resistance AND acquired (developed) resistance to contend with.

Drs. Osbourne and Schiff have done some interesting work on the fact that there are two kinds of ERalpha--one which sits in the nucleus and works slowly and one which sits on the cytoplasmic (outside) membrane of the cell and works quickly and there is ERbeta and there are two types of PR, PR1
and PR2 involved to complicate the picture further. Drs. Schiff and Osbourne have studied the crosstalk mechanisms between her2 and ER and these are multifactorial as well.

It is only with enough tissue samples and highly powered computers and a lot of brainpower, will, passion and "propeller grease" (Tom used to call me propeller head!) that this puzzle will be worked out. I have gone to lectures at Stanford by mathematical modellers of networks trying to simplify the way the computer structures its analysis of this complicated web that.

At SABCS Dr. Yigel Yarden, who performed some of the seminal work on her2, spoke about this network analysis very eloquently.

From the abstract it seems they believe their work points toward classifying patient's tumors by assessing ER and PR by methods other than the Allred method and that it may be necessary to subclassify them as well.

I have a suspicion from my readings that much of the effect of antihormonals may be via their effect on cyclooxygenase (see the paper I posted today by a researcher at Ohio State who has published a lot of great papers on making a tissue specific drug which works as an antiaromatase only in the breast (bypassing lots of sideeffects hopefully) by modifying celecoxib (celebrex). My suspicion is likely to be entirely wrong, but I would suggest you read the papers by Bruggemeister (sp?) from OSU and see what you think.

Estrogen seems to have effects besides stimulating growth of breast cancer cells-- including angiogenic effects, whether directly or indirectly.

Similary NSAIDs have both antiangiogenic and antiaromatase effects. Those two facts may be relate, they may not.

The point is we only have insight into a tiny corner of an enormous puzzle.

That is what makes donating the specimens to help speed things up so important. Many poster presenters based their results on 20-60 tumors and felt lucky they had so many. Of those very few were ER+PR+ or ER+PR-. Since this group thinks there are at least 3 subgroups of her2+ER+PR-, how are we ever going to tease out the answer without more data?

Hope I answered your question...off my soapbox for now!