ECE and Z11. Has anything changed?

[Mandelin Rain]

50 year old. Seeing a 1/1 SLN with microscopic ECE for radiation. No discussion of dissection by surgeon. Mammaprint low-risk so no chemo (gulp). Is this someone you just treat without discussing dissection? Just curious of opinions.

---

Members don't see this ad.

 

[thecarbonionangle]

Based on z11, no omitting surgery as ECE was exclusion criteria. Based on AMAROS, I believe ECE was not an exclusion criteria but all XRT patients got a SCV field. Similar outcomes. If clinically node negative, would feel comfortable with XRT after discussion.

 

[medgator]

Based on z11, no omitting surgery as ECE was exclusion criteria. Based on AMAROS, I believe ECE was not an exclusion criteria but all XRT patients got a SCV field. Similar outcomes. If clinically node negative, would feel comfortable with XRT after discussion.

There's ece and then there is ece. Microscopic ece was allowed in z11 iirc

 

[Palex80]

I would probably treat the axilla and SCV.
Like pointed out in the other thread, I like to put the data in the nomogram to see the risk for non-sentinel positive nodes.

 

[Mandelin Rain]

It's a whole new world of pre/peri- menopausal women with T1c and "at least" one node positive with extracapsular spread not getting chemo. Kind of freaks me out.

 

[evilbooyaa]

Treat axilla and SCV. Is there SLNB era stuff saying ECE is bad in breast cancer? I know we all freak out about it in H&N but in a SLNB'd axilla with ECE I'm not sure that I do anything differently.

Reviewing the Z11 paper, women were ineligible if they had matted nodes (N2) or gross extranodal disease (meaning cN1?) I don't see anything that says all ECE was excluded.

 

[medgator]

Treat axilla and SCV. Is there SLNB era stuff saying ECE is bad in breast cancer? I know we all freak out about it in H&N but in a SLNB'd axilla with ECE I'm not sure that I do anything differently.

Reviewing the Z11 paper, women were ineligible if they had matted nodes (N2) or gross extranodal disease (meaning cN1?) I don't see anything that says all ECE was excluded.

It wasn't. They had to be clinically suspicious at the time of surgery. Microscopic ece was fine

 

[scarbrtj]

Probably WAY more undiscovered microscopic ECE than discovered microscopic ECE, perhaps 100:1? Of all the thousands and thousands of fields-of-view of the nodal surface, the pathologist caught that one where there was ECE. A typical FOV is maybe, what, like 0.5x0.5mm on a slide view. And a 1cm node has a surface area of 3.14 sq cm. So you would need ~125,000 ~1250 fields of views, or slide views, to sample the whole nodal surface. Whatever the number might be, it won't ever be 100% sampling. So microscopic ECE is not ECE in the typical way we think about things from our rad onc perspective IMHO (ie there is not appreciable disease left in the axilla after surgery with microscopic ECE). That's first. Second is, the med oncs are definitely going molecular vs clinical in chemo decisions (in these "borderline" cases) nowadays and there is good data to back this up.

 

[nkmiami]

Treat axilla and SCV. Is there SLNB era stuff saying ECE is bad in breast cancer? I know we all freak out about it in H&N but in a SLNB'd axilla with ECE I'm not sure that I do anything differently.

Reviewing the Z11 paper, women were ineligible if they had matted nodes (N2) or gross extranodal disease (meaning cN1?) I don't see anything that says all ECE was excluded.

I dont know what I would. You can make a lot of good arguments either way. I dont think patient needs AND.

 

[Pointless]

Had the same patient recently. The discussion that I had with her is that her axilla should probably be addressed. There are data to suggest that ENE is predictive of higher nodal burden and since AMAROS tells us that nodal RT is equivalent to surgery for clearance of the axilla (outside of gross clinically positive nodes), that was my recommendation. I'm not to a point in my practice where I would feel comfortable going full Z11 (tangents only) with 1/1 with ENE. TBH, even without the ENE I would go high tangents with any SN positive.

 

[Palex80]

Second is, the med oncs are definitely going molecular vs clinical in chemo decisions (in these "borderline" cases) nowadays and there is good data to back this up.

Indeed, but the patients in the Oncotype & Co – trials probably had „standard“ axillary treatment, since they were enrolled in the trials prior to Z011 results becoming mature.

Perhaps we are deescalating a little too fast...

And as I pointed out in the other thread, Z011 had several flaws, both in its design (no strict RT recommendations) and the patients that were recruited (a lot pN1 mic). It was a surgical trial and in my view rather sloppy conceived and carried out.

 

[scarbrtj]

It was a surgical trial and in my view rather sloppy conceived and carried out.

Not being sarcastic, this is the truth: you must be very distressed that this single sloppily conceived trial has completely changed the standard of care in breast cancer over the last decade. Ten years ago, not doing ALND after SNB+ was considered malpractice; now, it's a standard of care. All this solely on the basis of Z0011. EDIT: It distresses others too.

 

[thecarbonionangle]

Had the same patient recently. The discussion that I had with her is that her axilla should probably be addressed. There are data to suggest that ENE is predictive of higher nodal burden and since AMAROS tells us that nodal RT is equivalent to surgery for clearance of the axilla (outside of gross clinically positive nodes), that was my recommendation. I'm not to a point in my practice where I would feel comfortable going full Z11 (tangents only) with 1/1 with ENE. TBH, even without the ENE I would go high tangents with any SN positive.

Most of the patients on z11 did not even get XRT “per z11”. They were supposed to get whole breast only but many got at least high tangents or even comprehensive nodal XRT.

 

[scarbrtj]

This woman has a <1% chance of axillary recurrence REGARDLESS of wonky, inside baseball radiation field considerations. (Was the tangent high? How high? Who defined how high? What is low? Is there "too low?" Please show your work.) We are moving no needles; all our axillary discussions about breast cancer patients with small tumors and single positive SLNs are truly sisyphean.

Radiotherapy was "all over the place" in Z0011. But... so? Out of ~900 women, ~450 got ALND and ~450 got SNB only; 2/450 recurred in the axilla, and 5/450 recurred in the axilla, respectively. Seven out of ~900 patients recurred in the axilla in the whole dang trial. Not much to work with in terms of post hoc mental masturbation. The biggest protocol deviation in Z0011 was no XRT at all: 107 out of 335 protocol violation patients had *no* RT, the other 228 did. And in that group, 1/107 no RT patients had an axillary recurrence; 4/228 who did get RT had an axillary recurrence.

Z0011 is only confusing or sloppy or worrisome until realizing that, no matter what, even when the approaches were as crappy as they come, axillary recurrences by themselves are rare, in the ~1% range. And the ALND did nothing for or against that natural history. More surgery, more radiation, not the answers to biological breast cancer questions. Breast cancer's biology drinks your high tangent milkshake.

 

[Pointless]

That is a fair point, but we don't really know what the numbers would have been if everyone truly got tangents only because, as you say, very few were actually treated per protocol. lymphedema rates after snl biopsy only are very low and high tangents don't add much additional toxicity. You make a compelling argument but my practice remains the same... and I'll sleep better for it at night.

 

[Palex80]

Quoting the axillary recurrence rates of Z0011 and saying that these rates apply to any "single node positive breast cancer patient without axillary dissection" is simply wrong. 38% of Z0011 patients only had microscopic involvement in their lymph nodes. This population alone is a "low risk" population, which certainly does not needs further axillary treatment. Including patients like these in the trial certainly "contaminated" the results.

And concerning the other retrospective trial Scrbtj quoted: All trials with follow-up times of around 5 years are inadequate to provide statements on recurrence rates. We know very well that in luminal A cancer recurrence rates can often pick up after antihormonal treatment has been terminated. This has been shown as well in the trials testing omission of post BCS radiation therapy to the breast. One does need 10 years of follow up to make a definitive statement.

There are subgroups that clearly do not benefit from axillary RT and to which Z0011 recurrence rates may very well fit. There are however other subgroups of patients that are at higher risks for recurrence. Z0011 was not powered to demonstrate differences between subgroups.

Concerning data quality in the trial, just have a look at the publication of the trial data.
Data is missing for tumor grading in more than a quarter of all patients (how can you not have that on file, since it's a surgical trial?), they are even missing AGE in 2% of all patients...

Effect of Axillary Dissection vs No Axillary Dissection on 10-Year Overall Survival Among Women With Invasive Breast Cancer and Sentinel Node Metastasis: The ACOSOG Z0011 (Alliance) Randomized Clinical Trial

The results of the American College of Surgeons Oncology Group Z0011 (ACOSOG Z0011) trial were first reported in 2005 with a median follow-up of 6.3 years. Longer follow-up was necessary because the majority of the patients had estrogen receptor–positive ...

www.ncbi.nlm.nih.gov

I have grave concerns regarding quality of reported data.

 

[Pointless]

I continue to be amazed and frustrated at the lack of quality and direction of the breast radiation oncology literature. We do so many things in breast treatment that are archaic and not at all evidence-based. Outside of hypofractionation, which seems to be a clear home run as it pertains to patient convenience (and likely toxicity), we have made frustratingly few advances. To wit, breast remains essentially the only disease site in which it is still quasi-acceptable to not contour the target structures! This thread has thankfully not been afflicted by all the doomsaying of the other threads, but this IS one of my concerns with our field. We are purportedly an evidence-based field led by the highest and brightest of academicians, yet we can't definitively answer simple questions such as the OP's clearly and concisely.

 

[scarbrtj]

Quoting the axillary recurrence rates of Z0011 and saying that these rates apply to any "single node positive breast cancer patient without axillary dissection" is simply wrong. 38% of Z0011 patients only had microscopic involvement in their lymph nodes. This population alone is a "low risk" population, which certainly does not needs further axillary treatment. Including patients like these in the trial certainly "contaminated" the results.

Let's assume for sake of argument that all 5 of the ~450 SNB-only Z0011 patients with LRR recurrence were in the 62% of patients who had gross involvement of nodes. Eliminate the 38% with microscopic, so now let's multiply: (100/62)*5 = 8. By eliminating all microscopic patients from the SNB arm, but keeping 38% microscopic in the ALND arm, a big trial imbalance in favor of ALND, this would have added 3 LRRs to the SNB arm, worst case, and done this to the statistical outcome:

So, unconvincing to me. And, as mentioned, the arms were well balanced, for the bad and the good.

And concerning the other retrospective trial Scrbtj quoted: All trials with follow-up times of around 5 years are inadequate to provide statements on recurrence rates. We know very well that in luminal A cancer recurrence rates can often pick up after antihormonal treatment has been terminated. This has been shown as well in the trials testing omission of post BCS radiation therapy to the breast. One does need 10 years of follow up to make a definitive statement.

There are subgroups that clearly do not benefit from axillary RT and to which Z0011 recurrence rates may very well fit. There are however other subgroups of patients that are at higher risks for recurrence. Z0011 was not powered to demonstrate differences between subgroups.

Concerning data quality in the trial, just have a look at the publication of the trial data.
Data is missing for tumor grading in more than a quarter of all patients (how can you not have that on file, since it's a surgical trial?), they are even missing AGE in 2% of all patients...

Effect of Axillary Dissection vs No Axillary Dissection on 10-Year Overall Survival Among Women With Invasive Breast Cancer and Sentinel Node Metastasis: The ACOSOG Z0011 (Alliance) Randomized Clinical Trial

The results of the American College of Surgeons Oncology Group Z0011 (ACOSOG Z0011) trial were first reported in 2005 with a median follow-up of 6.3 years. Longer follow-up was necessary because the majority of the patients had estrogen receptor–positive ...

www.ncbi.nlm.nih.gov

I have grave concerns regarding quality of reported data.

See, I knew you were distressed OK. But: all the data show regardless of any "grave concerns" (about any minor factoid like microscopic involvement or no, etc.) that T1-2 SNB+ patients without ALND, and also regardless of XRT approach, have an axillary recurrence risk of <1%. I kept saying around 1% risk above just to be non-confusing/quote the Z0011, but these trials (including Z0011) don't tease out axilla alone, they just lump everything as LRR ("Locoregional recurrence was defined as a tumor in the breast or in ipsilateral axillary, internal mammary, subclavicular, or supraclavicular nodes") which was ~5/450 in Z0011 SNB-only patients. In fact axillary recurrences rates are much closer to 0% than 1% ("no isolated axillary recurrences" in hundreds and hundreds of undissected, unirradiated-in-the-armpit patients). So I don't really care what "recurrence rates" are, like distantly or in the breast, when we have these how-to-treat-or-handle-axilla discussions.

So, show me anywhere, any data, that axillary recurrence rates are greater than 1-2% for T1/2 N1 patients in the modern era, because all the data I have seen says the rates are so low that no matter what approach any individual practitioner decides to take re: the axilla in this setting it will not appreciably affect therapeutic outcomes. Then the question becomes: what toxicity, and amount of toxicity, do you want the patient's axilla to countenance.

 

[nkmiami]

Not being sarcastic, this is the truth: you must be very distressed that this single sloppily conceived trial has completely changed the standard of care in breast cancer over the last decade. Ten years ago, not doing ALND after SNB+ was considered malpractice; now, it's a standard of care. All this solely on the basis of Z0011. EDIT: It distresses others too.

B04 originally proved radiation to a clinically negative, surgically unaddressed axilla had very low failure rates? been a while since i looked at it

 

[scarbrtj]

I continue to be amazed and frustrated at the lack of quality and direction of the breast radiation oncology literature. We do so many things in breast treatment that are archaic and not at all evidence-based. Outside of hypofractionation, which seems to be a clear home run as it pertains to patient convenience (and likely toxicity), we have made frustratingly few advances. To wit, breast remains essentially the only disease site in which it is still quasi-acceptable to not contour the target structures! This thread has thankfully not been afflicted by all the doomsaying of the other threads, but this IS one of my concerns with our field. We are purportedly an evidence-based field led by the highest and brightest of academicians, yet we can't definitively answer simple questions such as the OP's clearly and concisely.

And we never will. This has become so obvious to me that I have stopped worrying about the question(s), lumped them all in a container called "Dumb Questions," and stopped wasting my time. No other physician in oncology is worrying about these "Dumb Questions." Examples: how high is your breast tangent? Should I treat the axilla/treat differently with microscopic ECE in a single node? How many nodes must be dissected for an adequate axillary dissection? If not enough nodes have been dissected in the axilla, how much radiation should I add to the axilla? Etc etc. These questions are important inside our own little house only, and anybody who thinks he or she has the "right" answer is delusional. I mean I love a long-winded, I'm-so-smart walk down data lane as much as the next guy but c'mon folks. The axilla is super, super low yield. There can be no right answer when you can't even show how an answer is wrong.

 

[Pointless]

And we never will. This has become so obvious to me that I have stopped worrying about the question(s), lumped them all in a container called "Dumb Questions," and stopped wasting my time. No other physician in oncology is worrying about these "Dumb Questions." Examples: how high is your breast tangent? Should I treat the axilla/treat differently with microscopic ECE in a single node? How many nodes must be dissected for an adequate axillary dissection? If not enough nodes have been dissected in the axilla, how much radiation should I add to the axilla? Etc etc. These questions are important inside our own little house only, and anybody who thinks he or she has the "right" answer is delusional. I mean I love a long-winded, I'm-so-smart walk down data lane as much as the next guy but c'mon folks. The axilla is super, super low yield. There can be no right answer when you can't even show how an answer is wrong.

I see what you're saying but I am not as nihilistic about management of the axilla as you are. As ALND has become increasingly unused, it falls to use to manage the lymph nodes. I think we should take that seriously. Nodal failures are a thing, crappy Z0011 data notwithstanding.

 

[Palex80]

Let's assume for sake of argument that all 5 of the ~450 SNB-only Z0011 patients with LRR recurrence were in the 62% of patients who had gross involvement of nodes. Eliminate the 38% with microscopic, so now let's multiply: (100/62)*5 = 8. By eliminating all microscopic patients from the SNB arm, but keeping 38% microscopic in the ALND arm, a big trial imbalance in favor of ALND, this would have added 3 LRRs to the SNB arm, worst case, and done this to the statistical outcome.

Now you'd probably have to weigh in the other 30% that got some kind of axillary RT in the "SLNB-only arm"...

Z0011 was underpowered.
Look at AMAROS: The axillary recurrence rates there was 1.19% with RT vs. 0.43% with ALND and yet the trial could not prove non-inferiority. And that with even greater numbers than Z0011!
Z0011 was designed with the following hypothesis: 80% OS for ALND vs. 75% OS for SLNB at 5 years. And that was considered non-inferior! Does that sound logical to you? Would ANY of us say to a patient after SLNB: "Oh, your risk of dying at 5 years would be 5% lower if we treated your axilla (I am suggesting that ALND and RT have the same effect here), but hey, we don't think it's worth to treat your axilla just for a 5% benefit. Tomorrow you are going to the medical oncologist, he will discuss with your 6 months of chemotherapy for a 3% survival benefit!".

To my view, they chose an "easy" endpoint in order to limit the number of patients needed in the trial! Surgeons, what to expect...

Still 8/450 means roughly 2% excessive recurrence rate. Med. oncs. used to give 4 rounds of chemo for such a PFS-difference 20 years ago...
And we are giving SCV- and IM-irradiation for not "alot more" benefit based on theMA20- & EORTC-trials, which is potentially more toxic than just a high tangent.

So, show me anywhere, any data, that axillary recurrence rates are greater than 1-2% for T1/2 N1 patients in the modern era, because all the data I have seen says the rates are so low that no matter what approach any individual practitioner decides to take re: the axilla in this setting it will not appreciably affect therapeutic outcomes. Then the question becomes: what toxicity, and amount of toxicity, do you want the patient's axilla to countenance.

AMAROS had 1.19% with axillary RT. What AMAROS was missing was a third arm. One, where NOTHING would be done following BCS+SLNB, nothing other tangential irradiation of the breast (and that would be truly a tangential irradiation of the breast and not "do what you want" as was carried out in Z0011). Of course that would have been unethical when AMAROS was designed, that's why it never happened.
I do believe however, that the axillary recurrence rate in AMAROS would have been greater than 2% without any further intervention. We know that RT generally halves the risk of recurrence in breast cancer. Do I have data to back it up? Nope.

 

[thecarbonionangle]

Because the rad oncs got cold feet. We aren't exactly known for having much "balls" as rad oncs. We cower. Most people in trial got at least high tangents and maybe even more. Breast data is all a crapshoot, so much data, trials and no answer to questions. Institution dependent, but I have seen very few surgeons be enthusiastic about ALND these days, and they want to apply it to even patients that don't even fit any criteria in these trials (positive nodes which are negative now after neoadjuvant chemo, etc)...

 

[Pointless]

Because the rad oncs got cold feet. We aren't exactly known for having much "balls" as rad oncs. We cower. Most people in trial got at least high tangents and maybe even more. Breast data is all a crapshoot, so much data, trials and no answer to questions. Institution dependent, but I have seen very few surgeons be enthusiastic about ALND these days, and they want to apply it to even patients that don't even fit any criteria in these trials (positive nodes which are negative now after neoadjuvant chemo, etc)...

Exactly! Where is the guidance on how to manage these ypN0 patients that were initially node positive? We don't even know if radiation benefits them at all in the post-mastectomy setting (see NSABP B51). We have moved away from ALND, which is most likely for the best, but its left us as rad oncs with a lot of pieces to pick up and nobody designing trials to help us. I understand that the other disciplines don't give a **** about our little technical quandaries, but we should!

 

[thecarbonionangle]

Exactly! Where is the guidance on how to manage these ypN0 patients that were initially node positive? We don't even know if radiation benefits them at all in the post-mastectomy setting (see NSABP B51). We have moved away from ALND, which is most likely for the best, but its left us as rad oncs with a lot of pieces to pick up and nobody designing trials to help us. I understand that the other disciplines don't give a **** about our little technical quandaries, but we should!

Often quoted Mamounas et al data analysis from B18/B27 shows some patients who perhaps may benefit the least from adjuvant XRT in this setting but if you look at the N numbers, quite small. There's really no high level data.... We mostly treat all these patients with adjuvant XRT at my institution.

 

[scarbrtj]

BTW, not to cloud the issue (#sarcasm), anywhere from 6-12% of women with DCIS have a positive sentinel node. As we all know, we don't really debate these issues in DCIS (mercifully).

About 50,000 women per year in the US are diagnosed with DCIS, which mean we are leaving potentially thousands of women per year with DCIS with totally unaddressed/unthought-about axillas.

 

[scarbrtj]

I see what you're saying but I am not as nihilistic about management of the axilla as you are. As ALND has become increasingly unused, it falls to use to manage the lymph nodes. I think we should take that seriously. Nodal failures are a thing, crappy Z0011 data notwithstanding.

Eh, I am not a nihilist. But I'm also not an egoist (think radiation does anything here), nor a masochist (ask for more surgerization). I'm a cancer biologist heh.

 

[Pointless]

BTW, not to cloud the issue (#sarcasm), anywhere from 6-12% of women with DCIS have a positive sentinel node. As we all know, we don't really debate these issues in DCIS (mercifully).

About 50,000 women per year in the US are diagnosed with DCIS, which mean we are leaving potentially thousands of women per year with DCIS with totally unaddressed/unthought-about axillas.

Dude.... come on. You are quoting 20 year old retrospective series with <100 patients, seemingly to make the point that those of us who think about how to treated surgically unaddressed axillas are being hysterical, Tell me that you actually believe that 6-12% of DCIS patients have positive nodes. I dare you.

EDIT: I'm going to add a smiley emoji to soften what may have sounded overly aggressive. I'm enjoying the debate and always like to read your thoughts on things as you seem like a bright young chap, but I just disagree with you here and not even by a whole bunch.

 

[scarbrtj]

Dude.... come on. You are quoting 20 year old retrospective series with <100 patients, seemingly to make the point that those of us who think about how to treated surgically unaddressed axillas are being hysterical, Tell me that you actually believe that 6-12% of DCIS patients have positive nodes. I dare you.

Dang, it's like Christmas Story in here. I would quote more modern data, but you can't find any modern data because mysteriously (#sarcasm again), I guess, people quit looking at the axilla in DCIS even though they found non-zero rates of involvment. One thing is for sure: the rate of sentinel node involvement in DCIS is a helluva of a lot bigger than the rate of isolated axillary recurrence in early stage invasive patients (sans ALND, sans axillary RT) with SN involvement. (oh, and I never alleged hystericality; I take a more Socratic approach)

 

[Pointless]

Dang, it's like Christmas Story in here. I would quote more modern data, but you can't find any modern data because mysteriously (#sarcasm again), I guess, people quit looking at the axilla in DCIS even though they found non-zero rates of involvment. One thing is for sure: the rate of sentinel node involvement in DCIS is a helluva of a lot bigger than the rate of isolated axillary recurrence in early stage invasive patients (sans ALND, sans axillary RT) with SN involvement. (oh, and I never alleged hystericality; I take a more Socratic approach)

Please see my emoji-based edit from my previous response

 

[Mandelin Rain]

I agree with the one point made that the role of nodal management in breast cancer (a not infrequent diagnosis) has shifted to us. Yet, we can't say when, how, or where to treat nodes in breast cancer patients. This seems less than ideal, regardless of the perceived small benefits of doing things well for patients.

 

[Mandelin Rain]

Perhaps what this quandary needs is a well meaning resident to email blast a survey with the lure of a chance to win a $20 Starbucks gift card. Sure, we won't know what the "best" way to manage nodes in breast cancer, but we will know the most popular way (among the 17 rad oncs who respond).

 

[scarbrtj]

I agree with the one point made that the role of nodal management in breast cancer (a not infrequent diagnosis) has shifted to us. Yet, we can't say when, how, or where to treat nodes in breast cancer patients. This seems less than ideal, regardless of the perceived small benefits of doing things well for patients.

1) role of nodal management in breast cancer (a not infrequent diagnosis) has shifted to us
2) We can't say when, how, or where to treat nodes in breast cancer patients
3) This seems less than ideal

God grant me the serenity to accept the things I cannot change; courage to change the things I can; and wisdom to know the difference.

 

[thecarbonionangle]

Perhaps what this quandary needs is a well meaning resident to email blast a survey with the lure of a chance to win a $20 Starbucks gift card. Sure, we won't know what the "best" way to manage nodes in breast cancer, but we will know the most popular way (among the 17 rad oncs who respond).

Publish in Red Journal with 17 co-authors and Zeitman’s blessing.

 

[Mandelin Rain]

1) role of nodal management in breast cancer (a not infrequent diagnosis) has shifted to us
2) We can't say when, how, or where to treat nodes in breast cancer patients
3) This seems less than ideal

God grant me the serenity to accept the things I cannot change; courage to change the things I can; and wisdom to know the difference.

Our meathead colleagues in the OR have moved from Halsted, to MRM, to Lump + ALND, to Lump + SLNBx +/- ALND in the same time we've gone from 2D tangents +/- dealer's choice nodes to 3D tangents +/- dealer's choice nodes. All incremental progresses that I doubt anyone regrets working toward. As a field, we seem to be liberally conservative when it comes to breast cancer progress.

 

[scarbrtj]

Halstead was to breast what Ptolemy was to astronomy. Everything was empiric, and then deeper observations were made and hypotheses tested. Next thing you know Earth is not the center of the universe and breasts don’t have to be cut off down to the pectoral fascia. Almost everything that was known about “nodal management in breast cancer” was empiric even twenty years ago. The meathead colleagues in the OR implanted all these empiric concepts inside the eggheads’ minds down in the radiation basement. Radiation therapy started as a mop-up-the-slop branch of surgery. Like who says you have to cover the whole breast after a lumpectomy? That’s surgical empiricism. A radiation oncologist would’ve never thought that de novo. We did whole breast RT after lumpectomy so as to assuage surgeons’ fears; that one was recently dealt a blow. When data is pretty clear that axillary recurrences are <<1% in undisturbed (SN bx only, no surgery, no RT for T1/2pN1) axillae, wanting to "do something" to it is... meathead surgical empiricism. To paraphrase Nietzsche, he who fights surgeons should take care not to become a surgeon.

 

[KHE88]

I am going to rant a little bit. I really, really hate the breast literature. I despise it. And those who follow a dogma when no clear conclusions exist.

The question is what has changed with z11 and ECE.

The answer is nothing.
It was a garbage study then.
It remains a garbage study now.

The concept that anyone makes any treatment decisions off of this "study" blows my mind.
You have known disease in an axillary lymph node. Yet, you are just supposed to ignore that and do nothing?
How are you going to explain this to the patient?
Oh by the way, 20% of patients actually got a third field to the supraclav fossa and nearly half had the low axilla treated. And nearly a third had more axillary lymph nodes beyond the sentinel when the surgeons went hunting for them.
Are you kidding me?

I just want to LOL anytime someone mentions z11 in conference. Or ever, really.

The lack of good trials and clear treatment paradigms in breast oncology is one of the true travesties in modern cancer treatment IMO. We are innudated with myriad silly retrospective studies, meta-analyses of ancient trials, and re-analysis data hunting studies to memorize as residents.

There is no reason breast radiation should be this unclear especially given how stupidly common it is. There are so many basic outstanding questions that could easily be addressed with well-designed trials. But nope. We're hellbent on letting the surgeons drive everything with garbage trials and rad oncs focused on gimmicks and single institution cowboy-like paradigms to generate publications with highly dubious reproducibility.

 

[scarbrtj]

You have known disease in an axillary lymph node. Yet, you are just supposed to ignore that and do nothing?
How are you going to explain this to the patient?

It is tough to ignore obvious things, granted. "I just want to LOL anytime someone mentions z11 in conference." We'll be LOLing forever because a trial like Z0011 is not going to be repeated. It was controversial in its conception, controversial in its results, but it changed the standard of care, no denying that. So not mentioning it would be like denying history. The world, for better or worse no matter your predilections, has moved on. There is an orgy of data that axillary recurrence rates are super-low if doctors "ignore [the axilla] and do nothing." The odds are SO in disfavor of there being a non-confirmatory/contradictory Z0011-like trial that even if a contradictory trial appeared, it wouldn't change the current approach and trend.

In these early stage patients with positive SNs where we are "ignoring" the axilla otherwise...
Perhaps~20-30% have known disease in their IMNs.
Perhaps ~20% have known disease in their bloodstream.
Rad oncs: please ignore these two things. (EDIT: And even when you think the axilla is negative, there's a ~10% chance it's positive; please ignore that too.)

 

[gmsquid]

A couple of points.

Regarding the neoadjuvant setting the question of nodal therapy (not dissection) in ypN0 patients is going to be indirectly answered by B51. A large portion of those patients are going to be ypN0 by SLN and we will see if nodal RT has a benefit over none (a subgroup analysis in the SLN patients will be interesting and if axRT doesn’t have a benefit then that likely means the benefit of additional local therapy with a dissection is minimal). What’s going to be interesting is the subgroup analyses based on biology. How many Her2 patients on the trial versus TN versus HR positive. That will be the next series of arguments in breast cancer. Which biology does local therapy matter

Second point: I think people always talk low axillary recurrence risks as no benefit as if RTs only benefit is local. There was a DMFS and DFS benefit to comprehensive nodal RT on MA20 and the EORTC trial. Those patients on those trials actually had pretty low volume axillary disease (most were single node small volume I believe) Could it all be from the IMN treatment? Possibly. But could it also be that axillary recurrences are underreported in patients who have subsequent/concurrent distant metastasis? Don’t sell the local therapy short as therapy without systemic benefit

 

[scarbrtj]

A couple of points.

Regarding the neoadjuvant setting the question of nodal therapy (not dissection) in ypN0 patients is going to be indirectly answered by B51. A large portion of those patients are going to be ypN0 by SLN and we will see if nodal RT has a benefit over none (a subgroup analysis in the SLN patients will be interesting and if axRT doesn’t have a benefit then that likely means the benefit of additional local therapy with a dissection is minimal). What’s going to be interesting is the subgroup analyses based on biology. How many Her2 patients on the trial versus TN versus HR positive. That will be the next series of arguments in breast cancer. Which biology does local therapy matter

Second point: I think people always talk low axillary recurrence risks as no benefit as if RTs only benefit is local. There was a DMFS and DFS benefit to comprehensive nodal RT on MA20 and the EORTC trial. Those patients on those trials actually had pretty low volume axillary disease (most were single node small volume I believe) Could it all be from the IMN treatment? Possibly. But could it also be that axillary recurrences are underreported in patients who have subsequent/concurrent distant metastasis? Don’t sell the local therapy short as therapy without systemic benefit

I'll see your points and raise you a few more.

1) One's view of a particular study's findings is directly proportional to one's pre-study views/opinions on the questions the study is attempting to answer; B51 will be no different. (For example, EORTC 22845 showed adjuvant RT significantly improves local control and prevents seizures in low grade glioma... ordinarily this would "solidify" its status as a standard of care but one can use lack of OS benefit, availability of salvage, and worry of side effects, to "poo-poo" the results. No matter one's own view of EORTC 22845, you will not hear a discussion of EORTC 22845 without some personal rationalization of what the study "actually means"—you should give adjuvant RT! you shouldn't give adjuvant RT!— versus the findings of EORTC 22845 in and of itself.)
2) "Radiation oncologists are obsessed with locoregional recurrence of breast cancer." It's thinking like this which would have made SLN biopsy a standard of care in DCIS, an approach which from the armchair of 2019 we'd all call "wrong." Somewhere between an idée fixe and "who cares" is probably the right way to think about the nodes in breast cancer. I sure don't have the answer(s).
3) Radiation therapy is not a purely locoregional therapy in breast cancer. Or, probably, many other cancers. An "unknown unknown" as Donald Rumsfeld would've said.

 

[evilbooyaa]

I think we should all be sure we know the data, realize when patient meets inclusion criteria, realize when it doesn't give us the most clearcut answer, and then practice to our preference.

For example (getting back to OP) one should not say 'z11 did not include microscopic ENE, this patient should get ALND' because that is not correct. If you're going to treat the patient with SLNB + RT in setting of 1/1 LN w/ ECE, then I think doing standard tangents, high tangents, and full 3-field are all reasonable and 'defensible'. Saying that one should be 'dogma' is IMO, a misinterpretation of the data. That being said, if you 'under' treat the patient (let's say w/ standard tangents) and they recur in the axilla, there's going to be a lot of monday morning quarterbacking on your decision (either from yourself or others)

For me, if I don't have good data to safely de-escalate treatment (given that 20% of people got 3rd field and 1/3rd got high tangents) I'm going to focus on treating the cancer.

 

[Palex80]

Second point: I think people always talk low axillary recurrence risks as no benefit as if RTs only benefit is local. There was a DMFS and DFS benefit to comprehensive nodal RT on MA20 and the EORTC trial. Those patients on those trials actually had pretty low volume axillary disease (most were single node small volume I believe) Could it all be from the IMN treatment? Possibly. But could it also be that axillary recurrences are underreported in patients who have subsequent/concurrent distant metastasis? Don’t sell the local therapy short as therapy without systemic benefit

MA20 did report on the location of recurrences.
"The majority of regional treatment failures included the axillary nodes (in 63% of patients) or the supraclavicular nodes (in 27%)." There were 23 regional recurrences without RNI and 5 with RNI out 916 patients each. That means that there is a gain of 2% less regional recurrences based on MA20.
Please bear in mind that patients in the MA20 RNI-arm did not receive a "full" axillary irradiation. They merely had the internal mammary, the paraclavicular nodes and the level III axilla nodes irradiated. Now, in the old days prior to Z0011, surgeons seldom dissected level III axillary nodes in cN0 patients that came up as pN1 during surgery in frozen section of the SLN. Common practice, in my institution at least, was for them to dissect level I & II. I can only recall extensive level III dissections in patients with bulky cN2 disease and bad biology where the surgeons were and still are dissecting as many nodes as possible.

So let's see...
Based on MA20 lots of us have argued and continue to argue to conduct RNI in patients with positive nodes after ALND. Our motivation are the 2% gain in less regional recurrences through RNI (which according to our observations seem to have altered the course of the disease as well with less metastasis, although the mechanism is not completely clear).

At the same time based on Z0011 people say that cN0/pN1-patients do not need ALND or RNI.

Do you see the irony?

Some of us are calling for RNI based on MA20, treating SCV and MI on pN1 patients who have had ALND. Yet at the very same time some of us seem to be perfectly comfortable in not treating the axilla in the very same patients with pN1 and just a sentinel node dissection without ALND.

Now, who of these patients do you think has a higher risk for harboring microscopic disease in any non-dissected nodes?
The 65 year old patient with a pT2 pN1 (1/1) cM0 G2 IDC post BCS with SLN or the 65 year old patient with a pT2 pN1 (1/12) cM0 G2 IDC post BCS with ALND?
Who would you rather treat with any form of RNI and what would your targets be?
The first patient is a typical patient post surgery applying Z0011-practice. The second patient is a typical patient that would have been enrolled in MA20.

 

[Mandelin Rain]

MA20 did report on the location of recurrences.
"The majority of regional treatment failures included the axillary nodes (in 63% of patients) or the supraclavicular nodes (in 27%)." There were 23 regional recurrences without RNI and 5 with RNI out 916 patients each. That means that there is a gain of 2% less regional recurrences based on MA20.
Please bear in mind that patients in the MA20 RNI-arm did not receive a "full" axillary irradiation. They merely had the internal mammary, the paraclavicular nodes and the level III axilla nodes irradiated. Now, in the old days prior to Z0011, surgeons seldom dissected level III axillary nodes in cN0 patients that came up as pN1 during surgery in frozen section of the SLN. Common practice, in my institution at least, was for them to dissect level I & II. I can only recall extensive level III dissections in patients with bulky cN2 disease and bad biology where the surgeons were and still are dissecting as many nodes as possible.

So let's see...
Based on MA20 lots of us have argued and continue to argue to conduct RNI in patients with positive nodes after ALND. Our motivation are the 2% gain in less regional recurrences through RNI (which according to our observations seem to have altered the course of the disease as well with less metastasis, although the mechanism is not completely clear).

At the same time based on Z0011 people say that cN0/pN1-patients do not need ALND or RNI.

Do you see the irony?

Some of us are calling for RNI based on MA20, treating SCV and MI on pN1 patients who have had ALND. Yet at the very same time some of us seem to be perfectly comfortable in not treating the axilla in the very same patients with pN1 and just a sentinel node dissection without ALND.

Now, who of these patients do you think has a higher risk for harboring microscopic disease in any non-dissected nodes?
The 65 year old patient with a pT2 pN1 (1/1) cM0 G2 IDC post BCS with SLN or the 65 year old patient with a pT2 pN1 (1/12) cM0 G2 IDC post BCS with ALND?
Who would you rather treat with any form of RNI and what would your targets be?
The first patient is a typical patient post surgery applying Z0011-practice. The second patient is a typical patient that would have been enrolled in MA20.

This has always bothered me about those studies. One implies, do everything about the nodes. One implies do nothing about the nodes. Regardless... while I haven't reviewed Table 1 on these trials in some time, I assume (maybe a bad move) most of the perimenopausal women received chemotherapy in both. While not given for it's LRR effect, chemo must have SOME activity on microscopic disease in undissected nodes. For my patient, that's now off the table. So it's pretty much radiation or bust for this lady.

 

[evilbooyaa]

MA20 did report on the location of recurrences.
"The majority of regional treatment failures included the axillary nodes (in 63% of patients) or the supraclavicular nodes (in 27%)." There were 23 regional recurrences without RNI and 5 with RNI out 916 patients each. That means that there is a gain of 2% less regional recurrences based on MA20.
Please bear in mind that patients in the MA20 RNI-arm did not receive a "full" axillary irradiation. They merely had the internal mammary, the paraclavicular nodes and the level III axilla nodes irradiated. Now, in the old days prior to Z0011, surgeons seldom dissected level III axillary nodes in cN0 patients that came up as pN1 during surgery in frozen section of the SLN. Common practice, in my institution at least, was for them to dissect level I & II. I can only recall extensive level III dissections in patients with bulky cN2 disease and bad biology where the surgeons were and still are dissecting as many nodes as possible.

So let's see...
Based on MA20 lots of us have argued and continue to argue to conduct RNI in patients with positive nodes after ALND. Our motivation are the 2% gain in less regional recurrences through RNI (which according to our observations seem to have altered the course of the disease as well with less metastasis, although the mechanism is not completely clear).

At the same time based on Z0011 people say that cN0/pN1-patients do not need ALND or RNI.

Do you see the irony?

Some of us are calling for RNI based on MA20, treating SCV and MI on pN1 patients who have had ALND. Yet at the very same time some of us seem to be perfectly comfortable in not treating the axilla in the very same patients with pN1 and just a sentinel node dissection without ALND.

Now, who of these patients do you think has a higher risk for harboring microscopic disease in any non-dissected nodes?
The 65 year old patient with a pT2 pN1 (1/1) cM0 G2 IDC post BCS with SLN or the 65 year old patient with a pT2 pN1 (1/12) cM0 G2 IDC post BCS with ALND?
Who would you rather treat with any form of RNI and what would your targets be?
The first patient is a typical patient post surgery applying Z0011-practice. The second patient is a typical patient that would have been enrolled in MA20.

Excellent post. I think people have a tendency to focus on Z11 on its own and ignore the value of RNI. I know I did in my previous reply, and you've re-confirmed me why my practice in this scenario would be full 3-field treatment.

 

[scarbrtj]

MA20 did report on the location of recurrences.
"The majority of regional treatment failures included the axillary nodes (in 63% of patients) or the supraclavicular nodes (in 27%)." There were 23 regional recurrences without RNI and 5 with RNI out 916 patients each. That means that there is a gain of 2% less regional recurrences based on MA20.
Please bear in mind that patients in the MA20 RNI-arm did not receive a "full" axillary irradiation. They merely had the internal mammary, the paraclavicular nodes and the level III axilla nodes irradiated. Now, in the old days prior to Z0011, surgeons seldom dissected level III axillary nodes in cN0 patients that came up as pN1 during surgery in frozen section of the SLN. Common practice, in my institution at least, was for them to dissect level I & II. I can only recall extensive level III dissections in patients with bulky cN2 disease and bad biology where the surgeons were and still are dissecting as many nodes as possible.

So let's see...
Based on MA20 lots of us have argued and continue to argue to conduct RNI in patients with positive nodes after ALND. Our motivation are the 2% gain in less regional recurrences through RNI (which according to our observations seem to have altered the course of the disease as well with less metastasis, although the mechanism is not completely clear).

At the same time based on Z0011 people say that cN0/pN1-patients do not need ALND or RNI.

Do you see the irony?

Some of us are calling for RNI based on MA20, treating SCV and MI on pN1 patients who have had ALND. Yet at the very same time some of us seem to be perfectly comfortable in not treating the axilla in the very same patients with pN1 and just a sentinel node dissection without ALND.

Now, who of these patients do you think has a higher risk for harboring microscopic disease in any non-dissected nodes?
The 65 year old patient with a pT2 pN1 (1/1) cM0 G2 IDC post BCS with SLN or the 65 year old patient with a pT2 pN1 (1/12) cM0 G2 IDC post BCS with ALND?
Who would you rather treat with any form of RNI and what would your targets be?
The first patient is a typical patient post surgery applying Z0011-practice. The second patient is a typical patient that would have been enrolled in MA20.

Not to muddy the waters, but MA20 and EORTC 22922/10925 were co-reported in 2015; one on p.307 of the NEJM and the other p.317. (The reason I don't mention MA20 is because yes it confuses me, esp in light of its sister EORTC study. I tend to mangle what the study "means.") First, the elephant in the room: no survival advantages have ever been phIII reported specifically for RNI in breast cancer. (Whereas its much easier to "prove" some local therapy XRT of some sort improves survival in breast cancer.) Second, the European study was ostensibly a non-axillary-RT study, just sclav and IMN RNI only. And what did not irradiating the axilla do in the EORTC study? It decreased axillary recurrence rates by ~33%. Explain that without playing verbal/logical Twister; I can't. Fourth, re: "Please bear in mind that patients in the MA20 RNI-arm did not receive a full axillary irradiation... [t]hey merely had the internal mammary, the paraclavicular nodes and the level III axilla nodes irradiated," here's what MA20 said:

For patients who had fewer than 10 axillary nodes removed or more than 3 positive axillary nodes, the lateral aspect of the field was extended laterally to include the level I and II axillary lymph nodes.​

How many got level 3-only axillary RT in MA20? It seems 2/3 did, but 1/3 got the full axillary Monty I believe. Fifth, re: "There were 23 regional recurrences without RNI and 5 with RNI out 916 patients each. That means that there is a gain of 2% less regional recurrences based on MA20" and "The majority of regional treatment failures included the axillary nodes (in 63% of patients)"... at 10 years there was a 7.2% vs 4.5% of isolated locoregional recurrence in favor of RNI in MA20 (p=0.018), so in fact that 2% less may be ~2-3% less. But would this still be p<0.05 significant if we looked at isolated axillary recurrence only? I don't know; I have doubts.

Finally, no, I see no irony in Z0011 vs MA20. But like Forrest Gump said, I am not a smart man! Z0011 is a purely axillary concern, the other a much more hazy and nebulous axillary concern. And, as mentioned, toss in EORTC 22922 and that almost tosses any axillary insights out the window. Both MA20 and the EORTC show RNI lowers regional recurrences (again, the axillary issue by itself is hazy) in the neighborhood of needing to provide RNI to around 50 women to prevent one regional recurrence; and all that sans survival benefits. If you want to irradiate axillae, you better irradiate sclavs and IMNs at the same time or you'll lose therapeutic firepower (if you truly believe it exists). I still say, you can concentrate on the axilla if you want. I saw a woman in clinic yesterday who had ALND; her unsolicited complaint: "They ruined my arm." We can also concentrate on how much oil the terrorists use in their hummus.

The 65 year old patient with a pT2 pN1 (1/1) cM0 G2 IDC post BCS with SLN or the 65 year old patient with a pT2 pN1 (1/12) cM0 G2 IDC post BCS with ALND?
Who would you rather treat with any form of RNI and what would your targets be?

In my world, the second patient wouldn't exist in a malpractice-ish ("Your honor, this woman can no longer drive the bus OR iron clothes!" wink wink) sort of way, and definitely no RNI for the first patient. Aaaand... these are the same patients, it's just that second one went into a deep gravity well, recently emerging, in 1999 where she aged almost imperceptibly while time in our inertial frame marched onward.

 

[Palex80]

I urge you to have a look at the study protocols of the EORTC trial and the MA20 trial. You will then see HOW irradiation was performed in the RNI groups. You can find the study protocols in the supplementary material at NEJM...

Here's a hint from MA20

That's a lot of axillary nodes treated...

 

[scarbrtj]

I urge you to have a look at the study protocols of the EORTC trial and the MA20 trial. You will then see HOW irradiation was performed in the RNI groups. You can find the study protocols in the supplementary material at NEJM...

Here's a hint from MA20
View attachment 269804

That's a lot of axillary nodes treated...

If I were intimately aware of the field design from each and every patient from MA20, what difference would it make to me as an individual practitioner for the OP's case about which we're talking.

...........................Axillary recurrence rates
...........................axillary/nodal tx.......no axillary tx
EORTC(1).........1.3%..............................1.9%
MA20(2).........~0.3%.........................~1.6%
Z0011(3).........~0.5%.........................~0.9%
Z0011f/u(4)....0%...............................~1%

1: An IMN/sclav study, but some(?) patients got axillary RT
2: "a lot of axillary nodes treated" (see above); ~67% of regional recurrences were axillary
3: Difficult to be sure about extent of axillary RT, "no axillary-specific irradiation"; 4/446 recurred in axilla with no axillary tx, 2/445 in the other arm
4: ~793 pts, "no isolated axillary recurrences" @2.5y med f/u

5-∞: no change in OS or natural history by treating/not treating axilla in these studies; NNT for axillary therapy ~100-200 patients

 

[Palex80]

It will tell you that the 2% difference in axillary recurrence rate is probably "driving" the results of the trial.
And since it was a trial with ALND in most patients and YET STILL managed to produce such a benefit that led to NCCN changing its guidelines and many people doing RNI for pN1, you may consider doing axillary RT in patients with only SLN and witout ALND...
That's the point.

 

[scarbrtj]

It will tell you that the 2% difference in axillary recurrence rate is probably "driving" the results of the trial.
And since it was a trial with ALND in most patients and YET STILL managed to produce such a benefit that led to NCCN changing its guidelines and many people doing RNI for pN1, you may consider doing axillary RT in patients with only SLN and witout ALND...
That's the point.

Well first off I don't see, and haven't seen, 2% improvements cited anywhere ever. This may be quibbling over decimal points.
Second, re: "you may consider doing axillary RT in patients with only SLN and without ALND," what if we suggested ALND instead as the OP @Mandelin Rain opined: "Is this someone you just treat without discussing dissection?" If his patient had an ALND, you think there'd be a ~2% improvement in axillary recurrence? Seriously? The data strongly says otherwise, yes?
Third, if you accept the premise that axillary RT and ALND are roughly oncologically similarly therapeutically, and you accept the premise that ALND is not needed for most SLN+ T1/2 patients nowadays--which obviously has been a tough sell, mostly for doesn't-treat-much-breast-CA docs--thinking about adding RNI for SLN+/no ALND is illogical, Mr. Spock.
Fourth, the recurs-only-in -the-axilla patient is a rare, rare, about 1-in-1000 patient.
Fifth, I was taught (and I don't have a great citation for this) that for a radiation oncologist to have an good, measurable impact on something, he/she needs to be treating in an area with at least a 10-15% local recurrence risk.
Sixth, with #5 in mind, can we agree the NNT for axillary RT is about 100? At least? And if so, do you think the number-needed-to-harm with axillary RT is as high as 100? I do not. I think it's in the 25-50 range, maybe less. Ergo, risk/benefit considerations, every action has an equal/opposite reaction, and all that.
(Can you see my mind is unchangeable )

 

[Palex80]

Well first off I don't see, and haven't seen, 2% improvements cited anywhere ever. This may be quibbling over decimal points.
Second, re: "you may consider doing axillary RT in patients with only SLN and without ALND," what if we suggested ALND instead as the OP @Mandelin Rain opined: "Is this someone you just treat without discussing dissection?" If his patient had an ALND, you think there'd be a ~2% improvement in axillary recurrence? Seriously? The data strongly says otherwise, yes?

I wouldn't say they do. It depends on THE PATIENT. A patient with a macrometastasis and possible some other adverse features, like 1/1 and not 1/3 SLN and maybe pre-menopausal may very well have a risk of >2% for an axillary recurrence without further treatment to the axilla. Z0011-rates are not valid, at least not for me for several reasons: a) lots and lots of low-risk patients in the trial, b) axillary RT to many patients, c) serious doubts on completeness of data, d) study not designed to answer question of recurrence rates.

Third, if you accept the premise that axillary RT and ALND are roughly oncologically similarly therapeutically, and you accept the premise that ALND is not needed for most SLN+ T1/2 patients nowadays--which obviously has been a tough sell, mostly for doesn't-treat-much-breast-CA docs--thinking about adding RNI for SLN+/no ALND is illogical, Mr. Spock.

I never said that T1/T2 SLN+ dont benefit from axillary treatment. That's my point the whole time! We need to consider treating more lymphatics if the surgeons operate less!

Fourth, the recurs-only-in -the-axilla patient is a rare, rare, about 1-in-1000 patient.

Says who?
MA20 certainly doesn't.

Fifth, I was taught (and I don't have a great citation for this) that for a radiation oncologist to have an good, measurable impact on something, he/she needs to be treating in an area with at least a 10-15% local recurrence risk.

I wasn't taught the same.

Sixth, with #5 in mind, can we agree the NNT for axillary RT is about 100? At least? And if so, do you think the number-needed-to-harm with axillary RT is as high as 100? I do not. I think it's in the 25-50 range, maybe less. Ergo, risk/benefit considerations, every action has an equal/opposite reaction, and all that.

You probably need to treat 20-30 women to spare one from an axillary recurrence in a collective of patients who are not super-low-risk. That means ruling out the ones with micrometastasis or the old, G1 tiny T1s.
That's enough for me.
I deliver post-BCS RT every day to the breast for around 3-4% lower recurrence rates to the breast.
Why should I change my practice when it comes to the axilla?
In fact, a recurrence in the axilla can be actually more difficult to manage than a recurrence to the breast.