No more PMRT for cN1 ypN0 breast cancer?

[TheWallnerus]

Of course we are. These are two, totally distinct settings.
It's like saying you would be treating a cN0 prostate cancer with a moderare/high risk for nodal involvement the same way as you would be treating a pN1 prostate cancer.

We actually know very little about how the individual patients did, since the protocol was written so poorly that we could only guess what kind of doses were delivered to the axilla. Furthermore, the vast majority of Z0011 patients had luminal A tumors.

ALND is both.

I dont "want" something.
Standard of care for cN+ after neoadjuvant chemo is to perform ALND or TAS. Most probably do TAS. SLNB is certainly not supported by evidence.

The Dutch study results cannot be extrapolated to modern practice if you are going to do SLNB only. Not, when 80% of the patients had ALND on the trial. The trial basically delivered ALND to cN1/ypN0 patients and deferred RT. I would be fine omitting RNI in this situation too and would only treat chestwall if there were risk factors for local recurrence (pure response of the primary, L1, etc...)
Extrapolating this to SLNB is not backed up by these data.

There is quite good data that performing SLNB post-chemotherapy is not as reliable as performing it pre-chemotherapy.

Sentinel-lymph-node biopsy in patients with breast cancer before and after neoadjuvant chemotherapy (SENTINA): a prospective, multicentre cohort study - PubMed

Brustkrebs Deutschland, German Society for Senology, German Breast Group.

pubmed.ncbi.nlm.nih.gov

Of 1737 patients who received treatment, 1022 women underwent sentinel-lymph-node biopsy before neoadjuvant chemotherapy (arms A and B), with a detection rate of 99.1% (95% CI 98.3-99.6; 1013 of 1022). In patients who converted after neoadjuvant chemotherapy from cN+ to ycN0 (arm C), the detection rate was 80.1% (95% CI 76.6-83.2; 474 of 592) and false-negative rate was 14.2% (95% CI 9.9-19.4; 32 of 226). The false-negative rate was 24.3% (17 of 70) for women who had one node removed and 18.5% (10 of 54) for those who had two sentinel nodes removed (arm C). In patients who had a second sentinel-lymph-node biopsy procedure after neoadjuvant chemotherapy (arm B), the detection rate was 60.8% (95% CI 55.6-65.9; 219 of 360) and the false-negative rate was 51.6% (95% CI 38.7-64.2; 33 of 64).

So if you ommit PMRT in a patient with ycN0 and ypN0 on SLNB means you are leaving cancer in the axilla in 14% of all patients.
This is bad.

Leaving cancer in the axilla is not bad ("New 10-year results from a major clinical trial in breast cancer confirm that it does not compromise overall survival to leave behind minimal amounts of cancer that have spread to the underarm lymph nodes in certain patients"). This is the fundamental issue. I again refer to z0011. Is it counterintuitive. Of course. Does it strike fear in our hearts. Of course.

America is sometimes accused of being the world’s policeman. Rad oncs are not the breast cancer policemen. You do “good enough” with your locoregional treatment, and that as they say is that. One out of three women with cN1 have disease in their bone marrow which is “left behind” after surgery and adjuvant RT.

---

Members don't see this ad.

 

[Palex80]

That’s not what Bernie found.

He never carried out that trial. cN1/ypN0

Leaving cancer in the axilla is not bad. This is the fundamental issue. I again refer to z0011. Is it counterintuitive. Of course. Does it strike fear in our hearts. Of course.

America is sometimes accused of being the world’s policeman. Rad oncs are not the breast cancer policemen. You do “good enough” with your locoregional treatment, and that as they say is that. One out of three women with cN1 have disease in their bone marrow which is “left behind” after surgery and adjuvant RT.

This is not the same patient population. Z0011 was mostly small luminal A. In the Dutch trial you have women with a chemotherapy indication, solely.

 

[TheWallnerus]

He never carried out that trial. cN1/ypN0

Yes of course he never carried out THIS trial. Go and look at what Bernie said about local control and its interrelationship with survival and his comments on EBCTCG 2000. He would never run this trial.

Two questions. What sort of trial would you run here and what would be its hypotheses, and would there be equipoise? In EBCTCG 2014, for pathologically node negative women (who got ALND… but allow me not to perseverate on that for a moment), radiation therapy helped no oncological measure. But the number needed to kill with radiotherapy was about 16: approximately 47% of women who got radiotherapy were dead at 20y vs 41% of women who didn’t get RT. (For pN0 negative women who got “sampling,” radiotherapy helped LC and didn’t hurt survival but the unirradiated women had higher 5 and 10y LRF than we see today.)

We know radiotherapy can confer a survival detriment to improperly selected breast cancer patients so we have to be so judicious about when to give it and how much to irradiate. I mention the node negative/ALND women from the EBCTCG 2014 because their 5y local failures rates closely mirror cN1/ypN0 women from RAPCHEM. If we take it for a given that unirradiated cN1/ypN0 women have 5y 2% local failure rates, what trial is conceivable and what do you hypothesize with RT. How many women would you need to prove it. And would an IRB ever approve it. Would it be honest or salacious to state there could be a survival disadvantage with PMRT here.

 

[bigman1]

Interesting dialectic here between Walnerrus and Palex.
Is there any evidence NAC approach holds any survival advantage over maximal upfront LRC including some sort of axillary treatment?
I hate the idea of a maximal systemic therapy approach to breast.

 

[Palex80]

Two questions. What sort of trial would you run here and what would be its hypotheses, and would there be equipoise?

I would run a trial where patients with initially cN1 that converted to cN0, got either targetted axillary dissection or axillary RT only.

In EBCTCG 2014, for pathologically node negative women (who got ALND… but allow me not to perseverate on that for a moment), radiation therapy helped no oncological measure. But the number needed to kill with radiotherapy was about 16: approximately 47% of women who got radiotherapy were dead at 20y vs 41% of women who didn’t get RT. (For pN0 negative women who got “sampling,” radiotherapy helped LC and didn’t hurt survival but the unirradiated women had higher 5 and 10y LRF than we see today.)

Wait, wait, wait... I only see a non-significant difference at 20 years of 3.8%.

Appendix Page 36 of article: Effect of radiotherapy after mastectomy and axillary surgery on 10-year recurrence and 20-year breast cancer mortality: meta-analysis of individual patient data for 8135 women in 22 randomised trials - PubMed

 

[TheWallnerus]

I would run a trial where patients with initially cN1 that converted to cN0, got either targetted axillary dissection or axillary RT only.

This trial will never be done. Reasons:

1) Everyone will demand 10 year data for it to be practice-changing. This means at least about 15-20y of running the trial and accruing patients to reach a ~10y median f/u, usually. No one will have enthusiasm to do this trial. Of course, you can report 10y data as long as you've followed any patient in your huge study for at least ten years...
2) So assume 5y 2% local recurrence sans RT becomes ~4% at ~10y and RT might cut that in half to ~2%. You'll need a N=~500 sample (14y accrual, 7y median f/u, 2-sided surv prob, H0=0.96, H1: 0.98, power=0.8 at t=10y)
3) But you will need ~7000 patients to detect a 1% survival increase at 10y (14y accrual, 7y median f/u, 2-sided surv prob, H0=0.90, H1: 0.91, power=0.8 at t=10y)

It would be a waste of a career to try and run this study.

At 5y in RAPCHEM, here were the OS's:

There was a trend toward better OS with less RT, worse OS with more RT. Were they significant? No. But again, on the basis of these, that's why I would hypothesize a ~90% 10y OS without RT and a ~91% 10y OS w/ RT. And make it a two-sided hypothesis. A non-inferiority study would entail Women's Health Initiative-size samples.

Wait, wait, wait... I only see a non-significant difference at 20 years of 3.8%.

NNK (no. needed to kill) w/ PMRT at 20y for pN0 by ALND: 1/0.06 = 16.7 patients

 

[TheWallnerus]

Is there any evidence NAC approach holds any survival advantage over maximal upfront LRC including some sort of axillary treatment?

I think the way you have phrased I would answer no/unsure/debatable. HOWEVER, for women who are N+ upfront and then become ypN0, NAC seems to confer a significant survival advantage. And for women who become ypT0N0, NAC confers a powerfully significant survival advantage.

 

[Palex80]

Thank you, Wallrenus, for pointing out at the correct graph!

 

[bigman1]

I think the way you have phrased I would answer no/unsure/debatable. HOWEVER, for women who are N+ upfront and then become ypN0, NAC seems to confer a significant survival advantage. And for women who become ypT0N0, NAC confers a powerfully significant survival advantage.

Data on long term cardiotoxicity from anthraycyclines/Herceptin in HER2+ disease, long term disabling neuropathies, second malignancies, complications from VTE, etc? This is the breast version of TNT. No women can tolerate these regimens if there's an equivalent local control that is better tolerated.
I literally had a med onc say that re-irradiation to chest wall after early recurrence will cause second malignancy, but has zero qualms with extended course anthracyclines.
BTW your analysis of relevant data has been incredible here.

 

[TheWallnerus]

Data on long term cardiotoxicity from anthraycyclines/Herceptin in HER2+ disease, long term disabling neuropathies, second malignancies, complications from VTE, etc? This is the breast version of TNT. No women can tolerate these regimens if there's an equivalent local control that is better tolerated.
I literally had a med onc say that re-irradiation to chest wall after early recurrence will cause second malignancy, but has zero qualms with extended course anthracyclines.
BTW your analysis of relevant data has been incredible here.

Local control is one thing and survival is another. What have modern trials (such as Z0011, MA20, EORTC 22922, even the Korean IMN ASTRO plenary from last year) taught us… it’s that variations in locoregional therapies do not significantly affect overall survival in this disease. RAPCHEM hits the point home yet again. What does significantly affect survival in this disease? The systemic treatments. Anti HER2 therapies have saved thousands upon thousands of lives and have literally changed the natural history of HER2 positive breast cancer. Not to beat a dead horse but again:

 

[Reaganite]

And to add icing on the cake....they used forward planned imrt!

 

[deleted1111261]

And to add icing on the cake....they used forward planned imrt!

Forward planned OR inverse planned !

 

[TheWallnerus]

And to add icing on the cake....they used forward planned imrt!

Forward planned OR inverse planned !

Isn't it wild how Evicore never wrote letters to the various editors all these times "forward planned" and "IMRT" appeared together trying to debate/support their point in the medical literature but instead would write edicts from undisclosed locations "radsplaining" that forward-planned IMRT is just 3DCRT?

Wait. I think ASTRO did that too.

 

[evilbooyaa]

My outlook on this trial (edit: prospective registry) is that it suggests that B51 is (more likely than not) going to show us that the experimental arm of omitting nodal RT in ypN0 patients is going to do fine. That being said, I'll wait until B51 is released to change my practice rather than relying on 291 patients in a single arm fashion.

Most of these patients are not clinically getting ALND anymore. TAS is a heterogeneous process even amongst breast surgeons. SLNB alone without removal of the initial positive LN is not oncologic SOC at the current time.

This trial showed us that ALND in lieu of axillary RT gives excellent outcomes after NAC. Given the relative unreliability of SLNB after NAC (compared to SLNB in the upfront setting), I think it's reasonable to give axillary RT in lieu of ALND (this is being prospectively asked on the Alliance trial in ypN(sn)+ patients, of which I do not know where they stand in regards to accrual).

A post-menopausal female with cT1N1 ER+, Her2- disease doesn't even NEED chemotherapy if she undergoes upfront surgery and oncotype testing showing a low score as per Rx-PONDER trial (a randomized, ph III). I think this trial is going to make med oncs want to give NAC to all to justify giving their therapy while trying to find a way to undercut us.

That being said, are we going to get away from radiating the lymph nodes in patients with cN1-> ypN0 disease? Likely yes, but would await the ph III B51 trial to result before doing do.

Also, 37% of the low-risk patient cohort received 'more radiation than study guideline'. Z0011 Criticism all over again (albeit better non-compliance numbers...)

 

[TheWallnerus]

Most of these patients are not clinically getting ALND anymore. TAS is a heterogeneous process even amongst breast surgeons. SLNB alone without removal of the initial positive LN is not oncologic SOC at the current time.

A pic from a slide from a presentation I gave in residency ~two decades ago:

Taking the whole view of all the data, I would propose an idea that seems anathema to the breast rad onc: chemo removes positive axillary nodes. And when you have a lady who shows herself ypN0 (by SLN or ALND) it's a good sign that the node is gone (or "enough" of the axillary cancer burden has gone).

That being said, are we going to get away from radiating the lymph nodes in patients with cN1-> ypN0 disease?

Not just no ENI. No radiation at all (for PMRT patients). Because the only indication for PMRT in more than a few mastectomy patients is the presence or suspected presence of a node at diagnosis and not any feature of the primary per se.

would await the ph III B51 trial to result

As epilogue I thought the accompanying editorial was well-written and will include it here:
....

Neoadjuvant chemotherapy for breast cancer management is increasingly used in clinical practice for large or locoregionally advanced tumours. Tumour response after neoadjuvant chemotherapy has a prognostic impact on survival. Consequently, de-escalating subsequent treatment modalities could be considered based on tumour response to neoadjuvant chemotherapy, in particular adjuvant radiotherapy, to avoid unnecessary treatment-related complications.

The RAPCHEM (BOOG 2010–03) prospective cohort study was a pioneering step to de-escalate radiotherapy in clinical practice based on the response after neoadjuvant chemotherapy for cT1–2N1 breast cancers with up to three radiologically suspicious lymph nodes (of which one had been pathologically confirmed). The tailored approach proposed by the authors relied on the definition of three risk groups, based on tumour response to neoadjuvant chemotherapy evaluated after surgery. The 5-year locoregional recurrence rate (primary outcome) was 2·2% for the whole cohort, confirming the excellent outcomes of this patient-tailored strategy. High locoregional control was achievable without any regional radiotherapy for ypN0 tumours (ie, low-risk group), evaluated on the basis of either axillary lymph node dissection (ALND) or sentinel lymph node biopsy (SLNB), and for ypN1 tumours (ie, intermediate-risk group) after ALND. In particular, radiotherapy was completely avoided in patients with ypN0 tumours after mastectomy. The results of the ongoing NSABP B-51 trial (NCT01872975), which is evaluating the eventual added benefit of lymph node irradiation in patients with ypN0 tumour status after neoadjuvant chemotherapy, are expected in 2028; however, given the outstanding results of the RAPCHEM study, adjuvant radiotherapy might be safely avoided in these patients from now on. Although there is no doubt that development of newer, state-of-the-art, radiotherapy techniques will reduce radiotherapy-associated adverse events, avoiding unnecessary irradiation shall always be the safest way to reduce radiation-induced toxicities.

ALND is being widely replaced by SLNB, which reduces the risk of arm toxicity. However, the reliability of SLNB after neoadjuvant chemotherapy to affirm an ypN0 status is subject to intense debate: it has been suggested that the frequency of false-negative findings in this context was not negligible, reaching up to 12·6% in the ACOSOG Z1071 trial. However, combining SLNB and removal of the nodes, which were initially clipped during the baseline radiological evaluation, could lower the false-negative rate to 2·0%. Nevertheless, the results of the RAPCHEM study suggest that patients with a pathological complete response to neoadjuvant chemotherapy on SLNB could be safely treated without any axillary management, thus reducing to a minimum the risk of lymphoedema. For patients with two or fewer positive SLNBs, the ACOSOG Z11 trial showed that ALND could be safely replaced by axillary radiotherapy after breast-conserving surgery. Thus, the RAPCHEM protocol logically proposed to treat patients with ypN1 tumours on SLNB with axillary radiotherapy. The excellent locoregional control shows that this approach might eventually become the standard of care, but the formal response to this question will eventually be brought by the ongoing Alliance A011202 trial (NCT01901094), evaluating the non-inferiority of axillary radiotherapy compared with ALND in case of positive SLNB after neoadjuvant chemotherapy.

The development of new effective systemic treatments also questions the need for nodal irradiation. A retrospective analysis of the HERA trial, after a follow-up of 10 years, showed that patients treated with trastuzumab for HER2-positive breast cancers with one to three lymph nodes had a significant benefit of postmastectomy radiotherapy in terms of locoregional control. However, the 7% locoregional relapse absolute risk reduction observed in this trial was much smaller than the historical 17% risk reduction reported by the Early Breast Cancer Trialists' Collaborative Group (EBCTCG) meta-analysis.Additionally, subgroup analyses showed that there was no significant benefit in case of hormone-receptor-negative HER2-positive tumours.

By contrast, the survival benefit in the EBCTCG meta-analysis showed a reduction in breast cancer mortality after 5 years. Longer follow-up is needed for RAPCHEM trial to have a conclusion in terms of survival.

Although the trend towards radiotherapy de-escalation is relevant in breast radiotherapy, several biological and molecular considerations should be kept in mind to de-escalate treatment safely. Multivariate analyses of the RAPCHEM study suggested that grade 3 and triple-negative breast cancers were associated with a higher risk of local failure. Other biomarkers, such as the presence of disseminated tumour cells, are correlated with a clinical benefit of regional irradiation. Nomograms might consequently be proposed in the future to further identify the patient subgroups in which radiotherapy de-escalation can be safely conducted.

In the era of effective systemic treatments, biomarkers, new surgical procedures, and safer radiotherapy techniques, breast cancer management has substantially improved—with the aim to reduce treatment-related toxicity. However, it is important that the patients themselves are not forgotten and should be involved in the decision-making process regarding their treatment pathway.

 

[Palex80]

ALND is being widely replaced by SLNB, which reduces the risk of arm toxicity. However, the reliability of SLNB after neoadjuvant chemotherapy to affirm an ypN0 status is subject to intense debate: it has been suggested that the frequency of false-negative findings in this context was not negligible, reaching up to 12·6% in the ACOSOG Z1071 trial. However, combining SLNB and removal of the nodes, which were initially clipped during the baseline radiological evaluation, could lower the false-negative rate to 2·0%. Nevertheless, the results of the RAPCHEM study suggest that patients with a pathological complete response to neoadjuvant chemotherapy on SLNB could be safely treated without any axillary management, thus reducing to a minimum the risk of lymphoedema. For patients with two or fewer positive SLNBs, the ACOSOG Z11 trial showed that ALND could be safely replaced by axillary radiotherapy after breast-conserving surgery. Thus, the RAPCHEM protocol logically proposed to treat patients with ypN1 tumours on SLNB with axillary radiotherapy. The excellent locoregional control shows that this approach might eventually become the standard of care, but the formal response to this question will eventually be brought by the ongoing Alliance A011202 trial (NCT01901094), evaluating the non-inferiority of axillary radiotherapy compared with ALND in case of positive SLNB after neoadjuvant chemotherapy.

I am sorry, but this sentence does not make any sense.

80% of the patients got ALND across all groups, yet RAPCHEM results suggest that SLNB is fine? Sorry, but that is not a sound conclusion, at least not from this study.

To me, RAPCHEM implications for clinical practice mean that one can de-escalate radiotherapy when doing ALND. Extrapolating the trial results to patients who only had SLNB is not supported by the data.

 

[TheWallnerus]

I am sorry, but this sentence does not make any sense.
View attachment 359291

80% of the patients got ALND across all groups, yet RAPCHEM results suggest that SLNB is fine? Sorry, but that is not a sound conclusion, at least not from this study.

To me, RAPCHEM implications for clinical practice mean that one can de-escalate radiotherapy when doing ALND. Extrapolating the trial results to patients who only had SLNB is not supported by the data.

You’re a very strict textualist about some things in breast cancer lol. There were about 60 ladies in the study who were ypN0 and didn’t get ALND. What if zero or one out of sixty had LRF at 5 years? Joe Biden yesterday told a heckler “you’re free to be an idiot.” Is it idiocy to be swayed by a 5y <1% LRF risk… or idiocy to be unswayed. (I’d be swayed.) To the authors’ credit I think their wording is that the data is “suggestive.”

 

[Palex80]

You’re a very strict textualist about some things in breast cancer lol. There were about 60 ladies in the study who were ypN0 and didn’t get ALND. What if zero or one out of sixty had LRF at 5 years? Joe Biden yesterday told a heckler “you’re free to be an idiot.” Is it idiocy to be swayed by a 5y <1% LRF risk… or idiocy to be unswayed. (I’d be swayed.) To the authors’ credit I think their wording is that the data is “suggestive.”

So, this is a 60-patient trial?

 

[TheWallnerus]

So, this is a 60-patient trial?

I said "about 60." It was actually 57 to be more precise. I checked the supplementary material. I will ask you count this as zero points for either one of us! 4 out of 234 ALND ladies had recurrence at 5 years vs 2 out of 57 non-ALND ladies at five years (p=0.419).

I had hoped it would be zero or one out of 57! You have convinced me ALND is necessary if giving no PMRT in cT1-2N1/ypN0 disease

 

[evilbooyaa]

@TheWallnerus Yes, just as B51 will eliminate PMRT in cN1--> ypN0 in patients undergoing mastectomy. Assuming epxerimental arm has reasonable outcomes.
Again, making the leap based on a single arm prospective registry is.... not ideal. Are you de-escalating to 30Gy to elective nodes in p16+ H&N based on MSKCC's outstanding results? I mean, clearly, if the K-M curve looks SO GOOD it'd be consistent to just stick with it, right?

I did go look at the authors of the editorial to nitpick abou ttheir specialty, but appears they're both rad oncs from france. Maybe clin oncs.

I agree with Palex - saying there was no differences in 60 patients (which made up 15-20% of the population) and thus this is totally kosher (compared to, well not compare dto anything) is not ideal.

 

[communitydoc13]

80% of the patients got ALND across all groups, yet RAPCHEM results suggest that SLNB is fine? Sorry, but that is not a sound conclusion, at least not from this study.

To me, RAPCHEM implications for clinical practice mean that one can de-escalate radiotherapy when doing ALND. Extrapolating the trial results to patients who only had SLNB is not supported by the data.

So, this is a 60-patient trial?

RAPCHEM is an impossible study to draw inferences from except for one:

1. Employing a prospective therapeutic algorithm to escalate or de-escalate locoregional therapy in patients who received neoadjuvant chemotherapy, seems to result in reasonable outcomes within a 5 year period when compliance with said algorithm is not terribly good.

Among the problematic things in the paper is that they excluded synchronous locoregional and distant recurrences from locoregional recurrence. In a group of patients that are predominantly triple negative, this alone may have a big impact.

Lots of patients who weren't supposed to get RT got it.

But it is undoubtedly true that we give RT to many patients that don't need it. Historically, we felt good about giving radiation when we thought it might have a 10-20% chance of benefitting a patient. This is of course a much higher standard than most medical management.

Also true that in our field we are not trying out new therapeutics regularly. The easiest and most significant contribution most academic radoncs will make is to prospectively study de-escalation of our one therapeutic. It it much more likely to be viewed as successful than the converse.

Que sera sera

 

[Palex80]

I said "about 60." It was actually 57 to be more precise. I checked the supplementary material. I will ask you count this as zero points for either one of us! 4 out of 234 ALND ladies had recurrence at 5 years vs 2 out of 57 non-ALND ladies at five years (p=0.419).

I had hoped it would be zero or one out of 57! You have convinced me ALND is necessary if giving no PMRT in cT1-2N1/ypN0 disease

Oh, my!
So the axillary recurrence rate goes up by 100%, when you don't do ALND but only SLNB?

 

[TheWallnerus]

1. Employing a prospective therapeutic algorithm to escalate or de-escalate locoregional therapy in patients who received neoadjuvant chemotherapy, seems to result in reasonable outcomes within a 5 year period when compliance with said algorithm is not terribly good.

Among the problematic things in the paper is that they excluded synchronous locoregional and distant recurrences from locoregional recurrence. In a group of patients that are predominantly triple negative, this alone may have a big impact.

You guys. Total gainsayers! Speaking of triple negative. It was totally not on my radar these early stage TNBC patients are getting Keytruda now. And when that happens:

Are you de-escalating to 30Gy to elective nodes in p16+ H&N based on MSKCC's outstanding results?

Uh... no comment. But 30 Gy is a wonderful dose for cancer cells containing HPV DNA.

So the axillary recurrence rate goes up by 100%, when you don't do ALND but only SLNB?

If there had been just one less patient having LRF in the non-ALND arm, the rates would have been identical. Two less, the recurrence rate in the ALND arm would have been infinitely higher than the non-ALND arm. There's a 'This Is Spinal Tap' amplifier joke in here somewhere.

 

[Palex80]

If there had been just one less patient having LRF in the non-ALND arm, the rates would have been identical. Two less, the recurrence rate in the ALND arm would have been infinitely higher than the non-ALND arm. There's a 'This Is Spinal Tap' amplifier joke in here somewhere.

And if there were 2 more, it would be three times as high!

 

[Mandelin Rain]

And if there were 2 more, it would be three times as high!

This is why inferences from 57 patient subgroups in single arm trials don't change (my) practice.

 

[metallica81788]

You people sure do make treating breast seem really complicated