- Joined
- Apr 3, 2019
- Messages
- 4,280
- Reaction score
- 10,006
Yeah. I would be absolutely comfortable skipping PMRT in a lady who’s cT1 and a solitary node who goes through neoadj chemo and becomes pCR. Single lymph node positive ladies were always a soft call for PMRT anyways, the vast bulk of the “proven” benefit there coming from a certain Danish study. For women who are LN positive in one node and have a pCR? Radiation therapy doesn’t help them. (I said what I said!) I see the study applying to a somewhat narrow subset: small T, small N women who get ypN0 (and these are usually not ER+ ladies who achieve that). I would additionally point out though: no women (including ypN2-3) got sclav or IMN RT in the study and look at those LRRs.I am not sure this trial changes a great deal of what I would do for most patients. The majority of cN1 patients are postmenopausal and have ER+ and HER2 negative tumors. Neoadjuvant chemotherapy should only be given if oncotype is high. Are you comfortable skipping RT for those with oncotype high based on this? I wouldn't be without more data based on genomics.
When has a breast study changed after the 5 year ? I can’t think of one5 year follow up, 'nuff said
Americans: we need 10-15y dataWhen has a breast study changed after the 5 year ? I can’t think of one
I think it took longer fu with the ebctcg meta-analysis to show an OS benefit based on LC earlier on right?When is the last time a study showed no statistical difference between RT/no RT at 5y but statistical significance appeared at 10y in favor of RT. One of the log-rank’s assumptions (for the null) is that there is no difference between the curves at any and ALL time points. It’s … quite unlikely … that significant differences (between curves) appear at 10y vs 5y. More likely is what happened eg in the “Upfront Outback” trial; at first, differences, and later, not.
I think this is ... a good point. OTOH, we are talking meta-analysis on data from trials that started accruing in the 70s and 80s. Because lifestyle, staging, surgical philosophy, and systemics have changed so profoundly since then (and breast cancer survival has steadily improved on account of that), it's also not a bad point that one day we may have to stop referring to pro-RT studies in this disease from last century.I think it took longer fu with the ebctcg meta-analysis to show an OS benefit based on LC earlier on right?
Um, your example is a meta-analysis?I think it took longer fu with the ebctcg meta-analysis to show an OS benefit based on LC earlier on right?
A pretty well known one? Did this not make the headlines at most programs when it came out?Um, your example is a meta-analysis?
It's not only thr Danish studies. Both EORTC and MA.20 showed a benefit to RNI for women with 1 to 3 nodes. Like half the patients had 1 Ln. Granted only one of the trials allowed mastectomy, but the point is that RNI makes a difference even in small nodal burden. Someone who has a PCR likely has less benefit, but again my point is these days it would be pCR in a patient with high oncotype most of the timeYeah. I would be absolutely comfortable skipping PMRT in a lady who’s cT1 and a solitary node who goes through neoadj chemo and becomes pCR. Single lymph node positive ladies were always a soft call for PMRT anyways, the vast bulk of the “proven” benefit there coming from a certain Danish study. For women who are LN positive in one node and have a pCR? Radiation therapy doesn’t help them. (I said what I said!) I see the study applying to a somewhat narrow subset: small T, small N women who get ypN0 (and these are usually not ER+ ladies who achieve that). I would additionally point out though: no women (including ypN2-3) got sclav or IMN RT in the study and look at those LRRs.
There was no overall survival benefit in either study further confirming the "Alternative Hypothesis." The most statistically significant results from these studies were on the toxicity side. But I think you're mixing metaphors a bit (NB: my post was entitled "No more PMRT for cN1 ypN0 breast cancer?"). MA20/EORTC 22922 didn't appear until 2015 and were not the raison d'être for PMRT in pN1 disease. Prior to ~2010-15, it was not routine to give PMRT to cT1-2N1 patients. It was the (yet another, different) EBCTCG meta-analysis whose results* (if you dig a little deeply the results were driven strongly by the Danish data) caused the switcheroo as it were.t's not only thr Danish studies. Both EORTC and MA.20 showed a benefit to RNI for women with 1 to 3 nodes.
Flippant is fine lol. There was a slow, but steady, separation of the curves both for OS and LC seen in the meta-analyses that were not seen in the trials themselves. I will grant you that. But grant me this... here I have superimposed the RAPCHEM curve onto the 2014 EBCTG for the appropriate comparator subgroup of N1 patients. Breast cancer patients (specifically T1-2N1 patients) in the 21st century recur at very different rates than breast cancer patients in the 20th century, irrespective of the radiotherapy package.I wrote my initial response flippantly while sitting at a restaurant with my kids and mostly to provocate. I now follow up while watching TV on my couch, still sans literature review.
We are definitely not moving the needle much in these low volume disease treatment responders. I'm all about getting out the game whenever I can. You could just stop radiating regional nodes for the rest of your career and your damage done would be negligible (but not zero). A few comments though, many of these studies such as ma20, eortc 22922, and yes the much maligned but oft quoted ebctcg meta-analyses reported out their results at 10-15 years. Did they even report 5 year outcomes?? The whole shaking your fist at WHEN DID ANYTHING EVER CHANGE FROM 5 TO 10 seems hollow when some studies have just waited for their data to mature.
Breast can be the long game. I've seen many patients recur late late late with local or regional disease only. My grandmother had a 20 year regional recurrence that eventually took her life.
5 years is trash. These patients can be on endocrine therapy for that entire duration that smudges out any result.
I think we see eye to eye and alluded to that in my postFlippant is fine lol. There was a slow, but steady, separation of the curves both for OS and LC seen in the meta-analyses that were not seen in the trials themselves. I will grant you that. But grant me this... here I have superimposed the RAPCHEM curve onto the 2014 EBCTG for the appropriate comparator subgroup of N1 patients. Breast cancer patients (specifically T1-2N1 patients) in the 21st century recur at very different rates than breast cancer patients in the 20th century, irrespective of the radiotherapy package.
Bury this thread and resurrect it in 5 years!5 year follow up, 'nuff said
I believe some of the „omission of RT“ trials in early-stage, low risk breast cancer, which means tamoxifen +/- RT showed similar rates of locoregional recurrences at 5 years between both arms, but LRC-rates went up, once tamoxifen was terminated.Ok, let me try again
Has any individual breast study changed after the 5 year ?
Calgb and prime 2 both separated by 5 years and were statistically significant.I believe some of the „omission of RT“ trials in early-stage, low risk breast cancer, which means tamoxifen +/- RT showed similar rates of locoregional recurrences at 5 years between both arms, but LRC-rates went up, once tamoxifen was terminated.
Sometimes one needs 5 year data; sometimes one don't.Calgb and prime 2 both separated by 5 years and were statistically significant.
“But we need >5 years!”
It’s always the case for breast. I used to be one of them. But, 5 years tells the story. If you base practice on 5 year follow up for RCTs for breast cancer patients, you will be treating your patients with highest level of evidence.
You need a with-RT control arm when without RT arm had a 2% LRF at 5 years? Ok. If we are being super strict.All this talk of 5 vs 10 year follow up is funny. There is no control arm for it to become significantly different from.
The LRF goes up slowly over time in BOTH irradiated and unirradiated patients (see above). So while waiting to see a between-arm difference at ten years that didn’t appear at 5, the “runners” (arms) usually beginning “running” at the same speed and the between-arm-delta doesn’t change as much toward the tail of the curve as it did at the beginning.The results will absolutely be different at 10 years. We know from basically every breast study in existence that LRF slowly goes up over time.
It depends...Be all that as it may…
It has struck me, looking back into the history of pro-PMRT for ladies with 1-3 nodes positive and getting chemo (lots of melphalan!), that all the randomized data supporting PMRT comes from ~1200 ladies across ~20-30 studies treated between about 1965 and 1985. And even then when we cull all that data together (again as shown above in the EBCTCG), in the unirradiated ladies the incidence of per-year LRF in years 5-10 is <1% (was ~17% at 5y and ~21% at 10y… and, again, this number is now ~2% at 5y for women chemotherapized but unirradiated in the modern era… showing how much things have changed IMHO).
Top left graph germane to “argument” here and the Danish data is subsumed, I assume, in the EBCTCG graph.It depends...
The Dutch have shown some other data too, in the past:
View attachment 359166
Is the benefit of postmastectomy irradiation limited to patients with four or more positive nodes, as recommended in international consensus reports? A subgroup analysis of the DBCG 82 b&c randomized trials - PubMed
The survival benefit after postmastectomy RT was substantial and similar in patients with 1-3 and 4+ positive lymph nodes. Furthermore, it was not strictly associated with the risk of loco-regional recurrence, which was most pronounced in patients with 4+ positive nodes. The indication for RT...pubmed.ncbi.nlm.nih.gov
I do not think that looking only at locoregional control rates is the optimal endpoint for trials on PMRT.
You may find it odd, but there are two important arguments here:
a) PMRT may be more valuable in reducing distant metastases (MA.20 gave a hint in that direction)
b) Really distinguishing LR failures is tricky in breast cancer. These patients do not undergo routine imaging to look at nodes and disease progression if often picked up when symptomatic distant metastases arise. It is quite possible that the effect of PMRT is being missed, simply because we do not perform any imaging in those women (not that I am advocating in favor of doing any, I am just stating that it is difficult to measure an endpoint, if you are not examining it at all in a systematic way).
It shows where things are headed. Can we omit xrt today? I don’t know, but drugs are only improving and I am confident that at some point in my career these patients will not be getting xrt.I am not sure this trial changes a great deal of what I would do for most patients. The majority of cN1 patients are postmenopausal and have ER+ and HER2 negative tumors. Neoadjuvant chemotherapy should only be given if oncotype is high. Are you comfortable skipping RT for those with oncotype high based on this? I wouldn't be without more data based on genomics.
Local recurrences can be obnoxious to patients, can't they?Top left graph germane to “argument” here and the Danish data is subsumed, I assume, in the EBCTCG graph.
I get your points about LRF. If PMRT is going to improve survival it’s got to affect more than just LRF. (However we will never achieve OS advantages in arenas where LRF is less than 5% long term but I digress.) But we are radiation oncologists, and you know what Monica Morrow said about us…
View attachment 359168
Absolutely. Is it now a tautology that overall survival advantages from RT will always be less than LC advantages? I think so. And in this trial (a subset of) women with positive nodes and no RT had a one in fifty chance of LRF at 5y, which suggests to me a NNT of at best approximately 100 for LC advantage w/ RT in that subset. OS NNT… in the thousands?Local recurrences can be obnoxious to patients, can't they?
On the basis of this and many other reports, I would omit PMRT in a lady who had evidence of N1 pre chemo, a small primary, and who was ypN0 post chemo. I don’t need to be hit over the head, or the boob. Next up: HER2 patients (and triple negative even) with pCR. Like a quote I read from a med onc: “We’ve followed 2000 patients with HER2 disease and have yet to see a single local failure.”It shows where things are headed. Can we omit xrt today? I don’t know, but drugs are only improving and I am confident that at some point in my career these patients will not be getting xrt.
I generally agree, however some more questions need to be adressed.On the basis of this and many other reports, I would omit PMRT in a lady who had evidence of N1 pre chemo, a small primary, and who was ypN0 post chemo.
I have never seen the effect of RT to be greater for either DMFS or OS in an irradiated group than the LRFS effect. Never. When you say “benefit little… in terms of LR control but… have a meaningful OS adv,” I’m hearing you say *absolute* OS advantage can be > LR advantage, and I just have NEVER seen that. I don’t even know mechanistically how it would work (curing cancers that were never coming back is not magic, it’s illogical).The point I was trying to make quoting the old Durch data is that the general rule of "prevent 4 recurrences and save the life of 1 patient", does not seem to always uphold. Patients with small (absolute) locoregional recurrence riks benefit little from PMRT in terms of locoregional control, but they may have a meaningful OS advantage because of it.
I should however have rather quoted this analysis, sorry for the confusion:
The unirradiated cN1 women in this study got SLN analysis. In my brain, I have made SLN and ALND completely totally fungible at the oncologic level in breast cancer. We all should IMHO. You can not rationally make any radiotherapy decisions on the basis of “I wish the lady had gotten her axilla dissected” anymore. “Know[ing] a bit more about the exact initial disease extent in the nodes” is bootless; it’s falling back into the angsty “but what if” arguments in the immediate post Z0011 days. Know what you can know (by PET or US or MRI or whatever), let her get “maximal” (sans RT) treatment (which should not include axillary dissection), and go on with life.I generally agree, however some more questions need to be adressed.
1. What surgical axillary management would you see as adequate to follow this route? Would you need an ALND proving „true“ ypN0? Would you also accept a TAS or even a SLN-excision showing ypN0?
2. What would you accept as initial staging? Ultrasound of the axilla (with biopsy)? MRI? PET-CT? If you may be planning to skip PMRT you should probably know a bit more about the exact initial disease extent in the nodes, wouldn’t you agree?
This is true. But sometimes OS-benefit can be similar to LR benefit.I have never seen the effect of RT to be greater for either DMFS or OS in an irradiated group than the LRFS effect. Never. When you say “benefit little… in terms of LR control but… have a meaningful OS adv,” I’m hearing you say *absolute* OS advantage can be > LR advantage, and I just have NEVER seen that. I don’t even know mechanistically how it would work (curing cancers that were never coming back is not magic, it’s illogical).
2% sounds good for PMRT. Lot's of breast cancer patients used to get 6 months of chemo for a comparable OS benefit.The unirradiated cN1 women in this study had a 2% LR failure at 5 years. This certainly sets a hard limit on the LR control advantage RT could offer here at 5 years: it would be a 2% advantage, at most. And, it also sets an *OS* advantage limit: it can not conceivably be >2%, and we both know it would be A LOT less.
Not really.The unirradiated cN1 women in this study got SLN analysis.
Oh no! I refuse to go on with this discussion, you just brought up Z0011!!!In my brain, I have made SLN and ALND completely totally fungible at the oncologic level in breast cancer. We all should IMHO. You can not rationally make any radiotherapy decisions on the basis of “I wish the lady had gotten her axilla dissected” anymore. “Know[ing] a bit more about the exact initial disease extent in the nodes” is bootless; it’s falling back into the angsty “but what if” arguments in the immediate post Z0011 days.
Well, "maximal" treatment can potentially include ALND then. I would prefer RT over ALND.Know what you can know (by PET or US or MRI or whatever), let her get “maximal” (sans RT) treatment (which should not include axillary dissection), and go on with life.
Correct me if I’m wrong but every time ALND has been tested against SLN there have been no significant differences in major outcomes long term. Every single time. How much data will it take for people to move on from thinking ALND accomplishes anything. ALND is very morbid compared to SLN.This is true. But sometimes OS-benefit can be similar to LR benefit.
2% sounds good for PMRT. Lot's of breast cancer patients used to get 6 months of chemo for a comparable OS benefit.
Not really.
View attachment 359176
View attachment 359175
The unirradiated women, were the ypN0-group (Low Risk). 80% of them got ALND.
Oh no! I refuse to go on with this discussion, you just brought up Z0011!!!
Z0011 was in cN0, this is cN1. Totally different scenario. No study has yet established that you defer ALND in cN1 patients, many centers do TAS (which is fine probably), but SLNB only? Nope sorry, that's two steps too far, in my opinion.
Well, "maximal" treatment can potentially include ALND then. I would prefer RT over ALND.
SLNB without any sort of local adjuvant therapy for cN+ has never been tested against ALND.Correct me if I’m wrong but every time ALND has been tested against SLN there have been no significant differences in major outcomes long term. Every single time. How much data will it take for people to move on from thinking ALND accomplishes anything. ALND is very morbid compared to SLN.
Cause we irradiate the IMN, just like we irradiate the axilla (if it's not properly dissected).EDIT: Sorry to do a late edit. But it’s reasonable to think one out of five of these cN1 ladies would have positive IMNs on dissection of the IMNs. If ALND is needed for axillary cN1, why isn’t IMN dissection too?
Oh Lordy. Now we are parsing between cN1 and pN1. In Z0011 there were a lot of pN1 SLN+ women that got no “local therapy” (ALND) to the axilla, and even a lot that got no RT at all (in violation of protocol). We all know how they did long term. I thought we (the Bernie Fisherian we) were teaching everyone that the axilla (and ALND especially) is prognostic, not therapeutic. Guess not. I almost can’t believe you want a lady with a single node on ultrasound and later ypN0 to get her axilla majorly surgerized (or radiotherapized) during her mastectomy when all we can truly be sure of (we can’t be sure it helps oncologically) is that her lymphedema risk and QOL will be affected.SLNB without any sort of local adjuvant therapy for cN+ has never been tested against ALND
Who knows. Medoncs make that determination. Supposedly the vast majority of women would undergo toxic chemo for 4% absolute improvement in survival.Bit off subject here and not subtle, but do we really think it is a good idea for women to be having all this (neoadjuvant) systemic therapy? The long term toxicity is far worse than any local treatment could ever cause.
I think thats the problem with our specialty. We will cannibalise one another over seemingly anything but will allow the other specialties to do as they please with no focus on the actual patient.Who knows. Medoncs make that determination. Supposedly the vast majority of women would undergo toxic chemo for 4% absolute improvement in survival.
Of course we are. These are two, totally distinct settings.Oh Lordy. Now we are parsing between cN1 and pN1.
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.In Z0011 there were a lot of pN1 SLN+ women that got no “local therapy” (ALND) to the axilla, and even a lot that got no RT at all (in violation of protocol).
ALND is both.We all know how they did long term. I thought we (the Bernie Fisherian we) were teaching everyone that the axilla (and ALND especially) is prognostic, not therapeutic. Guess not.
I dont "want" something.I almost can’t believe you want a lady with a single node on ultrasound and later ypN0 to get her axilla majorly surgerized (or radiotherapized) during her mastectomy when all we can truly be sure of (we can’t be sure it helps oncologically) is that her lymphedema risk and QOL will be affected.
That’s not what Bernie found.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.
It’s how we are trained. Look at the big wig ASTRO heads. Quick to judge/look down on the community docs but first to kiss the rings of their colleagues and looking for any excuse to not have to give or look for any indication to give less radiation.I think thats the problem with our specialty. We will cannibalise one another over seemingly anything but will allow the other specialties to do as they please with no focus on the actual patient.