Breast: one and done

[scarbrtj]

Post-lump breast RT done in 5 fractions... old and busted.

But now, one and done (in the OR), we can confidently say. (Which takes us from about 3 patients/day on beam for breast to ~0).

Discuss.

"For patients with early breast cancer who met our trial selection criteria (T1-2 [0-3.5cm]/N0-1, 45y+, IORT N=1140, EBRT N=1158; ~9-19y followup), risk adapted immediate single dose TARGIT-IORT during lumpectomy was an effective alternative to EBRT, with comparable long term efficacy for cancer control and lower non-breast cancer mortality (0.59, 0.40 to 0.86, P=0.005). TARGIT-IORT should be discussed with eligible patients when breast conserving surgery is planned."

---

Members don't see this ad.

 

[OTN]

I completely disagree with the conclusions from the TARGIT-IORT trial, and so did this reviewer, who put it more eloquently than I have the time to do:

"Vaidya et al1 report 5-year outcomes in 1153 patients enrolled in the TARGIT-A trial of intraoperative radiotherapy (IORT) vs standard-of-care external beam radiotherapy (EBRT) for low-risk breast cancer in the postpathology setting. Their report provides a somewhat rare opportunity to revisit the debate in the literature at the time of the original publication of trial outcomes. The following points take stock of what we now know. First, not only is delayed second-procedure targeted IORT not noninferior to EBRT, but it is actually highly significantly inferior to EBRT with local recurrence rates at 5 years of 3.96% after IORT vs 1.05% after EBRT (2-tailed P = .002). Second, IORT may be “better than nothing,”2 but certainly not by much. The recurrence rate after IORT is 3.96% (95% CI, 2.7%-5.9%), very close to the 4.1% (95% CI, 2.4%-5.7%) local recurrence rate observed in those randomized to no radiotherapy in the PRIME trial3 after a median follow-up of 5 years. Interestingly, the rate after whole-breast radiotherapy in the control arm of PRIME was 1.3% (95% CI, 0.2%-2.3%), in very good agreement with the outcome in the EBRT arm of the TARGIT-A trial. Third, any—nonsignificant at the time—imbalances in noncancer-related mortality are long gone, and with it the suggestion that IORT should spare potentially fatal adverse effects of EBRT. To summarize, while hindsight is 20/20, the take-home message is that the warning flags were there4 already in the statistical analysis of the original publication from the TARGIT-A trial.5 Now we know."

 

[scarbrtj]

I completely disagree with the conclusions from the TARGIT-IORT trial, and so did this reviewer, who put it more eloquently than I have the time to do:

"Vaidya et al1 report 5-year outcomes in 1153 patients enrolled in the TARGIT-A trial of intraoperative radiotherapy (IORT) vs standard-of-care external beam radiotherapy (EBRT) for low-risk breast cancer in the postpathology setting. Their report provides a somewhat rare opportunity to revisit the debate in the literature at the time of the original publication of trial outcomes. The following points take stock of what we now know. First, not only is delayed second-procedure targeted IORT not noninferior to EBRT, but it is actually highly significantly inferior to EBRT with local recurrence rates at 5 years of 3.96% after IORT vs 1.05% after EBRT (2-tailed P = .002). Second, IORT may be “better than nothing,”2 but certainly not by much. The recurrence rate after IORT is 3.96% (95% CI, 2.7%-5.9%), very close to the 4.1% (95% CI, 2.4%-5.7%) local recurrence rate observed in those randomized to no radiotherapy in the PRIME trial3 after a median follow-up of 5 years. Interestingly, the rate after whole-breast radiotherapy in the control arm of PRIME was 1.3% (95% CI, 0.2%-2.3%), in very good agreement with the outcome in the EBRT arm of the TARGIT-A trial. Third, any—nonsignificant at the time—imbalances in noncancer-related mortality are long gone, and with it the suggestion that IORT should spare potentially fatal adverse effects of EBRT. To summarize, while hindsight is 20/20, the take-home message is that the warning flags were there4 already in the statistical analysis of the original publication from the TARGIT-A trial.5 Now we know."

I hadn't followed followed TARGIT 'til now, but this analysis sailed over my head....

 

[Gfunk6]

Since IORT improves your survival from non-breast cancer related causes, maybe we can use it on COVID patients?

 

[evilbooyaa]

Was going to say, Soren Bentzen ripped the 5-year results in JAMA Onc a pretty big one already. What is this thing published in BMJ as well? I was always confused by this concept of delayed IORT - that's not how its done in my experience.

I will continue to counsel my patients on a slightly higher (~2-3%) risk of LR when doing IORT as I have been.

 

[scarbrtj]

I will continue to counsel my patients on a slightly higher (~2-3%) risk of LR when doing IORT as I have been.

I mean, I thought so too, but the invasive recurrence rates are essentially the same (~10%) at 12y in ~2000 patients. That's pretty high-quality (absence of there being a significant difference) evidence. TBH I don't think IORT will ever be popular in the US, but I think that has purely to do with remuneration and not clinical/convenience advantages/disadvantages. Now, I can't see really even one meaningful clinical disadvantage of IORT, and the balance if anything is a little more on the advantage side. Certainly advantage on convenience side. And, an advantage on cost-to-society side. Just one more thing to fret about!

 

[Gfunk6]

Third, any—nonsignificant at the time—imbalances in noncancer-related mortality are long gone, and with it the suggestion that IORT should spare potentially fatal adverse effects of EBRT.

Could someone please expound on this comment for my benefit? I'm not quite getting the point that is being raised.

 

[StIGMA]

Could someone please expound on this comment for my benefit? I'm not quite getting the point that is being raised.

I think that person is arguing against the study conclusions. However that sentence is very poorly written.

From the TARGIT collaborative group, sent out last Friday:
"Benefits of TARGIT-IORT during lumpectomy for early stage invasive ductal breast cancer:

• less travel, thus less cost
• less pain
• better quality of life
• cosmetically superior
• fewer non-breast cancer deaths "

I think most would agree that the last point should not be the primary rationale for giving IORT. This is in large part because there is no clear mechanistic basis for it. I think the reviewer is arguing that because the groups are balanced, there is no rational reason for a non-breast cancer OS difference.

 

[Ray D. Ayshun]

What if 4 women on the ebrt arm who became friends got into a head-on mva with 4 dudes surveiling their PCA on the protect trial?

 

[Gfunk6]

I think that person is arguing against the study conclusions. However that sentence is very poorly written.

From the TARGIT collaborative group, sent out last Friday:
"Benefits of TARGIT-IORT during lumpectomy for early stage invasive ductal breast cancer:

• less travel, thus less cost
• less pain
• better quality of life
• cosmetically superior
• fewer non-breast cancer deaths "

I think most would agree that the last point should not be the primary rationale for giving IORT. This is in large part because there is no clear mechanistic basis for it. I think the reviewer is arguing that because the groups are balanced, there is no rational reason for a non-breast cancer OS difference.

That makes a lot more sense, thank you

 

[deleted605854]

The data on IORT is mixed IMO and the logistics of it plus the necessary equipment means it'll never be that popular.

It's going to be one of those things that one cancer center might get to advertise to patients to draw them in.

My money is on definitive rads for certain patients with breast cancer

ClinicalTrials.gov

clinicaltrials.gov

Come on Canada, win another one for the field.

 

[Palex80]

TARGIT has major flaws. One of the main issues in the trial was the inclusion of many patients who you would not consider suitable for PBI, for instance cN1, yes cN1.

This is the same problem that popped up with the Eliot trial.

https://www.thelancet.com/journals/lanonc/article/PIIS1470-2045(13)70497-2/fulltext

which also showed inferior results.

5 x 6 Gy external beam PBI every other day is a great schedule. You have ample time to pick out the patients that will really benefit from such a schedule, not having to worry about all the uncertainties of treating patients with PBI in the OR that would have actually needed WBRT. We worked with Intrabeam where I originally trained and I can recall several patients where we had to deliver WBRT after Intrabeam based on the final pathology report. I am not familiar with any data showing what the long-term cosmetic effect would be in patients who had to undergo both procedures and I am not aware of any data showing you can safely hypofractionate them --> 5 weeks of WBRT following a (in retrospect) wrongly indicated intraoperative Intrabeam treatment?

 

[scarbrtj]

This is in large part because there is no clear mechanistic basis for it [excess deaths from breast RT]. I think the reviewer is arguing that because the groups are balanced, there is no rational reason for a non-breast cancer OS difference.

Tell that to Ralph Weischelbaum and David Kirsch who say that **1 Gy** of thoracic irradiation (and breast irradiation is a form of thoracic irradiation) causes an extra 7 deaths per 100 people. This is what they should have written an editorial about: the thousands of women killed per year from breast RT! But in seriousness, I do believe that in a study of thousand(s) of people getting significant (ie far above background doses) external beam dose to an organ like the lung or heart, we could see a signal of excess deaths albeit a small signal.

TARGIT has major flaws. One of the main issues in the trial was the inclusion of many patients who you would not consider suitable for PBI, for instance cN1, yes cN1.

Flaw/bug, or feature? They included cN1, so what? As has been discussed here previously (wink wink), these patients can cogently/acceptably still just receive breast RT and zero ENI and have fine outcomes. "Suitability for PBI" for just hard core pN0 is more emotionally driven than data driven today... just ask these trialists.

 

[communitydoc13]

This is a critical table. It enforces what we already knew about ER+ breast cancer (the overwhelming portion of the cohort here). Namely, that local recurrences increase in time and become more significant after the first five years (often with endocrine therapy) after surgery. I would put observation up against anything for a 5 year only f/u with endocrine therapy, and this is why we of course no longer treat many older women with adjuvant XRT.

Looking at the >5 year data here (and considering the marked decrease in evaluable patients starting at 6 years) one could estimate an (absolute and not Hahnian) improvement in local control after the 5 year period of at least 4% in this cohort (with only about a third of patients evaluable at 10 years). I would expect this benefit to increase over time. Not sure that this data would sell me on IORT for a 50-60 year old.

In addition, excess deaths from whole breast XRT within first five years only and not subsequently, as I think mentioned above. I'm not convinced I'm killing patients by heart disease within five years with modern techniques and lots of population based data to make one question this.

Their inclusion criteria was very broad (undoubtedly for accrual purposes) and institutions could define exclusion criteria for the experimental arm (including positive nodes and high grade), which I believe ensured whole breast radiation after IORT for many of these patients. As such I don't really know how many higher risk patients (triple negative, node positive, LVSI) actually received IORT alone.

 

[evilbooyaa]

Could someone please expound on this comment for my benefit? I'm not quite getting the point that is being raised.

There was a musing by Vaidya in the original paper that this was due to cardiac sparing effects of IORT compared to EBRT, when in reality it was likely just a type II error.

@scarbrtj they just published these results in JAMA Onc just a few months ago (Effect of Delayed Targeted Intraoperative Radiotherapy vs Whole-Breast Radiotherapy on Local Recurrence and Survival: Long-term Results From the TARGIT-A Randomized Clinical Trial in Early Breast Cancer - PubMed) which makes me question this separate (but not separate?) publication in BMJ.

An issue from their pre-pathology cohort (meaning IORT done at time of initial surgery) back at their original 2/3-year f/u publication was that a significant proportion of those women then needed WBI on top of it (50/25, no boost as IORT was the boost).

Here it is again - nearly 25% of patients that received IORT ALSO received EBRT:

Critics made a huge stink of it 5-years ago when the difference in LR was < 2.5% but was statistically significantly different. I see they spend ample time in their BMJ paper pontificating on how only they know statistical analysis and the rest of us who poo-poo this are just jealous idjeets (honest to god this is copied and pasted from their BMJ paper's M&M):

We designed the trial as a non-inferiority trial. Non-inferiority trials in cancer are performed to test new treatments that have obvious non-oncological advantages, such as better access, convenience, or quality of life for the patient, or reduced costs for the healthcare system. The non-inferiority statistical test for such a comparison is not meant to check for superiority, but to assess if the difference is within an acceptable margin and the experimental treatment is not meaningfully worse than the control. Therefore, whether the difference seen between the two randomised arms is statistically significant is not relevant here. As long as the absolute difference is not clinically significant, the new treatment would be deemed non-inferior.27 Any chosen non-inferiority margin must be one that clinicians and patients agree is an acceptable difference for the sake of the other benefits. These benefits might include lower toxicity, better cosmetic outcome, better quality of life, and overall patient preference. Therefore, in the original protocol, non-inferiority was specified as being achieved if the difference in the binomial proportions of local recurrence rate at five years did not cross a stringent margin of 2.5% in absolute terms; that is, local recurrence risk with TARGIT-IORT minus local recurrence risk with EBRT should not be more than 0.025 (2.5%). In the 2013 analysis, an even more rigorous criterion was used, specifying that the upper 90% confidence interval of the absolute difference must not exceed 0.025 (2.5%).

The 2.5% non-inferiority margin in the TARGIT-A trial is a relevant, relatively stringent margin. Patient preference studies in the United States, Australia, and Europe suggest that 2.5% is an acceptable margin.282930 Importantly, it is widely regarded as a safe margin because it is well established that a local recurrence difference of less than 10% at five years does not worsen breast cancer survival31; that is, when the risk in arm A minus the risk in arm B is less than 0.1 (10%). A large increase in local recurrences (>10% at five years) is required to lead to increased mortality because they can be effectively treated. For example, a 20% increase in local recurrence (a risk increase by 0.2) would cause a 5% increase in deaths (a mortality risk increase by 0.05)31; this was the basis for the ethics approval of this trial. In the ELIOT trial, which also investigated intraoperative radiotherapy, the non-inferiority margin was set at 7%.15 In recently reported trials of systemic therapy, the margin of a 3% difference in disease-free survival was considered acceptable.32 /quote]

 

[RadRadRad]

TARGIT has major flaws. One of the main issues in the trial was the inclusion of many patients who you would not consider suitable for PBI, for instance cN1, yes cN1.

This is the same problem that popped up with the Eliot trial.

https://www.thelancet.com/journals/lanonc/article/PIIS1470-2045(13)70497-2/fulltext

which also showed inferior results.

5 x 6 Gy external beam PBI every other day is a great schedule. You have ample time to pick out the patients that will really benefit from such a schedule, not having to worry about all the uncertainties of treating patients with PBI in the OR that would have actually needed WBRT. We worked with Intrabeam where I originally trained and I can recall several patients where we had to deliver WBRT after Intrabeam based on the final pathology report. I am not familiar with any data showing what the long-term cosmetic effect would be in patients who had to undergo both procedures and I am not aware of any data showing you can safely hypofractionate them --> 5 weeks of WBRT following a (in retrospect) wrongly indicated intraoperative Intrabeam treatment?

seen poor cosmetic outcome with combo
Agree with 30 in 5.
Initial target report was statistical malpractice and had mean follow up of <2 years while arguing in discussion that most breast cancers recur in first 2 years. Of course this is wildly inaccurate for luminal a breast cancer with about half of recurrences occurring between years 5-10.
Wouldn’t trust their data

 

[scarbrtj]

There was a musing by Vaidya in the original paper that this was due to cardiac sparing effects of IORT compared to EBRT, when in reality it was likely just a type II error.

@scarbrtj they just published these results in JAMA Onc just a few months ago (Effect of Delayed Targeted Intraoperative Radiotherapy vs Whole-Breast Radiotherapy on Local Recurrence and Survival: Long-term Results From the TARGIT-A Randomized Clinical Trial in Early Breast Cancer - PubMed) which makes me question this separate (but not separate?) publication in BMJ.

I was gonna ask you to link to the "other" TARGIT; thanks, I hadn't seen it. I had heard of it (we've all heard of IORT), but this "new TARGIT" was of a different flavor, more compelling IMHO. Hard to know what to make of that JAMA TARGIT of N=~1100 and this BMJ TARGIT that is N=~2200 and the LRs are essentially the same between the arms.

 

[gmsquid]

It’s interesting that they used LRFS with Kaplan Meier instead of LRR by Fine-Gray with death as a competing risk. Usually LRFS would imply death as an event. If there was a statistical anomaly in terms of deaths between the two arms, like alluded to by others, that would negate the LRR difference especially given how small the LRR is in this very favorable group of patients.

 

[scarbrtj]

It’s interesting that they used LRFS with Kaplan Meier instead of LRR by Fine-Gray with death as a competing risk. Usually LRFS would imply death as an event. If there was a statistical anomaly in terms of deaths between the two arms, like alluded to by others, that would negate the LRR difference especially given how small the LRR is in this very favorable group of patients.

Really re: KM for LRFS? Educate me. Never heard that. I know the nomenclature is “KM survival” “instantaneous survival” “survival statistic” etc but that is affectational I thought. The KM statistic can analyze any event that occurs over time where you have to account for censored data. I’ve always(?) seen KM applied to LRFS (or DMFS etc). Getting more wonky, also was under impression that KM is non parametric and other things like Fine Gray less so. But HMU with some insight. (NB: Deaths before local recurrence get censored, so if death rates are equal between two groups in KM LRFS it’s a wash. However here if anything in TARGIT deaths were more common in the EBRT arm so they would have less chance to negatively skew LRFS, relative to IORT, by being “event contributors.”)

 

[Palex80]

I believe that in Kaplan Meier - progression free survival when the patient dies it's still an event, irrelevant of what the progression status of the disease is. Death is not censored.

 

[gmsquid]

Yeah there’s a difference in endpoint between locoregional recurrence free survival and freedom from locoregional recurrence. The latter censors death while the former counts it as an event. Think about the rtog 91-11 long term follow up. Laryngectomy free survival no different between induction and concurrent but laryngeal preservation better with recurrent. This was because there were more late deaths in the concurrent arm.

 

[gmsquid]

Also, in the original publication they reported lrr numbers not LRFS. Though they don’t clearly define if they include death as an event With LRFS, the change smells of statistical shadiness

 

[SneakyBooger]

Wouldn’t trust their data

Agreed.

Looking forward to the coming discussion this publication will generate in the coming weeks to months.

 

[Palex80]

We could claim 100% freedom from locoregional recurrence for many tumors. Just give 300 Gy to the site.
It only takes one patient that will survive the dose and you have proof.

 

[RollTideRadOnc]

Italian APBI protocol longterm results published in JCO today
Best APBI results by far.
This is clearly the go to regimen.
bye bye BID.
Sorry Dr. White.

 

[RollTideRadOnc]

ps. I’ve always thought the original Italian publication was criminally under discussed and the technique under utilized in the US

imho just shows our field’s “US” bias and how unscrupulous individuals will use bad techniques with ****ty cosmesis for career advancement while hardly acknowledging alternatives.

 

[scarbrtj]

ps. I’ve always thought the original Italian publication was criminally under discussed and the technique under utilized in the US

imho just shows our field’s “US” bias and how unscrupulous individuals will use bad techniques with ****ty cosmesis for career advancement while hardly acknowledging alternatives.

It's all lost in the fog of war now but the Italian data has been out a while, and these results now published in JCO were at either asco last year or this year. Very robust of course. Very believable. Has had "ring of truth" inside my mind for years. Since seeing it, I have been doing it a while. And then I mention 5 fraction once a day IMRT at ~6Gy a day to the academics and they're like "we are studying that." (UK IMPORT LOW also under-discussed.) It's as if sometimes there's Level 1 evidence... and then there's our Level 1 evidence. Other Level 1 evidence only counts when the powers that be say it counts. We should all strive to be as internally consistent as possible when choosing how we cherry-pick which phIII randomized data we love and which randomized data we say is fishy/suspect/not calculated correctly.

 

[Palex80]

It's all lost in the fog of war now but the Italian data has been out a while, and these results now published in JCO were at either asco last year or this year. Very robust of course. Very believable. Has had "ring of truth" inside my mind for years. Since seeing it, I have been doing it a while. And then I mention 5 fraction once a day IMRT at ~6Gy a day to the academics and they're like "we are studying that." (UK IMPORT LOW also under-discussed.) It's as if sometimes there's Level 1 evidence... and then there's our Level 1 evidence. Other Level 1 evidence only counts when the powers that be say it counts. We should all strive to be as internally consistent as possible when choosing how we cherry-pick which phIII randomized data we love and which randomized data we say is fishy/suspect/not calculated correctly.

They were shown at the San Antonio Breast Cancer Symposium, if I am not mistaken. I believe it was in year 1 B.C. (Before Corona) --> 2019.

10-Year Follow-up of Adjuvant APBI vs Adjuvant WBI - The ASCO Post

ascopost.com

 

[theradiator]

It's all lost in the fog of war now but the Italian data has been out a while, and these results now published in JCO were at either asco last year or this year. Very robust of course. Very believable. Has had "ring of truth" inside my mind for years. Since seeing it, I have been doing it a while. And then I mention 5 fraction once a day IMRT at ~6Gy a day to the academics and they're like "we are studying that." (UK IMPORT LOW also under-discussed.) It's as if sometimes there's Level 1 evidence... and then there's our Level 1 evidence. Other Level 1 evidence only counts when the powers that be say it counts. We should all strive to be as internally consistent as possible when choosing how we cherry-pick which phIII randomized data we love and which randomized data we say is fishy/suspect/not calculated correctly.

Am I allowed to declare TARGIT as fishy/suspect/not calculated correctly?

 

[scarbrtj]

Am I allowed to declare TARGIT as fishy/suspect/not calculated correctly?

I’ve Monday morning quarterbacked and driven a car from the backseat with the best of ‘em.

 

[Palex80]

Am I allowed to declare TARGIT as fishy/suspect/not calculated correctly?

Dr Vaidya has received a research grant from Photoelectron Corp (1996-1999) and from Carl Zeiss for supporting data management at the University of Dundee (2004-2008) and has subsequently received honoraria. Drs Vaidya, Tobias, Williams, Potyka, Ms Brew-Graves, and Mr Roberts receive funding from Health Technology Assessment Programme, National Institute for Health Research (NIHR), Department of Health for some activities related to the TARGIT trials. Dr Baum was on the scientific advisory board of Carl Zeiss and was paid monthly consultancy fees briefly before 2010. Dr Wenz has received a research grant from Carl Zeiss for supporting radiobiological research. Carl Zeiss sponsors some of the travel and accommodation for meetings of the international steering committee and data monitoring committee and when necessary for conferences where a presentation about targeted intraoperative radiotherapy is being made for all authors apart from Dr Eiermann who declares that he has no conflicts of interest. No other conflicts are reported.

 

[evilbooyaa]

I was gonna ask you to link to the "other" TARGIT; thanks, I hadn't seen it. I had heard of it (we've all heard of IORT), but this "new TARGIT" was of a different flavor, more compelling IMHO. Hard to know what to make of that JAMA TARGIT of N=~1100 and this BMJ TARGIT that is N=~2200 and the LRs are essentially the same between the arms.

Because nearly 25% of patients who got IORT in their n=2200 cohort also go EBRT. That was one of the main criticisms at initial publication, at how abysmal it was at actually getting patients to be able to avoid EBRT.

Soren Bentzen loves shredding the TARGIT authors:



I can also neither confirm nor deny that I am Soren Bentzen, for having the same hot take on SDN the day before he posted this on Twitter.

 

[scarbrtj]

Because nearly 25% of patients who got IORT in their n=2200 cohort also go EBRT. That was one of the main criticisms at initial publication, at how abysmal it was at actually getting patients to be able to avoid EBRT.

Soren Bentzen loves shredding the TARGIT authors:



I can also neither confirm nor deny that I am Soren Bentzen, for having the same hot take on SDN the day before he posted this on Twitter.

Maybe I’m just brain farting today but let’s assume IORT=holy water. Then this is a trial of EBRT vs radiotherapeutic nothing (nothing=IORT). One assumes that “high risk features” are balanced between the groups, so that effect will wash out (EBRT was “given” to high risk in the other arm too). So a logical follow-on of Soren’s argument is that EBRT has an effect of “exactly zero” in low risk patients? (Because LCs were similar between IORT and EBRT.) And that doesn’t seem right...
Also no one is addressing the mortality benefit of IORT. Whenever do we routinely say for a p=0.005 “I’m sure that’s a type 1 error.”

 

[Palex80]

Maybe I’m just brain farting today but let’s assume IORT=holy water. Then this is a trial of EBRT vs radiotherapeutic nothing (nothing=IORT). One assumes that “high risk features” are balanced between the groups, so that effect will wash out (EBRT was “given” to high risk in the other arm too). So a logical follow-on of Soren’s argument is that EBRT has an effect of “exactly zero” in low risk patients? (Because LCs were similar between IORT and EBRT.) And that doesn’t seem right...
Also no one is addressing the mortality benefit of IORT. Whenever do we routinely say for a p=0.005 “I’m sure that’s a type 1 error.”

The problem we have with trial interpretation are the broad inclusion criteria employed in those trials. This is a particular weakness that is striking.

My experience is that if you broaden the inclusion criteria to allow more "risky" patients (--> cN1, T2s, etc...) the investigators rarely enter such patients in the trial, because they know they may be put at risk with a deescalated treatment ("selection bias"). This is common for all de-escalation trials.
So these patients are the underrepresented in the trial but at the same time raise total recurrence rates anyway because there is enough of them.
But since noone is doing protocol-specified subgroup analyses (T1 vs. T2, pN0 vs. pN1) we cannot really claim that...

Net result are sh**ty data.
Patient recruitment was accelerated marginally because of the broader inclusion criteria but at the same time the diversity of the patients in the trial does not allow you to make a good statement on treatment efficacy in a well defined patient population.

Why couldn't the investigators design a trial with strict inclusion criteria, for instance like in this trial:
tumor size <3 cm, Grading 1 or 2, negative lymph nodes, positive estrogen and/or progesterone receptor status

 

[communitydoc13]

Maybe I’m just brain farting today but let’s assume IORT=holy water. Then this is a trial of EBRT vs radiotherapeutic nothing (nothing=IORT). One assumes that “high risk features” are balanced between the groups, so that effect will wash out (EBRT was “given” to high risk in the other arm too). So a logical follow-on of Soren’s argument is that EBRT has an effect of “exactly zero” in low risk patients? (Because LCs were similar between IORT and EBRT.) And that doesn’t seem right...
Also no one is addressing the mortality benefit of IORT. Whenever do we routinely say for a p=0.005 “I’m sure that’s a type 1 error.”

There was a significant difference in local failures between adaptive IORT and WBRT (24 vs. 60 at median 8.6 year f/u). Like gmsquid mentions above, they counted deaths as an event and as there was an imbalance favoring IORT regarding death from all causes, their endpoint ended up not being statistically significant.

The way the trial was designed makes it impossible for us to get what WBRT is doing in low risk patients as its a mixed population. I think we need to rely on other data for that. It's clearly doing something though, even in this trial.

Regarding how I would use this data to counsel patients in a real setting, I think that's complicated too. The 60 local failures in the IORT group is out of roughly 1000 patients but ~20+ percent also got whole breast. In addition, evaluable patients start disappearing around 6 years (mostly not because of death by the way), despite local failure events continuing. Could I really say to a patient "With IORT, your risk of local failure is going to be 6% at 10 years."? Of course not, and I would think the real risk to a patient with a long life expectancy would be markedly more than this.

 

[BobbyHeenan]

ps. I’ve always thought the original Italian publication was criminally under discussed and the technique under utilized in the US

imho just shows our field’s “US” bias and how unscrupulous individuals will use bad techniques with ****ty cosmesis for career advancement while hardly acknowledging alternatives.

I'll put in another plug for 30 in 5.

Been using it for a few years now. Used to do a lot of brachy APBI but started dabbling in 30/5 and I’m never going back.

Fantastic tolerance and cosmesis. Easy to plan and execute. Even for left sided patients that can’t do DIBH you have very low heart dose.

Would love to see 30/5 vs AI vs AI/30/5 for over age 70 T1s. I see all these ladies miserable on AI, getting prolia and what not when 5 sessions of the healing rays would do

 

[Palex80]

Would love to see 30/5 vs AI vs AI/30/5 for over age 70 T1s. I see all these ladies miserable on AI, getting prolia and what not when 5 sessions of the healing rays would do

Precisely!

The problem will be designing a trial like this.

From a purist point of view, you should do it as a non-inferiority trial but that would require thousands of patients to work.

But what if one thought about what "truly matters", as you very well outlined in your post, for these women. What about looking at quality of life?
Why not claim that all three treatment arms (like you described them) have the same recurrence rates more or less or that any difference between them is irrelevant, BUT you hypothesize that 30/5 is the arm with a superior quality of life, treatment satisfaction, no treatment regret and so on... Make it a superiority trial for that endpoint (could be a composite one) and look at the efficacy endpoints simply as secondary ones.
You would need far less patients for that. Needless to say, it is possible that it might even enhance accrual (I find it easier to explain and recruit patients in superiority trials than in non-inferiority ones).

We have learned that lesson before with AMAROS. Formally, it's a negative trial. AMAROS could not prove that lymphatics RT is not not-inferior compared to axillary dissection. Yet the difference is so small, that most people overlook that (primary!) endpoint and praise axillary RT because it produced such a better secondary endpoint: less arm lymphedema. Don't make the same mistake again! Design a trial with a revelant endpoint for the patients!

 

[dieABRdie]

I'll put in another plug for 30 in 5.

Been using it for a few years now. Used to do a lot of brachy APBI but started dabbling in 30/5 and I’m never going back.

Fantastic tolerance and cosmesis. Easy to plan and execute. Even for left sided patients that can’t do DIBH you have very low heart dose.

Would love to see 30/5 vs AI vs AI/30/5 for over age 70 T1s. I see all these ladies miserable on AI, getting prolia and what not when 5 sessions of the healing rays would do

Any issues getting IMRT approved for this?

 

[Burt Radnolds]

Any issues getting IMRT approved for this?

This has been discussed before, but yes, personally I have had issues with a select few payors. Usually able to get it through but they have been surprisingly resistant at times. Some just have a no imrt for partial breast or no imrt for right breast blanket policy that they will throw up like a shield. They don't care that the study exists.

The last one I had a problem with was going down the crapper until at the very end I said that due to covid I preferred to treat with 5fx IMRT, and that if they would not approve it I would inform the patient that I could not safely treat in this manner, and they would have to come in for 10-15 visits during the pandemic. Instantly approved.

 

[BobbyHeenan]

Any issues getting IMRT approved for this?

Rarely. I did have to fax the trial to the reviewer once but in last 6 months have had zero issues.

 

[Radonc90]

ASTRO just tweeted the TARGIT trial data: one and done.
This study, if the data holds up, will be great for patients but will be financially devastating for radonc...

 

[OTN]

ASTRO just tweeted the TARGIT trial data: one and done.
This study, if the data holds up, will be great for patients but will be financially devastating for radonc...

Not only will the data "hold up" in the future, it doesn't hold up now. Reposting my copying of a reviewer's statement, which describes why TARGIT was a pile o' crap

"Vaidya et al1 report 5-year outcomes in 1153 patients enrolled in the TARGIT-A trial of intraoperative radiotherapy (IORT) vs standard-of-care external beam radiotherapy (EBRT) for low-risk breast cancer in the postpathology setting. Their report provides a somewhat rare opportunity to revisit the debate in the literature at the time of the original publication of trial outcomes. The following points take stock of what we now know. First, not only is delayed second-procedure targeted IORT not noninferior to EBRT, but it is actually highly significantly inferior to EBRT with local recurrence rates at 5 years of 3.96% after IORT vs 1.05% after EBRT (2-tailed P = .002). Second, IORT may be “better than nothing,”2 but certainly not by much. The recurrence rate after IORT is 3.96% (95% CI, 2.7%-5.9%), very close to the 4.1% (95% CI, 2.4%-5.7%) local recurrence rate observed in those randomized to no radiotherapy in the PRIME trial3 after a median follow-up of 5 years. Interestingly, the rate after whole-breast radiotherapy in the control arm of PRIME was 1.3% (95% CI, 0.2%-2.3%), in very good agreement with the outcome in the EBRT arm of the TARGIT-A trial. Third, any—nonsignificant at the time—imbalances in noncancer-related mortality are long gone, and with it the suggestion that IORT should spare potentially fatal adverse effects of EBRT. To summarize, while hindsight is 20/20, the take-home message is that the warning flags were there4 already in the statistical analysis of the original publication from the TARGIT-A trial.5 Now we know."

 

[BobbyHeenan]

I haven't taken a deep dive into the IORT data....but can someone show me where in a randomized trial IORT is any better than just a lumpectomy alone? Are we sure it works?

I think that's what the reviewer/OTN is posting above.

 

[scarbrtj]

ASTRO just tweeted the TARGIT trial data: one and done.
This study, if the data holds up, will be great for patients but will be financially devastating for radonc...

Lol. ASTRO never retweets this in universes where APM doesn’t occur in Earth’s timeline.

 

[Radonc90]

So,

A researcher from Toronto took a shot at it lol...

 

[scarbrtj]

Single-Dose Radiation Therapy Can Counteract Serious Breast Cancer Treatment Compliance Issues During Pandemic

According to recent studies, an increase in the COVID-19 positivity rate correlates to an increase in mortality from non-coronavirus-related diseases including

www.newswise.com

 

[OTN]

“Unlike external beam radiation therapy (EBRT) which requires up to 30 visits to the radiotherapist, TARGIT-IORT achieves the same clinical outcomes with one dose of targeted radiation delivered from inside the breast during surgery immediately following the removal of the tumor,” explained Neil B. Friedman, M.D., FACS, Director of The Hoffberger Breast Center at Mercy.

SAME clinical outcomes, really?

 

[Palex80]

“Unlike external beam radiation therapy (EBRT) which requires up to 30 visits to the radiotherapist, TARGIT-IORT achieves the same clinical outcomes with one dose of targeted radiation delivered from inside the breast during surgery immediately following the removal of the tumor,” explained Neil B. Friedman, M.D., FACS, Director of The Hoffberger Breast Center at Mercy.

SAME clinical outcomes, really?

30 fractions...

At least he said "up to".

 

[Ray D. Ayshun]

Single-Dose Radiation Therapy Can Counteract Serious Breast Cancer Treatment Compliance Issues During Pandemic

According to recent studies, an increase in the COVID-19 positivity rate correlates to an increase in mortality from non-coronavirus-related diseases including

www.newswise.com

"researchers believe fear of contracting the virus compels patients to stay home instead of completing their post-lumpectomy radiotherapy regimens."

Didn't realize stage I breast cancer was such a quick killer. Maybe if breast cancer mortality is up in 2035.

 

[RadOncDoc21]

This goes back to one of my prior comments, rad oncs hurt each other in the worst ways just to gain on an independent level.

I also like the term “radiotherapist”