2018 ASTRO whole breast radiation guidelines

[scarbrtj]

It was a centralized path review, with a difference in how tumors were graded (not the way it was commonly done) that resolved the discrepancy.

Interesting. So a post-hoc analysis of their post-hoc analysis to change the subgroup compositions produced a different statistical outcome? Inconceivable!

"In the original study, an unplanned subgroup analysis suggested that the hypofractionated RT regimen was less effective in patients with high-grade tumors [4]. This observation was unexpected and appeared to contradict the experimental evidence that tumors with a high proliferative rate (high-grade) may be more sensitive to accelerated schedules of radiation therapy."

The devil's in the details. No one wants to hear my opinion but I'll give it anyway. The observation was not unexpected. There is no experimental evidence that high-grade tumors are more sensitive to accelerated BUT REDUCED DOSE schedules. If we assume a/b=10, back of envelope calc:
BEDGy10 42.5/16 = 42.5*(1+2.65/10) = 53.8
BEDGy10 50/25 = 50*(1+2/10) = 60
So we would assume high-grade to be more sensitive to 50/25 (time correction factors will prob modify this, granted). Just because it's accelerated doesn't overcome proliferation; you also have to factor in total delivered dose. A good reason they compared 45 Gy in 5 weeks vs 45 Gy in 3 weeks in the Turrisi trial e.g.

But anyway. It's "good" to know grade issues have been "debunked."

---

Members don't see this ad.

 

[evilbooyaa]

Interesting. So a post-hoc analysis of their post-hoc analysis to change the subgroup compositions produced a different statistical outcome? Inconceivable!

"In the original study, an unplanned subgroup analysis suggested that the hypofractionated RT regimen was less effective in patients with high-grade tumors [4]. This observation was unexpected and appeared to contradict the experimental evidence that tumors with a high proliferative rate (high-grade) may be more sensitive to accelerated schedules of radiation therapy."

The devil's in the details. No one wants to hear my opinion but I'll give it anyway. The observation was not unexpected. There is no experimental evidence that high-grade tumors are more sensitive to accelerated BUT REDUCED DOSE schedules. If we assume a/b=10, back of envelope calc:
BEDGy10 42.5/16 = 42.5*(1+2.65/10) = 53.8
BEDGy10 50/25 = 50*(1+2/10) = 60
So we would assume high-grade to be more sensitive to 50/25 (time correction factors will prob modify this, granted). Just because it's accelerated doesn't overcome proliferation; you also have to factor in total delivered dose. A good reason they compared 45 Gy in 5 weeks vs 45 Gy in 3 weeks in the Turrisi trial e.g.

But anyway. It's "good" to know grade issues have been "debunked."

I'm not going to argue with that as I agree with you in regards to the differences in BED between hypofrac and conventional regimens - however, this is the part that I meant for you to focus on:
"The grading system applied to the tumors in the trial was the Scharff Bloom Richardson (SBR) grading system [6]. This system has subsequently been modified and replaced by the more quantitative and reproducible Nottingham grading system (NGS) which has been demonstrated to be prognostic for LR following BCS and whole-breast RT"

"The grading system employed for the original analysis was the SBR system [6]. This system is no longer employed in the grading of breast cancers and has been replaced by the more reproducible and objective NGS which has repeatedly been demonstrated to be prognostic for LR, OS and DFS [8–10]. The NGS applied by the study pathologist (AB), an expert breast pathologist, correlated modestly with the local hospital SBR grade (κ = 0.48) and was prognostic for outcome; however, it was not shown to be predictive of a response to hypofractionation. The inability of the NGS (or the SBR system) to predict response to hypofractionated RT in this current analysis may be the result of the smaller sample size, or more likely, that the statistically significant result observed in the original study was a chance finding and not a real effect."

I don't buy the accelerated nonsense being better either with such a miniscule elevation from 2Gy to 2.6Gy/day.

 

[Gfunk6]

I haven't had a chance to slog through this thread, but I would make several brief observations:

1. This position paper is a good thing for our patients and a good thing for our field. It is no wonder that the hypofractionation studies that were the backbone of this consensus paper were all from countries with socialized medicine where there is not a *****ic incentive to over-treat patients when it is not warranted. Breast hypofractionation = less acute toxicity, less chronic toxicity, more convenient for patients, less out of pocket costs for patients; all win-win

2. The old studies from Sweden where patients were treated using cobalt with extreme hypofractionation DID result in substantial > 50% rates of brachial plexopathy decades later. This is important to remember and consider, but IMO it is apples and oranges with modern hypofractionation. First off, we use photons which are much more forgiving for superficial doses than cobalt. Second we use 3D planning so we know exactly what dose the brachial plexus gests. Third, we have highly robust radiobiological models which let us intelligently estimate risk to brachial plexus.

3. Hypofractionation is the logical consequence of years of research by many Radiation Oncology investigators, physicists, and accelerator manufacturers. It is the future of our field. With improved beam shaping, image guidance, dose delivery speed, etc. we will surely advance to extreme hypofractionation in virtually all treatment sites.

4. I am not blind to the fact that this leads to less radiation visits and therefore less reimbursement in the US. However, we are obligated by the Hippocratic Oath to do what is right and do what would result in the maximum benefit/least harm. To avoid practices from going under financially we must embrace payment reform and captiated payments and get away from this ridiculous dose/fraction arrangement. ASTRO, to its credit is trying to craft an Advanced/Alternative Payment Model (APM) with CMS; in addition they have gotten Congress to freeze Radiation Oncology reimbursement rates until the end of 2019 while this is being worked on.

Residency expansion is a valid issue but has nothing to do with the above position paper. I am sympathetic to this problem but regurgitating this topic in every thread is counter-productive.

 

[scarbrtj]

The grading system employed for the original analysis was the SBR system

Yeah I read all that. But from an honest-dealing statistics standpoint, this was not a convincing way to "debunk" their original finding. It in no way contravenes the original finding. I don't fault them for publishing the "new finding" but I'm not going to honor-bound to listen to their results chronologically; a result is a result is a result. The lack of a grade issue in START contravenes the grade issue more than Whelan et al's post-hoc analysis of their post-hoc analysis and subgroup rearrangement (no matter the compelling reason for doing so). In a perfect world, the standard treatment would not be recommended to be wholesale replaced had there been some concerning finding regarding its replacement until a trial to address the concerning finding was done to see if the concerning signal was "live or Memorex." Now with the guidelines being published we can't put that cat back in the bag. C'est la vie. Who wants to do a hypofx trial on Gr3 patients? You'd be laughed outta the room. Although I'd have liked to have seen that trial.

 

[deleted18755]

I haven't had a chance to slog through this thread, but I would make several brief observations:

1. This position paper is a good thing for our patients and a good thing for our field. It is no wonder that the hypofractionation studies that were the backbone of this consensus paper were all from countries with socialized medicine where there is not a *****ic incentive to over-treat patients when it is not warranted. Breast hypofractionation = less acute toxicity, less chronic toxicity, more convenient for patients, less out of pocket costs for patients; all win-win

2. The old studies from Sweden where patients were treated using cobalt with extreme hypofractionation DID result in substantial > 50% rates of brachial plexopathy decades later. This is important to remember and consider, but IMO it is apples and oranges with modern hypofractionation. First off, we use photons which are much more forgiving for superficial doses than cobalt. Second we use 3D planning so we know exactly what dose the brachial plexus gests. Third, we have highly robust radiobiological models which let us intelligently estimate risk to brachial plexus.

3. Hypofractionation is the logical consequence of years of research by many Radiation Oncology investigators, physicists, and accelerator manufacturers. It is the future of our field. With improved beam shaping, image guidance, dose delivery speed, etc. we will surely advance to extreme hypofractionation in virtually all treatment sites.

4. I am not blind to the fact that this leads to less radiation visits and therefore less reimbursement in the US. However, we are obligated by the Hippocratic Oath to do what is right and do what would result in the maximum benefit/least harm. To avoid practices from going under financially we must embrace payment reform and captiated payments and get away from this ridiculous dose/fraction arrangement. ASTRO, to its credit is trying to craft an Advanced/Alternative Payment Model (APM) with CMS; in addition they have gotten Congress to freeze Radiation Oncology reimbursement rates until the end of 2019 while this is being worked on.

Residency expansion is a valid issue but has nothing to do with the above position paper. I am sympathetic to this problem but regurgitating this topic in every thread is counter-productive.

I agree completely with GFunk's summary above. In no way should residency expansion or the job market influence this position paper or even be mentioned in it but I certainly think that if ASTRO can commission such a task force for something like this then they could easily do so to address the residency situation if they really wanted to (quoting their own data/papers demonstrating a genuine need for less treatments per patients and therefore less doctors in the future!)

 

[Palex80]

Would someone mind offering a clarifying bit of info. Last time I saw Whelan et al, the grade 3 patients had a LR of 15.6% with hypofx as compared to 4.7% with standard fx (p=0.01). While this was post-hoc analysis, it was still concerning to me, and radiobiologically would make sense as a higher grade tumor theoretically would have higher alpha/beta and that would be a tumor for which hypofractionation coupled with attendant total dose-lowering would be infaust.

And yet...

The guidelines say: "The decision to offer HF-WBI should be independent of tumor grade. Recommendation strength: Strong. Quality of evidence: High. Consensus: 100%."

Why? Wouldn't there be a tiny pause for concern here for grade? Seems a little absolutist, scientifically.

No, no, no.

This "G3-concern" has been proven wrong.
Centralized pahology review showed no difference
https://pdfs.semanticscholar.org/9f08/15662b4929ce53abbf522c937dc86ff63cae.pdf

Edit: someone else pointed that out already, Thank you

 

[xrt123]

Do you have a link? That sounds pretty cool.

Tumor Boards and the Quality of Cancer Care | JNCI: Journal of the National Cancer Institute | Oxford Academic

The irony is now I got to about 6 tumor boards a week. Maybe someone should put out a Guideline on the lack of need to have tumor boards.
Level of Evidence: Strong
Consensus: 100% of my opinion.

 

[scarbrtj]

Centralized pahology review showed no difference

In actuality, changing the path classification so that prev "Gr3" were not "Gr3" anymore (ie changing the composition of the subgroups into a new post hoc analysis vs the previous post hoc analysis) showed that there was no difference in the new post hoc grouping. Semantics, maybe; good from a biostats approach, prob not. But the previous (as described in NEJM 2010) difference re: grade still stands. I don't know why I find this 2014 re-analysis/re-classification so weaselly and no one else does, but I don't wanna be quixotic. The lack of finding in START is more compelling.

 

[Burt Radnolds]

I know of no one in America who's using START fractionation schemes; do you?

I exclusively use 4005/15 (START B) and know many who practice likewise. It is the more robust trial and also demonstrated improved cosmesis compared to Whelan's equivalency. It's basically Canadian minus 1 which in my mind is the sweet spot.

Another thing that has not really been discussed is that START allowed and stratified by boost (1000/5). If there was in fact something behind that later debunked unplanned Whelan analysis, perhaps the appropriate use of boost in START washed it out and made it dissapear.

 

[Palex80]

I exclusively use 4005/15 (START B) and know many who practice likewise. It is the more robust trial and also demonstrated improved cosmesis compared to Whelan's equivalency. It's basically Canadian minus 1 which in my mind is the sweet spot.

Another thing that has not really been discussed is that START allowed and stratified by boost (1000/5). If there was in fact something behind that later debunked unplanned Whelan analysis, perhaps the appropriate use of boost in START washed it out and made it dissapear.

Very good poimts.

In Europe START B is also the most commonly used schedule.

 

[scarbrtj]

I exclusively use 4005/15 (START B) and know many who practice likewise. It is the more robust trial and also demonstrated improved cosmesis compared to Whelan's equivalency. It's basically Canadian minus 1 which in my mind is the sweet spot.

Ha, yes, true. Fractionation Jenga. It's that one. Extra. Fractional. False. Move. That kills the good cosmesis.

 

[Burt Radnolds]

Ha, yes, true. Fractionation Jenga. It's that one. Extra. Fractional. False. Move. That kills the good cosmesis.

Not sure if serious.... You appear to be a disciple of radiation biology, as am I. I would think you would embrace these compelling results from extremely robust trials with open arms. It is very possible that 1 fraction does indeed cross the threshold of impacting cosmesis negatively. It is also possible that the linear quadratic equation does not fully explain what we experience in vivo. While you are willing to engage in hard core rad bio pretzel logic, you appear to make light of the fact that one extra fraction may matter. Apologies if I misunderstood your intent, I do appreciate your willingness to engage in high level scientific exploration of these ideas.

 

[scarbrtj]

Not sure if serious.... You appear to be a disciple of radiation biology, as am I. I would think you would embrace these compelling results from extremely robust trials with open arms. It is very possible that 1 fraction does indeed cross the threshold of impacting cosmesis negatively. It is also possible that the linear quadratic equation does not fully explain what we experience in vivo. While you are willing to engage in hard core rad bio pretzel logic, you appear to make light of the fact that one extra fraction may matter. Apologies if I misunderstood your intent, I do appreciate your willingness to engage in high level scientific exploration of these ideas.

I am neither prepared not to fully accept the results nor disallowing myself from appreciating the irony that perhaps in one trial (no matter how well done) we found "the sweet spot." How propitious. I'm making light of it and simultaneously not throwing shade on it. Some say one of the marks of intelligence is holding two conflicting thoughts in one's mind at the same time...

 

[medgator]

How do you guys approach active smokers in terms of post-lumpectomy XRT? Any data re: hypofractionation in the active PPD smoker who is 50-60 years old?

 

[seper]

I personally think lung toxicity of standard tangential WBRT is too low to justify any plan adjustments based on smoking history.

How do you guys approach active smokers in terms of post-lumpectomy XRT? Any data re: hypofractionation in the active PPD smoker who is 50-60 years old?

 

[medgator]

I personally think lung toxicity of standard tangential WBRT is too low to justify any plan adjustments based on smoking history.

I was actually concerned more about soft tissue and skin toxicity. Agree lung isn't a big deal, although I read somewhere that there can be a higher risk of secondary lung malignancy in that population of patients.

 

[evilbooyaa]

I was actually concerned more about soft tissue and skin toxicity. Agree lung isn't a big deal, although I read somewhere that there can be a higher risk of secondary lung malignancy in that population of patients.

It is a higher risk of developing a lung cancer (than nonsmokers receiving whole breast RT), but how much of that is from their generalized smoking history and how much is from the radiation?

 

[MDACCRules]

I was actually concerned more about soft tissue and skin toxicity. Agree lung isn't a big deal, although I read somewhere that there can be a higher risk of secondary lung malignancy in that population of patients.

And what in radiobiology tells you that acute toxicity is determined by fraction size rather than total dose? And what in radiobiology tells you that there is a higher risk of secondary malignancy in patients treated with higher fractional doses but lower total doses? And why this means you should use different fractionation in smokers? Goodness gracious.

This is just faulty logic, and continuing to try to find subsets where you can give 33 fractions.

 

[nkmiami]

Partial breast 15 x 266 is always a consideration....

 

[2014XRT]

Partial breast 15 x 266 is always a consideration....

yeah IMPORT LOW it

 

[evilbooyaa]

And what in radiobiology tells you that acute toxicity is determined by fraction size rather than total dose? And what in radiobiology tells you that there is a higher risk of secondary malignancy in patients treated with higher fractional doses but lower total doses? And why this means you should use different fractionation in smokers? Goodness gracious.

This is just faulty logic, and continuing to try to find subsets where you can give 33 fractions.

1) He didn't say acute toxicity.
2) You've been warned for being directly harassing at this point, in a thread that I specifically posted a warning regarding to keep it civil. Continued verbal disparaging against other members of SDN in this forum will lead to removal from the site.

On a side note - I'm done having this forum be a lot of directly confrontational stuff due to the anonymity of it all.

 

[medgator]

1) He didn't say acute toxicity.

Which is why I made a point of stating her age, but I digress

You've been warned for being directly harassing at this point, in a thread that I specifically posted a warning regarding to keep it civil. Continued verbal disparaging against other members of SDN in this forum will lead to removal from the site.

On a side note - I'm done having this forum be a lot of directly confrontational stuff due to the anonymity of it all.

Amen.

 

[medgator]

http://ascopubs.org/doi/full/10.1200/JCO.18.00317

All patients treated with chemotherapy in this study received an anthracycline- or taxane-based regimen, and among these patients, the proportion with an adverse cosmetic outcome was 4.1% higher with HF-WBI than with CF-WBI, although it was below the 10% noninferiority margin, with a 90% UCL of 15.0%. This finding, likely the result of having a moderate sample size, shows that the trial was inconclusive regarding noninferiority in this strata-specific subgroup. Conversely, our finding that adverse cosmetic outcome was 18.6% lower with HF-WBI than with CF-WBI among patients with D cup or larger breasts offers strong reassurance that HF-WBI should not compromise cosmetic outcome among large-breasted patients. Such data may facilitate shared decision making regarding the choice to proceed with HF-WBI for patients with large breasts. Additional study of current-era chemotherapy combined with HF-WBI could be beneficial to decrease the uncertainty surrounding cosmetic outcome.

Nice to finally see some prospective/randomized data regarding the "modern" chemo + Hfx question using modern 3D planning techniques....

Everything else looked good... dcis, large breast size, use of a boost etc.

 

[evilbooyaa]

Mehhhh their non-inferiority margin was 10% worse, and this 4.1, with 90% CI went to 15% which is why it's not non-inferior. Given where the ASTRO WBI guidelines are currently, I don't think this is going to be enough to change practice - those who are comfortable with HF with chemo will continue to use it (as per ASTRO guideliens) and those who continue to poo-poo HF have a data point in chemo patients.

 

[Palex80]

Of course it's not enough to change practice...
The trial had 287 patients, out of which merely 85 received chemotherapy.
You cannot expect to draw conclusions from a subgroup of 85 patients...

And the graphs speak volumes...