SABR-COMET published in Lancet

[PointA]

The manuscript is now available in The Lancet: phase 2 trial with 99 patients showing OS benefit with SABR. There will be much hoopla for sure but we will have to wait and see what the phase 3 trial shows. It's an all too common and unfortunate story in oncology when a small phase 2 trial shows a large OS benefit that does not translate to the confirmatory phase 3, for example. Nonetheless, exciting for the field no doubt.

---

Members don't see this ad.

 

[Phantom1]

This is a great study. Why do you automatically say well we need to wait and see for a Phase 3? This was a randomized study. There is also a similar phase 2 randomized study which was stopped early because it showed a benefit to SBRT in oligometastatic lung cancer, which is a highly agressive disease.

My interpretation would be to treat oligometastic patients now with SBRT who fit either study criterion, and if the Phase 3 does not show a benefit that would be a surprise. Also what if there is a progression free survival benefit in the phase 3 but no OS benefit. Is that enough reason to treat? What if the Phase 3 shows a benefit only in patients with 1-3 mets but not >3? Patients with 1-3 mets in the interim would miss out from the benefit of SBRT if one waits to use SBRT in oligometastatic disease until the Phase 3 reports out. Other issue to weigh is that SBRT is really well tolerated with minimal side effects. Up for discussion...

FYI I was not involved in either study, but these are fantastic for the field. As systemic therapies get better, local therapy becomes more usefull (for example XRT to prostate in oligometastatic disease) similarly SBRT to sites of metastatic disease has now shown a benefit in two randomized Phase 2 studies..

 

[Palex80]

It is striking how everyone is cheering about SABR-COMET. Yet what did the trial do? Perform RT on oligometastatic disease, something we have been doing for decades (?) now... Yet, all the radiation oncology community was able to put together was this phase II trial with roughly 100 patients. Is it not sad?

 

[deleted18755]

It is striking how everyone is cheering about SABR-COMET. Yet what did the trial do? Perform RT on oligometastatic disease, something we have been doing for decades (?) now... Yet, all the radiation oncology community was able to put together was this phase II trial with roughly 100 patients. Is it not sad?

I would have called you an “SDN doom and gloomer” but you are correct - treating limited disease with SBRT makes sense and is what so many of us have been doing for years butthe best evidence to support it is a study with 100 patients of mixed histology?

That’s the criticism and push back I’m getting after enthusiastically forwarding this study to my non Rad Onc colleagues ... is it just me?

 

[Haybrant]

It is striking how everyone is cheering about SABR-COMET. Yet what did the trial do? Perform RT on oligometastatic disease, something we have been doing for decades (?) now... Yet, all the radiation oncology community was able to put together was this phase II trial with roughly 100 patients. Is it not sad?

Could someone w a lot of clinical trial design and study background tell us their thoughts about this. It was my first instinct as well but I know accruing onto something like this may not be easy either. But give how many met patients there are 100 sounded very paltry. Then again you’re just looking for a signal and this definitely has it - the phase 3 I assume would be a ton more patients

 

[xrthopeful]

We needed evidence to do it in a non-haphazard fashion, which is how it’s been done by people wanting to be aggressive for oligomets historically.

This is the first step. And It’s an exciting one

More trials are needed.

it costs a lot of money to run big trials and radiation trials dont have big pharma money/backing to pay for them.

Palma deserves a ton of credit for how productive he has been able to be with all of his trials, including ORATOR, MISSILE etc.

This trial would have been way harder to do in the US, and was hard enough in Canada. You think many med oncs or rad oncs want to RANDOMIZE a Patient they believe has oligomets to whether by or not they get sbrt? Like palex said - many are doing it anyways so why would they want to randomize some to no treatment?

This is the same reason LU002 is going to Be difficult to accrue

 

[Palex80]

This trial would have been way harder to do in the US, and was hard enough in Canada. You think many med oncs or rad oncs want to RANDOMIZE a Patient they believe has oligomets to whether by or not they get sbrt? Like palex said - many are doing it anyways so why would they want to randomize some to no treatment?

This is the same reason LU002 is going to Be difficult to accrue

Well on the other side, look at STAMPEDE Arm H. Stampede Arm H was asking a different question (treating the primary and not the mets), yet the theme was the same: "will local therapy lead to OS-benefit in metastatic disease?"

Did anyone think that treating the prostate in metastatic prostate cancer patient leads to an OS benefit? Perhaps some of us did. I personally had only done it in handful of patients before STAMPEDE Arm H came out, in >10 years of clinical practice. As far as I can recall the patients I treated only had a solitary metastatic lesion.

Yet the STAMPEDE investigators managed to recruit over 2000 patients in that arm comparing SOC vs. SOC + RT to the prostate. And many of those patients had quite advanced metastatic disease with multiple lesions. At the end they showed that low-burden patients benefit from local therapy to the prostate and the magnitude of that OS-benefit was at the same level of Docetaxel or Abiraterone (hazard-ratio-wise).
RT to the prostate is worth as much as chemotherapy or abiraterone in low burden metastatic prostate cancer!

Was a randomized phase II trial conducted before STAMPEDE Arm H looking at the same question? No.

 

[xrthopeful]

If your point is for us to be impressed by the ability of the UK to consistently carry out large scale trials that are difficult to do in North America (especially the US) - then it's too late, I already am impressed.

to be fair, it is fairly easy to enroll there with the NHS system - essentially any patient coming through with prostate cancer in the STAMPEDE era is enrolled on to whatever arm is open/relevant.

same thing for CHIIP or START years before. If you're eligible, you're on.

plus there is the part where COMET includes any histology so it's not just one gatekeeper, there are dozens.

I'm not sure exactly what your point is, but a larger phase III is open - COMET-3

 

[Radiated_GuineaPig]

The data at least in the med onc literature is littered with phase II studies that were incredibly promising and turned out to be negative in the phase III. Therefore it’ll be so important that we as a field enroll on these larger trials to show benefit especially with the looming bundled payment system that everyone is afraid of. However it’s a double edged sword because as these phase IIs come out it’ll be harder to get our rad onc colleagues to randomize. Unlike pharma studies in which pharma controls drug access we control the keys to the linac, and turning the key happens to give us a good chunk of change.

Anyways hopefully we do better as field than we did with protons, where everyone says we cannot now randomize, leaving fanatics and non believers with no way of reconciliation. All the while insurance keeps tightening the restrictions on who gets reimbursed.

 

[MegaVoltagePhoton]

Wow...people hating on SABR-COMET because you "already know" this would prolong survival or you "do this already anyway" are seriously missing the point. And potentially are part of the problem.

Palma himself has said that med oncs refer patients for SBRT for oligomet disease. When he offers randomization to his trial, both the patient and the med onc get pissed because they too "already know" SBRT would work. And then same med oncs will later tout in academic conferences that rad oncs dont have good trials or big phase III trials to show the benefit of their interventions, while med oncs do

Ask yourself -- how would your med onc referral base or the patients react if you offered a 50/50 randomization to SBRT for oligometastatic disease?

The trial was written as a phase II designed to find if a phase III would be worth it. That is, anticipating these problems, Palma wrote it to be a small trial.

I do commend STAMPEDE, but ARM H wasn't nearly the same thing. Those were patients facing something much more murky, with (ostensibly) more of a trade off (while SBRT was pretty toxic in the SABR-COMET trial, one could argue that's surprising...no one would find local side effects from prostate directed RT to be a surprise). so the STAMPEDE Arm H patients had a choice of a much less supported/commonly done intervention with definitive risk of side effects, which no doubt aided the randomization. Plus the UK is a monster with these things.

Even then, there isn't an oligomet protocol within the STAMPEDE schema currently, showing that not even they were trying this randomization for oligomet directed RT or not yet.

 

[scarbrtj]

My interpretation would be to treat oligometastic patients now with SBRT who fit either study criterion, and if the Phase 3 does not show a benefit that would be a surprise. Also what if there is a progression free survival benefit in the phase 3 but no OS benefit. Is that enough reason to treat?

DISCLAIMER:
This is a cool trial. I like this trial. God bless this trial.

KEEPING IT STATISTICALLY REAL:
The p-value for this trial for median OS differences of ~28 mos for standard vs ~41 mos for SABR was 0.09; either trial's too underpowered or the effect is spurious.
It will not be a "surprise" if a replicating PhIII study shows no benefit. It will actually be expected.
For effects in the p-value ~0.05 range, they reproduce about 50% of the time, so p-values above 0.05 will reproduce less commonly than 50/50.
In Vegas, you would get rich betting on non-reproducibilty of trials with p-values of 0.05 or higher in other words.
This reproducibility crisis has been hacked over ad nauseum in the press and won't redo that here.

That said, carry on and SABR.
(I would not give local therapies with any even mild risks of side effects in metastatic disease unless survival was improved. Patient selection of those Stage IV patients with very low risk of local side effects from SABR is an art and doable.)

 

[medgator]

Palma himself has said that med oncs refer patients for SBRT for oligomet disease. When he offers randomization to his trial, both the patient and the med onc get pissed because they too "already know" SBRT would work. And then same med oncs will later tout in academic conferences that rad oncs dont have good trials or big phase III trials to show the benefit of their interventions, while med oncs do

Ask yourself -- how would your med onc referral base or the patients react if you offered a 50/50 randomization to SBRT for oligometastatic disease?

Outside of brain mets, MOs in my neck of the woods aren't thinking with an oligometastatic paradigm (outside of maybe liver mets with CRC). So not necessarily true everywhere... Their knee jerk is to give chemo alone

 

[Radiated_GuineaPig]

Wow...people hating on SABR-COMET because you "already know" this would prolong survival or you "do this already anyway" are seriously missing the point. And potentially are part of the problem.

Palma himself has said that med oncs refer patients for SBRT for oligomet disease. When he offers randomization to his trial, both the patient and the med onc get pissed because they too "already know" SBRT would work. And then same med oncs will later tout in academic conferences that rad oncs dont have good trials or big phase III trials to show the benefit of their interventions, while med oncs do

Ask yourself -- how would your med onc referral base or the patients react if you offered a 50/50 randomization to SBRT for oligometastatic disease?

The trial was written as a phase II designed to find if a phase III would be worth it. That is, anticipating these problems, Palma wrote it to be a small trial.

I do commend STAMPEDE, but ARM H wasn't nearly the same thing. Those were patients facing something much more murky, with (ostensibly) more of a trade off (while SBRT was pretty toxic in the SABR-COMET trial, one could argue that's surprising...no one would find local side effects from prostate directed RT to be a surprise). so the STAMPEDE Arm H patients had a choice of a much less supported/commonly done intervention with definitive risk of side effects, which no doubt aided the randomization. Plus the UK is a monster with these things.

Even then, there isn't an oligomet protocol within the STAMPEDE schema currently, showing that not even they were trying this randomization for oligomet directed RT or not yet.

FYI the STAMPEDE group has circulated schemas of their next study which will randomize the mets in oligometastatic prostate cancer to definitive treatment or not. This is coming.

Speaking of STAMPEDE, I find it interesting that SWOG is starting S1802 where patients, including those with oligometastatic disease will be randomized to local treatment or no local treatment to the tune of ~1,500 patients. This study of course allows surgery and is the urologists play at keeping relevant in the metastatic arena. I think the question is fair albeit unexciting for polymetastatic patients, but i'm curious what people think of enrolling on this study for their oligometastatic prostate patient.

 

[MegaVoltagePhoton]

DISCLAIMER:
This is a cool trial. I like this trial. God bless this trial.

KEEPING IT STATISTICALLY REAL:
The p-value for this trial for median OS differences of ~28 mos for standard vs ~41 mos for SABR was 0.09; either trial's too underpowered or the effect is spurious.
It will not be a "surprise" if a replicating PhIII study shows no benefit. It will actually be expected.
For effects in the p-value ~0.05 range, they reproduce about 50% of the time, so p-values above 0.05 will reproduce less commonly than 50/50.
In Vegas, you would get rich betting on non-reproducibilty of trials with p-values of 0.05 or higher in other words.
This reproducibility crisis has been hacked over ad nauseum in the press and won't redo that here.

That said, carry on and SABR.
(I would not give local therapies with any even mild risks of side effects in metastatic disease unless survival was improved. Patient selection of those Stage IV patients with very low risk of local side effects from SABR is an art and doable.)

Mmmm....well, read the methods section, which explains rather clearly the power calculations and the design of the trial as a "signal finding" trial, rather than a definitive trial. Your interpretation of it being underpowered or spurious due to the p-value is incorrect.

 

[MegaVoltagePhoton]

FYI the STAMPEDE group has circulated schemas of their next study which will randomize the mets in oligometastatic prostate cancer to definitive treatment or not. This is coming.

Speaking of STAMPEDE, I find it interesting that SWOG is starting S1802 where patients, including those with oligometastatic disease will be randomized to local treatment or no local treatment to the tune of ~1,500 patients. This study of course allows surgery and is the urologists play at keeping relevant in the metastatic arena. I think the question is fair albeit unexciting for polymetastatic patients, but i'm curious what people think of enrolling on this study for their oligometastatic prostate patient.

Yes, and good on them. Just pointing out it's not trivial.

 

[MegaVoltagePhoton]

Outside of brain mets, MOs in my neck of the woods aren't thinking with an oligometastatic paradigm (outside of maybe liver mets with CRC). So not necessarily true everywhere... Their knee jerk is to give chemo alone

Fair enough then. But the poster above had clearly said he's doing it anyway (so either med oncs are pissed, or they buy in for that particular poster).

 

[Palex80]

Wow...people hating on SABR-COMET because you "already know" this would prolong survival or you "do this already anyway" are seriously missing the point. And potentially are part of the problem.

Palma himself has said that med oncs refer patients for SBRT for oligomet disease. When he offers randomization to his trial, both the patient and the med onc get pissed because they too "already know" SBRT would work. And then same med oncs will later tout in academic conferences that rad oncs dont have good trials or big phase III trials to show the benefit of their interventions, while med oncs do

Ask yourself -- how would your med onc referral base or the patients react if you offered a 50/50 randomization to SBRT for oligometastatic disease?

The trial was written as a phase II designed to find if a phase III would be worth it. That is, anticipating these problems, Palma wrote it to be a small trial.

I do commend STAMPEDE, but ARM H wasn't nearly the same thing. Those were patients facing something much more murky, with (ostensibly) more of a trade off (while SBRT was pretty toxic in the SABR-COMET trial, one could argue that's surprising...no one would find local side effects from prostate directed RT to be a surprise). so the STAMPEDE Arm H patients had a choice of a much less supported/commonly done intervention with definitive risk of side effects, which no doubt aided the randomization. Plus the UK is a monster with these things.

Even then, there isn't an oligomet protocol within the STAMPEDE schema currently, showing that not even they were trying this randomization for oligomet directed RT or not yet.

1. Define "oligometastatic disease". I think many of us get referrals for patients with 1 or 2 metastatic lesions from med oncs and surely these patients are considered oligometastatic by most of us, but look at COMET: up to 5 lesions and you are still oligometastatic?
I don't know, but I do not consider a patient with 3 liver lesions and 2 bone lesions with an NSCLC "oligometastatic". Something tells me that it's just the tip of the iceberg...
Therefore perhaps for a number of patients that would fit into COMET standard of care would have been in many places to deliver SBRT, but I do think that a sizable number of patients would be treated systemically only.

2. I don't know if SBRT is that toxic. Surely, a few people died in COMET, but I've said it before and will say it again: I still wonder why it happened. We shouldn't be killing patients SBRT and I see no reason to put them at harms way for a palliative treatment. Constraints should have been tighter to ensure safery. It's different for Stage I NSCLC obviously, I would risk toxicity there since treatment is curative.

On the other hand, I am not sure if it was more easy to "sell" prostate RT to patients in Stampede than it was to "sell" them SBRT in Comet.
1. Patients often are scared of mets more than they are of the primary tumors. So if I have a controlled primary (inclusion criteria in COMET) and 1-5 uncontrolled mets, I'd love to go on a trial to get those treated. Not the same in Stampede. I have 20 untreated bone lesions, why should I care about my primary tumor in the prostate, it's not the one giving me all the back pain...
2. Treating with SBRT in Comet is quicker than 4 weeks of EBRT as in Stampede.

 

[scarbrtj]

Mmmm....well, read the methods section, which explains rather clearly the power calculations and the design of the trial as a "signal finding" trial, rather than a definitive trial. Your interpretation of it being underpowered or spurious due to the p-value is incorrect.

"Signal finding" trial. I know, I know. Then laud it for that. When Ellie Arroway found a signal she didn't immediately claim extraterrestrial life exists. But here we are saying the aliens have arrived. (Although I just laughingly thought the abscopal effect might be our specialty's Fermi Paradox.) What's up with a "signal-finding" trial by the way. I'm not sure we want medical literature to turn into a plethora of positive trials, finding signals, setting p-values at 0.2 (you can find even more "signals" at a pre-specified p-value of 0.3!), etc. Palma is a smart guy, but a statistical purist he's not. Again, with a p-value of 0.09, this study is technically actually MORE helpful to say that SABR WON'T affect survival (in additional replicating trials), binary p< or >0.05 odds-wise, regardless of Palma's sui generis verbiage and non-standard pre-specified p-value. But I'm just some guy on the internet, ignore me if you want.

It's easy to think about. I collected some data from two groups looking at a single metric. I compared the metric of the two groups; they were different (P=0.09). Now I go and get completely new data from a new sample from the population. What are the odds I will find a difference that is significant at p<0.05 or less. Should be obvious it's less than 50%. What are the odds they will be different at p<0.005? Very, very low.

Not finding a very suggestive difference between groups is usually a "signal" that there is not a very suggestive difference. Hope not. But again, I won't be "surprised" any more than when I lose money at the blackjack table. Stay Bayesian.

EDIT:
Also Palma just had another PhII study out: Measuring the Integration of Stereotactic Ablative Radiotherapy Plus Surgery for Early-Stage Non-Small Cell Lung Cancer.
Because SABR usually has about a 90% LC rate, he "pre-specified" a 90% pCR rate at 10 weeks, calc'd the power necessary etc to find that.
The trial found the pCR was only 60%. Some people called that "eye-popping" level bad.
It wasn't bad. Palma just made a bad guess, and the trial overall was poorly conceived. Just like his guess that a p-value of 0.2 or less was OK was a bad guess. I'm not criticizing these results. I'm criticizing the guesses.

 

[deleted18755]

Outside of brain mets, MOs in my neck of the woods aren't thinking with an oligometastatic paradigm (outside of maybe liver mets with CRC). So not necessarily true everywhere... Their knee jerk is to give chemo alone

I think this is the issue that is causing a lot of differing opinions in this thread: there is incredible variability with regards to how our referring physicians regard the concept of "oligometastatic disease" and when SBRT is indicated (vs just pumping more and more, newer and newer, systemic agents).

I can imagine there are radiation oncologists who work more collaboratively with medical oncologists who refer appropriately who can't imagine that there are still many medical oncologists who will keep pumping systemic therapies in a patient with excellent performance status and a single liver met vs referring for SBRT. I work in a region with both extremes and it is mind blowing sometimes...

 

[evilbooyaa]

1. Define "oligometastatic disease". I think many of us get referrals for patients with 1 or 2 metastatic lesions from med oncs and surely these patients are considered oligometastatic by most of us, but look at COMET: up to 5 lesions and you are still oligometastatic?
I don't know, but I do not consider a patient with 3 liver lesions and 2 bone lesions with an NSCLC "oligometastatic". Something tells me that it's just the tip of the iceberg...
Therefore perhaps for a number of patients that would fit into COMET standard of care would have been in many places to deliver SBRT, but I do think that a sizable number of patients would be treated systemically only.

2. I don't know if SBRT is that toxic. Surely, a few people died in COMET, but I've said it before and will say it again: I still wonder why it happened. We shouldn't be killing patients SBRT and I see no reason to put them at harms way for a palliative treatment. Constraints should have been tighter to ensure safery. It's different for Stage I NSCLC obviously, I would risk toxicity there since treatment is curative.

On the other hand, I am not sure if it was more easy to "sell" prostate RT to patients in Stampede than it was to "sell" them SBRT in Comet.
1. Patients often are scared of mets more than they are of the primary tumors. So if I have a controlled primary (inclusion criteria in COMET) and 1-5 uncontrolled mets, I'd love to go on a trial to get those treated. Not the same in Stampede. I have 20 untreated bone lesions, why should I care about my primary tumor in the prostate, it's not the one giving me all the back pain...
2. Treating with SBRT in Comet is quicker than 4 weeks of EBRT as in Stampede.

1. Majority (nearly 100%) of the patients had 3 or less metastases. While the inclusion criteria did go up to 5, I think a cut-off of 3 (similar to what is used in other oligometastatic studies) is not unreasonable.
You could only have 2 lesions within an organ to be eligible IIRC, so 3 liver lesions would be an exclusion criteria.

2. Agreed, it was a focus of his ASTRO presentation at the plenary this year to ensure treatment is safe. I believe, in part, LU-002 has decreased the dose to mediastinal lymph nodes to 34/5 in response to this.

In regards to prostate RT - the data is there IMO with STAMPEDE to offer a 'palliative' hypofractionated course (not 9 weeks) in patients with metastatic disease. It doesn't seem like common sense to us or patients but it's there. Lots of patients go "well the cancer was resected completely why do I need chemo or RT" and it's because we have trials showing their benefit. Ditto for this scenario.

Many oncologists treat based on phase II trial results ALL the time. TMZ instead of PCV in low grade Glioma. Changing chemotherapy regimens from what was done on phase III trials. IMO, It is time for our field as a whole to stop living in the world of "Phase III RCT or Bust", and either randomize these patients (if US based) to LU-002 (if applicable) or treat them (safely) based on the results of SABR-COMET.

In the interim, we will await results of the phase III version of this trial. I am also interested in SABR-COMET 10 to determine if it is safe to treat up to 10 sites of metastatic disease.

 

[scarbrtj]

In regards to prostate RT - the data is there IMO with STAMPEDE to offer a 'palliative' hypofractionated course (not 9 weeks) in patients with metastatic disease. It doesn't seem like common sense to us or patients but it's there. Lots of patients go "well the cancer was resected completely why do I need chemo or RT" and it's because we have trials showing their benefit. Ditto for this scenario.

Exactly & it's not a subtle point. Should be such... common sense! Actually common sense can be fickle; it misleads us at the most inopportune times. Data is data. Unless you're Warner Bezwoda, data doesn't lie. If yoga improves outcomes for breast cancer patients tested in a randomized trial, I don't have to explain why it does; I can just prescribe yoga and know (hope) it works. But... What the data says can be interpreted. So far, I'm interpreting SABR-COMET and STAMPEDE that there could be a role for local therapy in metastatic disease. but our local therapies are not likely to affect overall survival. One also not subtle point: could I have predicted that metastatic patients have >2y median overall survivals when I started in rad onc in the late 1990's? No! But now they do, sans XRT, per SABR-COMET. So local XRT has quite the OS row to hoe as systemics have (evidently) gotten so damn good.

I can imagine there are radiation oncologists who work more collaboratively with medical oncologists who refer appropriately who can't imagine that there are still many medical oncologists who will keep pumping systemic therapies in a patient with excellent performance status and a single liver met vs referring for SBRT. I work in a region with both extremes and it is mind blowing sometimes...

... which is why that there can be these extremes in approaches... SABR is not a clear-cut, unimpeachable winner as overall survivals haven't improved enough to be a "homerun" (like this). 'Til then, it's "mind-blowing" and annoying for the rad oncs!

 

[Fractionated]

Vinay Prasad recently featured SABR-COMET on his podcast. Discussion begins around 26 minutes in. Won't regurgitate his points here, but he doesn't believe this is practice changing and that while further investigation with phase iii trial is warranted, it will likely be negative. Thoughts?

Listen to 1.54 Association of Patient Characteristics and Tumor Genomics with Clinical Outcomes & SABR-COMET by Plenary Session #np on #SoundCloud

 

[evilbooyaa]

One of his main beefs seems to be sensational headlines as a result of the trial - not in the control of the authors.
Otherwise I don't overtly disagree with him. I think, ideally, lung cancer primary patients should enroll on LU-002 which includes this treatment paradigm (initially treated for localized disease, then recur with oligometastases) and the paradigm from the Gomez trial (upfront oligometastatic).

I think the remaining histologies should have, at bare minimum, a discussion that there is evidence it seems to be beneficial, but it is still considered somewhat experimental. I suppose I'm taking a step back in how definitive I think this is - I think not even OFFERING it to a patient or that it's not practice changing (in terms of offering no SBRT vs offering SBRT) because it's 'only a phase II trial' is silly, given that we treat based on non-randomized phase II compared to historical controls (as I mentioned above) all the time.

I hesitate about extrapolating this to histologies NOT included on SABR-COMET.

All that being said, many places are (and have been) doing this without any randomized trial evidence for a while - by this being not practice changing, it means they will continue doing so.

 

[Palex80]

From all histologies, apart from SCLC, I personally think NSCLC is the least promising. I rather think prostate, breast and renal hold more momentum. Perhaps colorectal too. But not NSCLC.

Why NSCLC has been studied so much is clear: high incidence, lots of patients presenting with primary stage IV disease and rather short follow-up necessary, since (until recently) most patients died within 2 years after being diagnosed with stage IV.


But that‘s just my belief...

 

[scarbrtj]

Vinay Prasad recently featured SABR-COMET on his podcast. Discussion begins around 26 minutes in. Won't regurgitate his points here, but he doesn't believe this is practice changing and that while further investigation with phase iii trial is warranted, it will likely be negative. Thoughts?

Listen to 1.54 Association of Patient Characteristics and Tumor Genomics with Clinical Outcomes & SABR-COMET by Plenary Session #np on #SoundCloud

Damn he stole my talking points. Hate when I agree with the med oncs aka our overlords. Follow Vinay on Twitter btw good topics discussions etc. I also predicted a non-sign OS advantage. The signs are there. As he says, the control arm differed from the study arm by only two patient events (deaths). Wish that it were not so but...

 

[wandering star]

Palma is a smart guy, but a statistical purist he's not. Again, with a p-value of 0.09, this study is technically actually MORE helpful to say that SABR WON'T affect survival (in additional replicating trials), binary p< or >0.05 odds-wise, regardless of Palma's sui generis verbiage and non-standard pre-specified p-value. But I'm just some guy on the internet, ignore me if you want.

It's easy to think about. I collected some data from two groups looking at a single metric. I compared the metric of the two groups; they were different (P=0.09). Now I go and get completely new data from a new sample from the population. What are the odds I will find a difference that is significant at p<0.05 or less. Should be obvious it's less than 50%. What are the odds they will be different at p<0.005? Very, very low.

I actually think that the article you quoted ("Scientists rise up against statistical significance") is making the opposite point. The biostatisticians who signed that letter are trying to *discourage* binary categorization into "p>0.05, that's trash" versus "p<0.05, it's good to go." Rosnow put it as follows:

"... surely, God loves the .06 nearly as much as the .05. Can there be any doubt that God views the strength of evidence for or against the null as a fairly continuous function of the magnitude of p?”

Whether a Phase III trial would succeed or fail really depends on all the other information in the universe that is NOT in this trial itself. If, instead of SABR, it was a Phase II trial evaluating different flavors of jellybeans for curing cancer, and we knew that 100 trials had been kicked off simultaneously and 10 of them had reached P = 0.09, then we would be very reluctant to expect that a Phase III trial to be successful. (And in general, we are right to be skeptical. Lots of people try plausible but useless "jellybean" interventions, some of them work out in Phase II and we read headlines, and many of these do fail at Phase III)

If, on the other hand, secondary analyses from other trials are consistent with SABR-COMET -- to the extent that we say, look at STAMPEDE, a couple of tumors and the body and we SABR a few of 'em and see benefits, maybe that's the same story? To the extent that multiple trials point the same direction, we might be more hopeful for a positive result.

In a world with no other prior information, if we see P = 0.09 and we repeat the same trial with the same umber of patients, I agree that we're unlikely to get P < 0.05. But if we use many more patients, our confidence intervals close in.

In summary, I agree with everyone else in this thread and think a phase III trial is warranted.

 

[scarbrtj]

I actually think that the article you quoted ("Scientists rise up against statistical significance") is making the opposite point. The biostatisticians who signed that letter are trying to *discourage* binary categorization into "p>0.05, that's trash" versus "p<0.05, it's good to go." Rosnow put it as follows:

"... surely, God loves the .06 nearly as much as the .05. Can there be any doubt that God views the strength of evidence for or against the null as a fairly continuous function of the magnitude of p?”

Whether a Phase III trial would succeed or fail really depends on all the other information in the universe that is NOT in this trial itself. If, instead of SABR, it was a Phase II trial evaluating different flavors of jellybeans for curing cancer, and we knew that 100 trials had been kicked off simultaneously and 10 of them had reached P = 0.09, then we would be very reluctant to expect that a Phase III trial to be successful. (And in general, we are right to be skeptical. Lots of people try plausible but useless "jellybean" interventions, some of them work out in Phase II and we read headlines, and many of these do fail at Phase III)

If, on the other hand, secondary analyses from other trials are consistent with SABR-COMET -- to the extent that we say, look at STAMPEDE, a couple of tumors and the body and we SABR a few of 'em and see benefits, maybe that's the same story? To the extent that multiple trials point the same direction, we might be more hopeful for a positive result.

In a world with no other prior information, if we see P = 0.09 and we repeat the same trial with the same umber of patients, I agree that we're unlikely to get P < 0.05. But if we use many more patients, our confidence intervals close in.

In summary, I agree with everyone else in this thread and think a phase III trial is warranted.

Bayesian, yes.

surely, God loves the .06 nearly as much as the .05 Can there be any doubt that God views the strength of evidence for or against the null as a fairly continuous function of the magnitude of p?

Who doth truly know the mind of God? Actually, God spoke to me and says he does love the 10E-6 much more than the 0.05.

 

[scarbrtj]

An update.

https://ascopubs.org/doi/10.1200/JCO.20.00818

 

[Chartreuse Wombat]

An update.

https://ascopubs.org/doi/10.1200/JCO.20.00818

"A post hoc sensitivity analysis that excluded patients with prostate cancer was consistent with a treatment benefit, with 5-year OS rates of 16.2% (95% CI, 5% to 32%) v 33.1% (95% CI, 20% to 47%), respectively (stratified log-rank test P = .085)."

Will there be enough equipoise to answer the question in a Phase III format?

 

[Krukenberg]

"A post hoc sensitivity analysis that excluded patients with prostate cancer was consistent with a treatment benefit, with 5-year OS rates of 16.2% (95% CI, 5% to 32%) v 33.1% (95% CI, 20% to 47%), respectively (stratified log-rank test P = .085)."

Will there be enough equipoise to answer the question in a Phase III format?

I think with this many positive phase 2s, do a Bayesian trial with weakly informative prior probability of benefit. Lower sample sizeneeded and we can get a more definitive answer faster

 

[Chartreuse Wombat]

The problem I have with SABR is the enormous heterogeneity of primary sites included. The study is so small that testing for interaction is woefully underpowered. There are primary specific similar trials ongoing. In the prostate realm the next iteration of STAMPEDE will provide the answer of whether to treat the mets; probably within a few years.

 

[deleted774703]

SABR COMET great and amazing OS signal, but def not practice changing on its own.

Don't get feelings hurt by hearing that

66 pts in SABR arm. extremely heterogenous as @Chartreuse Wombat pointed out

Need P3

 

[theradiator]

The problem I have with SABR is the enormous heterogeneity of primary sites included. The study is so small that testing for interaction is woefully underpowered. There are primary specific similar trials ongoing. In the prostate realm the next iteration of STAMPEDE will provide the answer of whether to treat the mets; probably within a few years.

Let me push back a little on this. Drug companies and medical oncologists have been running basket trials of heterogeneous disease sites based on a biomarker. Since we currently do not have a validated biomarker for durable disease control with RT for metastatic disease, it is necessary to use phenotypic selection of good performance status and limited sites of disease agnostic of primary site. Obviously, Palma and colleagues should be congratulated on their study that was pushed through without the deep pockets of pharma.

 

[OTN]

SABR COMET great and amazing OS signal, but def not practice changing on its own.

Don't get feelings hurt by hearing that

66 pts in SABR arm. extremely heterogenous as @Chartreuse Wombat pointed out

Need P3

If you were a patient, would you want SBRT to oligomets based on the COMET data? I can say with complete confidence that I would.

 

[deleted774703]

If you were a patient, would you want SBRT to oligomets based on the COMET data? I can say with complete confidence that I would.

I 100% agree

But I also want to be fair in my rad Onc paper critique scientifically

I do the same with drug papers

 

[MegaVoltagePhoton]

We are too hard on ourselves as a speciality. You can’t have it both ways either with whining about hypofrac/sbrt for prostate Or breast but then attacking something that will clearly expand our role.

there’s being dogmatic and there’s being pragmatic.

Also anyone who treats these patients off protocol instead of on protocol can’t then come back and whine about how the only studies that exist are small.

 

[deleted774703]

We are too hard on ourselves as a speciality. You can’t have it both ways either with whining about hypofrac/sbrt for prostate Or breast but then attacking something that will clearly expand our role.

there’s being dogmatic and there’s being pragmatic.

Also anyone who treats these patients off protocol instead of on protocol can’t then come back and whine about how the only studies that exist are small.

It’s not both ways

I’m 100% in support of SBRT in metastatic disease

But I’m also in support of P3 disease site-specific SBRT

That’s why the SINDAS trial at ASCO is truly practice changing in my mind

 

[MegaVoltagePhoton]

I 100% agree

But I also want to be fair in my rad Onc paper critique scientifically

I do the same with drug papers

And this is good but at the same time no one cares about your drug paper critiques because you don’t prescribe drugs, whereas as purveyors of radiation people also might not care what we say but recognize our expertise in this setting.

this doesn’t mean that we should accept bad studies, at all. But there’s also no need to be a constant wet blanket “welll it’s just phase II” especially if you yourself are not opening one of the confirmatory phase iii trials are your site

 

[deleted774703]

And this is good but at the same time no one cares about your drug paper critiques because you don’t prescribe drugs, whereas as purveyors of radiation people also might not care what we say but recognize our expertise in this setting.

this doesn’t mean that we should accept bad studies, at all. But there’s also no need to be a constant wet blanket “welll it’s just phase II” especially if you yourself are not opening one of the confirmatory phase iii trials are your site

I’m enrolling into phase 3 so I’m comfortable with myself

Enrolled 2 ppl onto iyengar trial on my attending behalf

 

[MegaVoltagePhoton]

I’m enrolling into phase 3 so I’m comfortable with myself

Enrolled 2 ppl onto iyengar trial on my attending behalf

that is good to hear

 

[evilbooyaa]

If there is an acceptable Ph III trial version of SABR-COMET to offer my patient, I will offer enrollment. Otherwise I will treat as per excellent ph II data.

This concept of 'wait for a phase III' is an imbalanced expectation for something that has so dramatic of a benefit.

 

[medgator]

If there is an acceptable Ph III trial version of SABR-COMET to offer my patient, I will offer enrollment. Otherwise I will treat as per excellent ph II data.

This concept of 'wait for a phase III' is an imbalanced expectation for something that has so dramatic of a benefit.

Correction: you'll treat "when insurance permits"?

 

[Chartreuse Wombat]

If there is an acceptable Ph III trial version of SABR-COMET to offer my patient, I will offer enrollment. Otherwise I will treat as per excellent ph II data.

This concept of 'wait for a phase III' is an imbalanced expectation for something that has so dramatic of a benefit.

The history of RPII trials is that the estimates of benefit are NEARLY always overestimates if a PhIII is done. I suggest tempering your confidence for a "dramatic" benefit.

 

[xrt123]

An arm of Stampede and HORRAD trials showed a survival benefit to treating just ONE of the sites of oligometastatic disease. Both phase 3. For now I would believe good phase 2 evidence such as COMET even if it’s an overestimate of benefit. I think many would agree it is likely an overestimate of a benefit to your point.

 

[Chartreuse Wombat]

HORRAD and STAMPEDE support treating the PRIMARY in prostate with small volume metastases. The STAMPEDE group will answer the "treat the metastases" question in short order.

 

[Palex80]

HORRAD and STAMPEDE support treating the PRIMARY in prostate with small volume metastases. The STAMPEDE group will answer the "treat the metastases" question in short order.

Add to that that ONLY the low-burden metastatic subset of STAMPEDE showed an improvement. HORRAD was negative (although perhaps underpowered).
STAMPEDE adressed patients with a quite favorable prognosis: metastatic hormone-sensitive prostate cancer at first line treatment (with 3+ lines of effective systemic treatment left). Extrapolating SABR-COMET results to "any" first-line oligometastatic scenario of any tumor should be done very carefully. Patients wi high-burden mHSPC did not have an OS-benefit in STAMPEDE.

 

[evilbooyaa]

Correction: you'll treat "when insurance permits"?

I suppose, yes, although this is something I would at least attempt a p2p aggressively on, as I would feel as strongly about it if it was my family member.

 

[medgator]

I suppose, yes, although this is something I would at least attempt a p2p aggressively on, as I would feel as strongly about it if it was my family member.

Good luck...cigna thinks it's experimental.... They pretty much abhor igrt on anything palliative too

 

[evilbooyaa]

Good luck...cigna thinks it's experimental.... They pretty much abhor igrt on anything palliative too

I'm planning to be bright-eyed attending who is willing to p2p and throw insurance companies under the bus in a chart on a daily basis as necessary for treatments that I feel very strongly about.

How long I'll last at that level of enthusiasm before the medical insurance establishment beats it out of me will be determined.

Them wanting to deny IGRT on a palliative case is small beans to me in comparison to denying an entire course of life-prolonging treatment.

 

[Neuronix]

I'm planning to be bright-eyed attending who is willing to p2p and throw insurance companies under the bus in a chart on a daily basis as necessary for treatments that I feel very strongly about.

Just pissing in the wind my friend, unfortunately.