New Carbon Ion phase III RCT

[xrthopeful]

I think this trial is super ambitious and may be hard to accrue, but I'm glad people in the field are taking big swings like this.

---

Members don't see this ad.

 

[Neuronix]

Interesting. I wonder how that trial is powered and expected time to accrual?

 

[Palex80]

ClinicalTrials.gov

clinicaltrials.gov

OS at 2 years is primary endpoint and they plan to include 103 patients. It is going to be a statistical disaster... They probably "suggested" that carbons are going to add something like 25% to the OS-rate at 2 years, which is illusional. Not even surgery does that...

 

[nkmiami]

What is the dose in the trial?

 

[tjquinn]

What is the dose in the trial?

Radiation: Carbon ion radiotherapy
Japan: 55.2 GyE in 4.6 GyE per fraction in 12 fractions delivered 4 days a week.
European: 59.4 GyE in 4.95 GyE per fraction in 12 fractions delivered 4 days a week
Other Name: CIRT

Drug: Gemcitabine
Concurrent gemcitabine at 1000 mg/m2 will be delivered during weeks 1-3 of CIRT. 1000 mg/m2 weekly, three on/one off for 4 cycles following chemoradiotherapy
Other Name: Gemzar

Drug: nab-paclitaxel
125 mg/m2 weekly, three on/one off for 4 cycles following chemoradiotherapy
Other Name: Abraxane

Radiation: Photon radiotherapy
50.4 Gy in 1.8 Gy per fraction in 28 fractions delivered 5 days a week.
Other Name: IMRT

Drug: Gemcitabine
Concurrent gemcitabine at 600 mg/m2 will be delivered during weeks 1-5 of IMRT. 1000 mg/m2 weekly, three on/one off for 4 cycles following chemoradiotherapy
Other Name: Gemazr

Drug: nab-paclitaxel
125 mg/m2 weekly, three on/one off for 4 cycles following chemoradiotherapy
Other Name: Abraxane

 

[Chartreuse Wombat]

So much not to like here.

2:1 randomization favoring carbon. Non equal allocation is inefficient method and makes it more likely that the results will be inconclusive (but probably makes it easier to accrue but ethically I have big problem-can you say therapeutic misconception..

Primary Outcome Measures :

1. 2 year overall survival [ Time Frame: 2 years ]
   Overall survival outcomes will be estimated using the Kaplan-Meier approach and compared between the arms using the log-rank test. The Cox proportional hazard regression model will be used to determine hazard ratios and 95% confidence intervals for the treatment difference in overall survival. Unadjusted ratios and ratios adjusted for stratification variables and other covariates of interest will be computed.

OK well enough but the underlying assumption must be that the HR is 0.5 or something like it.

Just in case we don't get the result we want-

Other Outcome Measures:

1. Quality-adjusted survival time [ Time Frame: 3 years ]
   The quality-adjusted survival time estimates need to account for the presence of censoring. Due to the induced informative censoring problem, the ordinary survival method (e.g., Kaplan-Meier estimator) cannot be applied in this case. Accordingly, the investigators will use the inverse-probability weighted method of Zhao and Tsiatis to carry out the survival time analysis to estimate quality adjusted survival time, data from EQ-5D will first be translated into utility measures. These measures are obtained at discrete time points, so they will be interpolated into the time intervals between the visits. The quality-adjusted survival time is just an integration of the utility measures over a patient's survival time, or until the time limit similar as the cost calculation, whichever occurs earlier

Hmmm

 

[SneakyBooger]

So this is where we are "valuing" our allocated dollars...

Who has carbon ion in USA? No one. So they are flying those randomized to carbon ion elsewhere.

 

[thecarbonionangle]

SAD. Choose wisely folks!!

 

[scarbrtj]

I think this trial is super ambitious and may be hard to accrue, but I'm glad people in the field are taking big swings like this.

CIPHER? They intentionally called the trial CIPHER?

 

[Krukenberg]

So much not to like here.

2:1 randomization favoring carbon. Non equal allocation is inefficient method and makes it more likely that the results will be inconclusive (but probably makes it easier to accrue but ethically I have big problem-can you say therapeutic misconception..

Primary Outcome Measures :

1. 2 year overall survival [ Time Frame: 2 years ]
   Overall survival outcomes will be estimated using the Kaplan-Meier approach and compared between the arms using the log-rank test. The Cox proportional hazard regression model will be used to determine hazard ratios and 95% confidence intervals for the treatment difference in overall survival. Unadjusted ratios and ratios adjusted for stratification variables and other covariates of interest will be computed.

OK well enough but the underlying assumption must be that the HR is 0.5 or something like it.

Just in case we don't get the result we want-

Other Outcome Measures:

1. Quality-adjusted survival time [ Time Frame: 3 years ]
   The quality-adjusted survival time estimates need to account for the presence of censoring. Due to the induced informative censoring problem, the ordinary survival method (e.g., Kaplan-Meier estimator) cannot be applied in this case. Accordingly, the investigators will use the inverse-probability weighted method of Zhao and Tsiatis to carry out the survival time analysis to estimate quality adjusted survival time, data from EQ-5D will first be translated into utility measures. These measures are obtained at discrete time points, so they will be interpolated into the time intervals between the visits. The quality-adjusted survival time is just an integration of the utility measures over a patient's survival time, or until the time limit similar as the cost calculation, whichever occurs earlier

Hmmm

Trial is really underpowered to demonstrate a survival difference. Maybe 30-40% of patients die of local progression, so it would seem in order to get a positive trial they’d have to prevent all of those.

 

[xrthopeful]

If the trial even accrues it will be a success. Enrolling 100 patients in this era where many believe there is no role for RT at all for pancreatic cancer, let alone the part where 2/3 of them will have to go to japan, will be a challenge to say the least.

Kudos to them for taking on this venture. Will be interesting to see what happens.

 

[scarbrtj]

So much not to like here.

2:1 randomization favoring carbon. Non equal allocation is inefficient method and makes it more likely that the results will be inconclusive (but probably makes it easier to accrue but ethically I have big problem-can you say therapeutic misconception..

Primary Outcome Measures :

1. 2 year overall survival [ Time Frame: 2 years ]
   Overall survival outcomes will be estimated using the Kaplan-Meier approach and compared between the arms using the log-rank test. The Cox proportional hazard regression model will be used to determine hazard ratios and 95% confidence intervals for the treatment difference in overall survival. Unadjusted ratios and ratios adjusted for stratification variables and other covariates of interest will be computed.

OK well enough but the underlying assumption must be that the HR is 0.5 or something like it.

Just in case we don't get the result we want-

Other Outcome Measures:

1. Quality-adjusted survival time [ Time Frame: 3 years ]
   The quality-adjusted survival time estimates need to account for the presence of censoring. Due to the induced informative censoring problem, the ordinary survival method (e.g., Kaplan-Meier estimator) cannot be applied in this case. Accordingly, the investigators will use the inverse-probability weighted method of Zhao and Tsiatis to carry out the survival time analysis to estimate quality adjusted survival time, data from EQ-5D will first be translated into utility measures. These measures are obtained at discrete time points, so they will be interpolated into the time intervals between the visits. The quality-adjusted survival time is just an integration of the utility measures over a patient's survival time, or until the time limit similar as the cost calculation, whichever occurs earlier

Hmmm

In other words, like in a lot of trials nowadays,
"We plan to bribe a statistician so we can get the result we want."

 

[Lamount]

Does anyone know the accrual goal? I didn’t see it and I can’t get a sense of the power without it.

 

[xrthopeful]

Accrual goal 100

 

[SubserviantToABR]

All things considered, you gotta give credit where it is due. At least someone is stepping up to generate prospective data on the utility of particle therapy, in this case, carbon ions, unlike all of the proton people who s*** the bed with their unwillingness to produce data, building machines everywhere, bilk payors like they were an ATM, and complain when there is impending slashes to their reimbursement.

It took how long before they thought it was a good idea to commence with phase III studies? I mean, I'm just a dumb recent graduate, *EDITED BY MODS*, but besides the Loma Linda/Mass Gen dose escalation prostate proton study, are there any other randomized, phase III studies that have been published? And if so, has it changed your practice?

At least, if it works with carbon for pancreatic cancer, it can be a potential treatment option for those with this deadly disease. And if it doesn't, maybe there is not a need for heavy ion therapy, and we can realign our focus in radiation oncology on more important things. But hey, at least there will be data out there in the future if it accrues. But what do I know? I've been busy these days perusing the ASTRO Career Board looking for a custodial job, hopefully, I can clean the desk of a sarcoma expert.

 

[Palex80]

I wonder who funds trials like these... Flying 70 patients to Japan to receive carbon ion treatment for several weeks must be quite costly.

 

[SubserviantToABR]

Once the UKs proton machines are up and running lots of trials are planned !

I can't really hate on people who want to generate data, but it is hard to care later when the enthusiasm for protons is soured by unscrupulous radiation oncologists. Hopefully, they can do something like a proton breast or prostate study to show its utility, and when it is read out and published, I have already retired and jumped off of this runaway train. Fortunately, I have thoughts and prayers .

At least, and remember, I am a lowly, ignorant graduate, but my perception is that in the US, the research priorities have changed. I mean, look at any of the latest issues in our premier journal. Why push forward with running phase III studies when you can do survey studies with 2.1% response rates or write hit articles about SDN?

 

[SubserviantToABR]

I wonder who funds trials like these... Flying 70 patients to Japan to receive carbon ion treatment for several weeks must be quite costly.

You know, I've lurked on these forums for years and always enjoyed reading your posts, as you seemed like you have a handle on the data. It is always well-thought out, and as a bright-eyed, bushy-tailed trainee, I have learned a lot. Truly, I have.

So, it really bothers me, at least from the tone of your sentence, there is some disdain for this study. I do not know who funded this study, but from the sounds of it, I don't think it was done with public funds, so why hate on a study that when the data comes out, will be available to you and the public (well, of course, at that that is after you pay some journal a ridiculous amount of money for access because that is the way the world works...you do work for the publishers and journals) and answer a question for which we do not have an answer to?

You can keep hating on other radiation oncologists because I love to do the same , as you can tell with my recent posts, but keep in mind, in the medical oncology world, industry pays for new trials, most of the time, and broaden new indications for their drugs. So, we can hate on phase III trials in radiation oncology, meanwhile, the medical oncologists can find new ways to make sure we lose at the seat at the table with our patients:

 

[xrthopeful]

I share your opinion on it being a good thing to do this trial. But to be fair to Palex I don’t think he was being negative - just curious. I am as well because funding for rad onc trials hasn’t always been plentiful

 

[xrthopeful]

But YES. There is currently very limited role for RT in pancreatic cancer if you ask a med onc. We need to prove our worth in these trials or med oncs will just give chemo as long as they can before the patient either refuses or dies

 

[SubserviantToABR]

I share your opinion on it being a good thing to do this trial. But to be fair to Palex I don’t think he was being negative - just curious. I am as well because funding for rad onc trials hasn’t always been plentiful

In that case, I apologize to Palex for misinterpreting the tone of their comments. Keep on fighting the good fight and contributing to the forum. For us American trainees, please be sure to include vague genetic alterations in rare malignancies as our board feels that is much more important than being competent to take care of cancer patients.

 

[Palex80]

You know, I've lurked on these forums for years and always enjoyed reading your posts, as you seemed like you have a handle on the data. It is always well-thought out, and as a bright-eyed, bushy-tailed trainee, I have learned a lot. Truly, I have.

So, it really bothers me, at least from the tone of your sentence, there is some disdain for this study. I do not know who funded this study, but from the sounds of it, I don't think it was done with public funds, so why hate on a study that when the data comes out, will be available to you and the public (well, of course, at that that is after you pay some journal a ridiculous amount of money for access because that is the way the world works...you do work for the publishers and journals) and answer a question for which we do not have an answer to?

You can keep hating on other radiation oncologists because I love to do the same , as you can tell with my recent posts, but keep in mind, in the medical oncology world, industry pays for new trials, most of the time, and broaden new indications for their drugs. So, we can hate on phase III trials in radiation oncology, meanwhile, the medical oncologists can find new ways to make sure we lose at the seat at the table with our patients:

Thank you for your comment.

1. I do not "hate" this trial. I find the question asked a relevant one. I do not think the methods used will result in a statement that we can uphold. It is clear, that the role of local RT in this setting (locally advanced, non-metastastic, irresectable pancreatic cancer) AT ALL is questionable, based on the data from the French trial which showed no difference whatsoever with our without RT after response to chemo.
Dose-escalated RT with carbon ions may indeed increase local control with hopefully favorable toxicity BUT patients are still going to die due to metastatic disease. Metastatic disease which was already there at presentation and was not picked up during staging (please bear in mind that according to the study's protocol they will be using CT to stage, meaning a lot of small metastases, for example in the liver, may be overseen).
The primary endpoint is overall survival at 2 years which is not good in these patients and will be influenced only little by local treatment. You would need a very big trial to show that. I presume that local treatment may raise OS from 25% to 30% at 2 years... 100 patients randomized will not prove that.
These are my concerns.

2. The comment was purely out of curiosity.
Flying 70 patients to Japan for treatment, paying for their medical expenses, lodging, food there (which will not be reimbursed by the US insurers, I assume) and bearing in mind that patients with pancreatic cancer do tend to develop complications / problems requiring hospitalization (especially when you are treating in the upper abdomen with 5 Gy fractions/day) is going to be expensive, not to mention the cost of the carbons (which I presume are state-funded by Japan?). I have no idea what they plan to do with the relatives (do they get plane tickets to Japan too?), I assume that would be at least nice bearing in mind that a quarter of these patients will be dead in less than a year? I would assume costs >100.000$ / patient --> >7 million for 70 patients. That's a lot of money. That was the point. I have no idea, who funds trials like these.
Prostate cancer trials for instance which enjoy no industry support occasionally get funded by the Movember Foundation.
I am not aware of something like that in pancreatic cancer too.

Now, yes, I do think that you can run a lot better trials with 7 million dollars in radiation oncology which will actually generate data and allow us to draw conclusions that will be practice changing. Want some examples?
1. A randomized trial testing observation versus stereotactic fractionated RT versus radiosurgery versus whole brain RT for patients with a resected brain metastasis. We have trials that have looked into some of these subsets, but not one trial that has looked in all.
2. A randomized trial on volumes in head & neck cancer. Do we really need to treat the bilateral neck in all cancers (except confined tonsilar cancer and early glottic cancer)? Non-inferiority design looking at PFS.
3. A randomized trial on volumes in preoperative rectal cancer RT. Do we really need to treat the pelvis as we do or would treating the mesorectum only suffice? Non-inferiority design looking at PFS.
4. A trial on standard of care RCT in glioblastoma multiforme patients. Do the fit patients really need 60/2? So far, we only give 40/2.67 or whatever else hypofractionated scheme for the less-fit patients. But do we really need 60/2 in the fit ones? Can't we use moderate hypofractionation to shorten therapy for the "fit" patients too (of which again a quarter will be dead by the end of year 1?).
5. A trial on preoperative RT with a shorter schedule for extremity sarcomas. Do we need 50/2 in the neoadjuvant setting? Perhaps a shorter regime would be feasible and produce the same PFS-benefits with same toxicity?

And guess what... We probably need less than 7 millions for each of these trials.

 

[xrthopeful]

The next NRG HN trial is looking at elective volume de-escalation for HPV oropharynx patients.

 

[xrthopeful]

Also there are multiple institutions prospectively studying 30/5 in the pre-op setting for sarcoma

 

[medgator]

Also there are multiple institutions prospectively studying 30/5 in the pre-op setting for sarcoma

Great, yet another thorn in the side of the specialty grappling with residency expansion. I don't see 30/5 for a niche disease really moving the needle for the specialty, personally.

 

[xrthopeful]

Great, yet another thorn in the side of the specialty grappling with residency expansion. I don't see 30/5 for a niche disease really moving the needle for the specialty, personally.

Just addressing one of the potential interesting questions raised by Palex

 

[SneakyBooger]

I wonder who funds trials like these... Flying 70 patients to Japan to receive carbon ion treatment for several weeks must be quite costly.

"Who cares who funds what. We represent Wall Street now and could care less about the opinions of rad oncs. We own the training programs through donations. Shareholders are currently the most important component of healthcare dollars. Get over it."

Is that type of response correct?

Of course not.

 

[evilbooyaa]

ClinicalTrials.gov

clinicaltrials.gov

OS at 2 years is primary endpoint and they plan to include 103 patients. It is going to be a statistical disaster... They probably "suggested" that carbons are going to add something like 25% to the OS-rate at 2 years, which is illusional. Not even surgery does that...

Oof, no way they're going to see that big of a difference.

 

[evilbooyaa]

I thought UTSW was building a Carbon center? Are they really going to be sending patients out of the country for this trial?

I look forward to the conclusion: Carbon ion did not improve outcomes, but that's because it wasn't done in the hands of UTSW! Now that we have a machine we will be treating all pancreatic cancers with Carbon ion despite the results of this trial as it is feasible and shows similar toxicity to photons.

 

[deleted774703]

Should have chosen PFS as primary endpoint.

That’s what most of the garbage immunotherapy trials use and then make 100 million dollars.

Only radonc is stupid enough to pigeonhole themselves with OS as endpoint.

 

[xrthopeful]

Lol this thread is almost a parody in demonstrating how people here will find a way to get mad about literally anything. Even when one of the things they were mad about before is how our field never does anything - someone even literally said like ‘we don’t even bother to look at carbon ions more’.

Never change, SDN

 

[medgator]

Lol this thread is almost a parody in demonstrating how people here will find a way to get mad about literally anything. Even when one of the things they were mad about before is how our field never does anything - someone even literally said like ‘we don’t even bother to look at carbon ions more’.

Never change, SDN

Yeah carbon ions. That's the ticket.. how many of us trained or are currently training with them BTW?

 

[Lamount]

Yeah carbon ions. That's the ticket.. how many of us trained or are currently training with them BTW?

Probably could have made the same argument about IMRT in the 90s. This is what we SHOULD be doing as field... and if it doesn't work, try something else like FLASH or radiation sensitivity assays. Or better yet, try everything. Just because you don't know how to do something, it doesn't make it a bad idea.

 

[medgator]

Probably could have made the same argument about IMRT in the 90s. This is what we SHOULD be doing as field... and if it doesn't work, try something else like FLASH or radiation sensitivity assays. Or better yet, try everything. Just because you don't know how to do something, it doesn't make it a bad idea.

Did it cost the same? Is the therapeutic benefit going to be the same going from photons to carbon ion vs going from 3D to imrt?

The proton guys still haven't proven protons are superior to photons at this point in places like lung and prostate

 

[xrthopeful]

Do you know how ridiculous you sound? We’re not talking about people doing carbon ions and getting paid more for it without data like proton.

This is a 100 patient phase III RCT. The way it’s supposed to be done - on a trial.

And it’s in a disease site where there isn’t much of a role for RT anyways so they’re not stealing patients from you (or the theoretical you) - but rather trying to see if there may be a new space.

There shouldn’t be any backlash to trying to expand what we can do as a field. If everyone just sticks with the status quo we will be dinosaurs faster than you think.

 

[Lamount]

Did it cost the same? Is the therapeutic benefit going to be the same going from photons to carbon ion vs going from 3D to imrt?

Don't know, let's investigate (old school, I know).

The proton guys still haven't proven protons are superior to photons at this point in places like lung and prostate

...and they may never prove that it is. But there is a scientific basis to investigate, so investigate we shall. And maybe, it will turn out that protons are only helpful for NSCLC when V20 > 30 or V5 > 50 with IMRT... and THAT will be the indication.

It's reasonable to skeptical that a new idea will be successful, but it is not reasonable pigeonhole scientists for having the audacity to look. Med Oncs don't eat their own for exploring innovative new therapies...so why do we?

 

[Palex80]

Probably could have made the same argument about IMRT in the 90s. This is what we SHOULD be doing as field... and if it doesn't work, try something else like FLASH or radiation sensitivity assays. Or better yet, try everything. Just because you don't know how to do something, it doesn't make it a bad idea.

I do not recall a randomized trial in the 90s adressing IMRT vs. 3D designed to prove a survival benefit in any disease and certainly not on of the most lethal ones, where the overall role of RT is debateable.

 

[Palex80]

Do you know how ridiculous you sound? We’re not talking about people doing carbon ions and getting paid more for it without data like proton.

This is a 100 patient phase III RCT. The way it’s supposed to be done - on a trial.

And it’s in a disease site where there isn’t much of a role for RT anyways so they’re not stealing patients from you (or the theoretical you) - but rather trying to see if there may be a new space.

There shouldn’t be any backlash to trying to expand what we can do as a field. If everyone just sticks with the status quo we will be dinosaurs faster than you think.

The problem is not the aim itself, IMHO. The question is interesting and should be adressed: „Can escalation if local treatment cure more patients witg unresectable pancreatic cancer?“

The problem is how they do it. With a 100–patient randomized phase 3 trial, that is based on either a ridiculous hypothesis or severely underpowered. This is the issue.
And it can also backfire. Med. Oncs. are going to say, that if carbons dont work, we shouldn’t be doing any RT... since carbons are „so much better“ than photons.

 

[xrthopeful]

I do not recall a randomized trial in the 90s adressing IMRT vs. 3D designed to prove a survival benefit in any disease and certainly not on of the most lethal ones, where the overall role of RT is debateable.

Are you arguing for the opposite of medgator - that we should be using carbon ion routinely and have it paid for without studying it prospectively at all?

 

[xrthopeful]

I agree that study design is difficult but also hard to do much else if you’re going to do a randomized trial. I doubt it will be a positive trial for OS but there will be other secondary outcomes as well as the fact that we will have some proper prospective data to understand how carbon works in the clinic. All sorts of things will come out of this. I’m glad the study leads were able to get funding for this. We need more funding for Rad Onc trials that don’t include us begging at pharma company teats


We should not let perfect become the enemy of good

 

[Lamount]

The problem is not the aim itself, IMHO. The question is interesting and should be adressed: „Can escalation if local treatment cure more patients witg unresectable pancreatic cancer?“

The problem is how they do it. With a 100–patient randomized phase 3 trial, that is based on either a ridiculous hypothesis or severely underpowered. This is the issue.
And it can also backfire. Med. Oncs. are going to say, that if carbons dont work, we shouldn’t be doing any RT... since carbons are „so much better“ than photons.

This is certainly not meant to be the definitive study and was likely conceived knowing full-well that it was underpowered for its "primary endpoint". However, if for example, this trial failed to demonstrate superiority of CIT compared to the control arm, but was able to demonstrate that CIT performed significantly better than historical controls, it would likely stimulate further investigation and may provide the necessary preliminary data for additional funding.

Clearly there is a catch 22 because, as of right now, there isn't enough data to get funding for the sort of RCT that would be powered to ask this question the right way.... thus an underpowered trial with a 2:1 randomization seems like a reasonable first step.

As for your concerns about perceptions, I think that this is completely the wrong mentality. Do they base their preliminary trial designs around your opinion? If it turns out that CIT is no better (i.e. not even a signal suggestive of better outcomes), then skepticism is appropriate. If it turns out that there is a signal but it fails to meet statistical significance, we will have more ammunition for the next study. Med oncs will be the last people to support anything and we shouldn't bother engaging them in this discussion or considering their opinion until we actually have something to talk about... and if we have a new therapy that offers better outcomes, we should present it to med oncs as if their support is a forgone conclusion, not like we are begging for the privilege of their approval.

 

[OTN]

I thought UTSW was building a Carbon center? Are they really going to be sending patients out of the country for this trial?

I look forward to the conclusion: Carbon ion did not improve outcomes, but that's because it wasn't done in the hands of UTSW! Now that we have a machine we will be treating all pancreatic cancers with Carbon ion despite the results of this trial as it is feasible and shows similar toxicity to photons.

My understanding is that the funding for UTSW's very ambitious particle program has gone away.

 

[nkmiami]

One issue that I have with this trial is that it may be possible to deliver a similar dose with IMRT. From what I understand Chris Crane and MSKCC are treating most of the unresectable pancreatic cancers to slightly higher beds, but someone should correct me if I am wrong. No matter what the modality, margin, imaging and motion, are going to dictate you not cover the full PTV or gtv to this dose. (at least 1/3 of locally advanced pancreatic cancers involve the duodenum. Even in prostate the anterior rectal wall does not get a different dose with photon than protons due to having a PTV?
If this is the case, why not just have the patient fly to New York. The trial is testing dose escalation, not photon vs carbon. It very well maybe easier to deliver higher doses to the gtv with an mri/linac or existing linac.

 

[Palex80]

I agree that study design is difficult but also hard to do much else if you’re going to do a randomized trial. I doubt it will be a positive trial for OS but there will be other secondary outcomes as well as the fact that we will have some proper prospective data to understand how carbon works in the clinic. All sorts of things will come out of this. I’m glad the study leads were able to get funding for this. We need more funding for Rad Onc trials that don’t include us begging at pharma company teats
We should not let perfect become the enemy of good

Clearly there is a catch 22 because, as of right now, there isn't enough data to get funding for the sort of RCT that would be powered to ask this question the right way.... thus an underpowered trial with a 2:1 randomization seems like a reasonable first step.

That is not ethical.
You don't simply "randomize" people in a trial, when you know that your hypothesis is wrong all along or your trial is underpowered.
That's not how trials should be run!

I wonder what these guys wrote in the protocol and made their sample size calculation. Die they simply assume a huge benefit through carbon ions to justify the 103 patients?

It doesn't have to do with "perfect". this is simply not good clinical practice. If you can't fund a randomized trial with a decent hypothesis, then don't run the trial in that way and simply change your study design!

 

[nkmiami]

You "doubt" that the trial will be positive, yet you think we will find out many things about carbon ions. Great. Then why do the trial in a randomized matter, why do it with an endpoint many (if not all of us) doubt, will be positive. What has happened to good clinical practice?

We have data that without surgery and 50 Gy radiation, you get 0-5% (closer to 0) long term survivors. We have nonrandomized data that with around 100 Gy bed, 25% pts with 3 year and possibly longer survival, so it wouldnt surprise me trial is positive.

Focal Radiation Therapy Dose Escalation Improves Overall Survival in Locally Advanced Pancreatic Cancer Patients Receiving Induction Chemotherapy and Consolidative Chemoradiation - PubMed

Radiation dose escalation during consolidative chemoradiation therapy after induction chemotherapy for LAPC patients improves OS and local-regional RFS.

www.ncbi.nlm.nih.gov

Ablative radiation therapy for locally advanced pancreatic cancer: techniques and results

Standard doses of conventionally fractionated radiation have had minimal to no impact on the survival duration of patients with locally advanced unresectable pancreatic cancer (LAPC). The use of low-dose stereotactic body radiation (SBRT) in 3- to 5-fractionshas ...

www.ncbi.nlm.nih.gov

 

[Palex80]

We have data that without surgery and 50 Gy radiation, you get 0-5% (closer to 0) long term survivors. We have nonrandomized data that with around 100 Gy bed, 25% pts with 3 year and possibly longer survival, so it wouldnt surprise me trial is positive.

Focal Radiation Therapy Dose Escalation Improves Overall Survival in Locally Advanced Pancreatic Cancer Patients Receiving Induction Chemotherapy and Consolidative Chemoradiation - PubMed

Radiation dose escalation during consolidative chemoradiation therapy after induction chemotherapy for LAPC patients improves OS and local-regional RFS.

www.ncbi.nlm.nih.gov

Ablative radiation therapy for locally advanced pancreatic cancer: techniques and results

Standard doses of conventionally fractionated radiation have had minimal to no impact on the survival duration of patients with locally advanced unresectable pancreatic cancer (LAPC). The use of low-dose stereotactic body radiation (SBRT) in 3- to 5-fractionshas ...

www.ncbi.nlm.nih.gov

Indeed, there is potential in there. However those were highly selected patients with favorable anatomy and response (or at least stable disease) to neoadjuvant chemo. Not exactly the same population compared to those that will be treated in the randomized trial.

 

[Lamount]

That is not ethical.
You don't simply "randomize" people in a trial, when you know that your hypothesis is wrong all along or your trial is underpowered.
That's not how trials should be run!

I wonder what these guys wrote in the protocol and made their sample size calculation. Die they simply assume a huge benefit through carbon ions to justify the 103 patients?

It doesn't have to do with "perfect". this is simply not good clinical practice. If you can't fund a randomized trial with a decent hypothesis, then don't run the trial in that way and simply change your study design!

A very quick calculation demonstrates that they would have 80% power to detect an HR of 0.55 with their planned accrual. This is an ambitious goal (perhaps even a little pie-in-the-sky) but I don't think it is unethical.

There are two questions to consider
1) Are large RTCs with OS endpoints the best clinical trial model going forward?
-My personal opinion: In the US, probably not for novell/controversial treatments (i.e. CIT). Nonetheless, there are many who would refuse to use a parachute if it weren't found to be superior in an RTC, so here we are.

2) If an RTC is inconclusive regarding its primary endpoint, can one still draw important conclusions regarding the outcomes?
-My personal opinion: It depends. While statistical orthodoxy tells us that anything and everything short of a statistically significant outcome on the primary endpoint is a "negative" result, I think that there can be some laxity in this "rule" depending upon the question being asked and the reason that the endpoint could not be met. For example, if the CIPHER trial demonstrated a HR of 0.60 with p = 0.07, I personally would find this result interesting, especially if survival rates were much better than historical controls. Additionally, such a result COULD move the needle on more basic/translational research, larger clinical trials, and/or perhaps make a US CIT center slightly more likely.


I guess my question to you is: given that CIT has been used to treat abdominal tumors for decades and the toxicities are known, to whom is such a trial unethical?... the patients who (at bare minimum) will have very close follow up?... those who funded it (likely approved of the trial design)?... the reader?

I do agree with you that, with a purist view of statistics and clinical trial design, CIPHER dances on the edge of the rules... and I won't assume that you are calling the designers "unethical" but were merely proposing that it is "unethical" to design a trial that is destined to fail. I, however, think it is very reasonable to design a trial with a very ambitious endpoint when a negative result can still be fruitful.

 

[nkmiami]

I wouldnt overthink the stats. In the control arm, unless you have a pt with a rare BRCA or lynch mistmatch type mutation, you should have about 0% survival at 5 years. In the original chemorads/herskovic esophagus rtog trial, you also had a control arm with 0% survivors, and I think it had very few pts... I just think with margin and proximity to duodenum and moving target, you will not be able to cover 95% of gtv, just like with high dose photon techniques. Can they even do cone beam on carbon, because if you cant, you need to add extra margin because you may have bowel in the field based on daily variations. My guess is that Image guidance very well may be more important than particles vs photons. This is supposedly the advantage of an MRI linac, as seeing the exact position of bowel and replanning right before treatments.

 

[ramsesthenice]

One issue that I have with this trial is that it may be possible to deliver a similar dose with IMRT. From what I understand Chris Crane and MSKCC are treating most of the unresectable pancreatic cancers to slightly higher beds, but someone should correct me if I am wrong. No matter what the modality, margin, imaging and motion, are going to dictate you not cover the full PTV or gtv to this dose. (at least 1/3 of locally advanced pancreatic cancers involve the duodenum. Even in prostate the anterior rectal wall does not get a different dose with photon than protons due to having a PTV?
If this is the case, why not just have the patient fly to New York. The trial is testing dose escalation, not photon vs carbon. It very well maybe easier to deliver higher doses to the gtv with an mri/linac or existing linac.

Yes and no. He isn’t giving slightly higher doses. They are doing around 57/15 to the GTV plus margin and 90/15 to a 1 cm contracted volume. Patients have to be very carefully selected (> 1 cm from bowel, etc). You have to have excellent IGRT and some form of effective motion management as well as an insatiable sense of adventure/risk taking. You could use his approach to argue carbon ions make sense. At the very least, it suggests getting a higher BED might matter.

The design of this trial is deeply flawed and it makes some rather lofty assumptions. I doubt it teaches us too much in the end. But, the central hypothesis probably is scientifically sound.

 

[Palex80]

I guess my question to you is: given that CIT has been used to treat abdominal tumors for decades and the toxicities are known, to whom is such a trial unethical?... the patients who (at bare minimum) will have very close follow up?... those who funded it (likely approved of the trial design)?... the reader?

A very quick calculation demonstrates that they would have 80% power to detect an HR of 0.55 with their planned accrual. This is an ambitious goal (perhaps even a little pie-in-the-sky) but I don't think it is unethical.

There are two questions to consider
1) Are large RTCs with OS endpoints the best clinical trial model going forward?
-My personal opinion: In the US, probably not for novell/controversial treatments (i.e. CIT). Nonetheless, there are many who would refuse to use a parachute if it weren't found to be superior in an RTC, so here we are.

2) If an RTC is inconclusive regarding its primary endpoint, can one still draw important conclusions regarding the outcomes?
-My personal opinion: It depends. While statistical orthodoxy tells us that anything and everything short of a statistically significant outcome on the primary endpoint is a "negative" result, I think that there can be some laxity in this "rule" depending upon the question being asked and the reason that the endpoint could not be met. For example, if the CIPHER trial demonstrated a HR of 0.60 with p = 0.07, I personally would find this result interesting, especially if survival rates were much better than historical controls. Additionally, such a result COULD move the needle on more basic/translational research, larger clinical trials, and/or perhaps make a US CIT center slightly more likely.


I guess my question to you is: given that CIT has been used to treat abdominal tumors for decades and the toxicities are known, to whom is such a trial unethical?... the patients who (at bare minimum) will have very close follow up?... those who funded it (likely approved of the trial design)?... the reader?

I do agree with you that, with a purist view of statistics and clinical trial design, CIPHER dances on the edge of the rules... and I won't assume that you are calling the designers "unethical" but were merely proposing that it is "unethical" to design a trial that is destined to fail. I, however, think it is very reasonable to design a trial with a very ambitious endpoint when a negative result can still be fruitful.

I do agree with you that, with a purist view of statistics and clinical trial design, CIPHER dances on the edge of the rules... and I won't assume that you are calling the designers "unethical" but were merely proposing that it is "unethical" to design a trial that is destined to fail. I, however, think it is very reasonable to design a trial with a very ambitious endpoint when a negative result can still be fruitful.

I fully agree with you. I am not callng the designers unethical, I just find it not GCP-conformal to design a trial, which is highly probable to fail.
And the ethical committee / scientific review committee should have probably picked that up.

We don't know what's written in the patient information, but usually, you should make a statement in there concerning the benefit to the patient.
Not numbers, but you do need to write that you think the experiemental treatment is either better in terms of efficacy or bears less side effects or is more convenient/cheaper. These are the 3 things we aim for, when testing an experimental treatment.

And although the HR may be 0.55, the point is what these 0.55 represent. Is the trial designed to show an OS improvement roughly
a) from 10% to 20% at 2 years or
b) from 25% to 50% at years

If a) is the case, then 100 patients are not enough to show it and I hope that someone picked that up during the review process. If not, it was a bad review process.
If b) is the case, then I think that many of us think it's illusional to suggest that 50% are achievable. Not with this patient population at least.

So when the patient asks you, what improvement the investigators think carbon ion treatment may bring, you would have to make some kind of estimation / statement. And although it's a trial and you need to be very careful about what you say (no numbers!), the patient still has the right to know what you expect or hope to show with the experimental treatment. He may even ask to read the protocol.
Making illusional assumptions about the benefit of a treatment and then communicating these to the patient, raising his hopes for cure and making him more eager to join a trial is NOT GCP-conform. That's the point I am trying to make.

A patient who reads that the trial is designed to prove that the new treatment increases his chance of cure from 25% to 50% will join and hope to fly to Japan. A patient who reads that his chance of cure will increase from 10% to 20% will not be as eager probably.



So tell me doctor, I've read in the internet less than a quarter of patients like me are still alive 2 years after diagnosis. You suggest that I should enter this trial and maybe I'll get the chance to go to Japan. How much better do you think that treatment may be?