The Upcoming Randomized Trials in Oropharynx Cancer are Critically Important for the Proton Therapy Industry.
Looking at 2023, the best hope for a huge win for proton therapy in the field of radiation oncology is in the now closed to accrual Oropharynx trial.
Omissions or errors: I’m an author of one so, if I missed something please or this is a big trial that should have been included, comment and I’ll do some work and clean anything up.
The year ahead for Proton Therapy:
As the New Year begins and we head off into 2023. You look back to reflect and look forward for things to come. On my radar is the Oropharyngeal Cancer (OPC) trial comparing proton therapy to photon therapy. (REF1)
In simple terms the trial compares traditional radiation to proton therapy in a randomized prospective fashion. It is a high quality trial in a subset of patients that I think is one of, if not the best groups to show a benefit within. The trial objectives are very strong with plenty of outcome metrics to illustrate a strong difference in outcome IF there is one.
It is a non-inferiority trial which, in general, I don’t like. In this case in particular, protons can’t be equal and win - they have to win to win. They are too expensive and too complex relative to really good IMRT treatments if they are only “as good” and maybe in regions around the edge “a little better”. The reference really addresses this process and how they got there far better than I can summarize - it is a large point of that reference article and will be part of my ongoing posts on this topic. That said, to me it is a negative and opens the door to criticism.
But I’ve looked at this population a lot and have had a significant experience treating head and neck cancers. I was fortunate to train at MD Anderson and we often saw 10-15 consults 2 days per week in the afternoon - yep - afternoon clinic 1-4pm or something - say 12 new HN consults - hand written chart notes - we were busy. Under beam, between the 3 attendings we’d have 60 or patients so plenty of experience in training. In my North Little Rock practice, we had 8 radiation oncologist across several cities and me and one other MD treated the majority of our HN cases for many years.
I picked up my practice and moved to OKC due to the volume of HN cases I treated, the toxicity associated with the treatments, and the thinking that proton therapy would be a way for us to reduce toxicity. While in OKC, I’ve seen and treated some cases with uniform scanning unilateral treatments - not a ton but some. To date, I certainly think they perform dramatically better but the data will tell. I could relay anecdotes but after the years of COVID, let’s not.
Realistically, Head and Neck cancer, to me, is the primary option on the current horizon with the ability to meaningfully achieve the large scale victory over photons that protons need.
“The Promise of Protons”
Before we finish that thought, let’s back up and look at our other front running options for delivering upon the “promise of protons”. I coined this phrase a few years ago but I think it works. It summarizes my vision and approach to evaluating data in the proton space. To me, and I fully believe that they are better for some cases, the “promise of protons” is not to use in them in very selected cases - not to cherry pick odd ball cases where there will never be randomized data. To me, the “promise of protons” is in the Bragg peak and in the dosimetry slides. They provide less dose to normal tissues and therefore must end with less toxicity across a broad spectrum of disease sites. It is the center of the marketing campaigns for these facilities. That is bar that has been self-set by this industry and one that needs to be cleared.
And remember, to date, protons have not been able to show this in randomized prospective trials with a large obvious gain. There are many reasons for this - far to many and too complicated to address here, but in the end, the data doesn’t exist today.
Randomized Lung Cancer Trial:
The first really strong attempt was the Lung cancer trial. (REF2) It has some problems from my perspective with the design: 1) it had this odd Bayesian Adaptive Randomization, 2) it also reduced dose from 74Gy to 66Gy for the patient if either approach failed to meet plan objectives - basically a shift that should move the two approaches towards a similar outcome, 3) it is a passive scatter approach - that is what was available within proton therapy at the time - pencil beam scanning (PBS) just wasn’t available and, 4) lung is a site where range effects and range uncertainty are harder to manage than in some other sites.
All that to say that the trial was negative. The primary endpoints of radiation pneumonitis and local failure were both negative with protons actually having a bit more pneumonitis in the trial if anything - 6.5% IMRT vs. 10.5% for Protons. They did an ad hoc analysis looking at time - basically showing the proton group improved more later in the trial, but to summarize, it is a negative trial. You can pick it apart to try and achieve a different answer, but definitely not a big win.
Randomized Esophageal Cancer Trial:
This is the second large randomized prospective trial to look at protons - this time in esophageal cancer. (REF3) It is a Phase IIB trial that randomized patients to pre-operative treatment using either IMRT or PBT (including both PBS and passive scanning « 80% with passive). Dosing was the same for the two arms. Primary outcomes measures were progression free survival and a new metric developed at MD Anderson - total toxicity burden (TTB).
This new metric attempts to capture a level of total toxicity for the patient - rather than focusing on 1 toxicity - let’s say pneumonitis - they included 11 distinct adverse events with weighting based on severity and/or link to treatment and then measured a score over a year. It is, to me, an elegant metric to help pick up toxicity differences but it is complex and, in a way, fails what I call the “promise of protons” where if protons are clearly superior, you should have pretty straightforward standard metrics to illustrate the differences.
Results from this trial showed that the total toxicity burden 2.3 times higher with 7.6 times the post-operative complications in the IMRT arm. Progression free survival was the same at 51% and overall survival was the same at 44.5% so the early toxicity differences didn’t translate into either progression free survival (PFS) or overall survival (OS).
It’s an interesting trial because it did show a significant reduction in toxicity but, in my view of the US practice of radiation oncology, the newly created metric (TTB) along with the lack of any trend in PFS or OS, blunted this trials impact. I think if they had clearly defined a simpler metric like post-operative complications as the primary objective, it would have been received as stronger data, but TTB was the primary objective. I was in the room live for the presentation, and even then, it missed being a “big win”.
The Closest Trial to Achieving the Promise:
Comparative Effectiveness of Proton vs. Photon Therapy
For me, the closest trial is still a trial published 3 years ago. It has inherent problems that are significant. It is retrospective in nature with relatively poorly matched treatment arms. So from the start it is far worse quality data than the above two references, but to me, it address the promise - that protons can and do improve outcomes across a wide spectrum of cases. This trial looked at patients treated with concurrent chemoradiation and the primary end point was 90-day adverse events grade 3 or higher using the very standard CTCAE toxicity scale (REF 4).
The arms were imbalanced. In favor of the IMRT arm for some broad metrics - the patients had less comorbidities and were younger (so big advantages) but “balancing that” was a much greater number of HN patients in the photon arm (a very toxic treatment site). These imbalances speak to the issues inherent in all retrospective studies - those that are acknowledged and the unknown confounders.
But it worked - at least in this retrospective study. Over a 90 day period the hospitalization rate (Gr3 or higher toxicity) was 27.6% vs. 11.5%) with a p<0.001). It was a simple look and a powerful conclusion. It served as the basis for my work towards a Proton Dosimetry Comparison Trial which I will begin to post here in the future. If this trial is correct, this is an easy to run national trial with massive impact potential. It is a huge difference in outcomes in patients treated with concurrent chemoradiation - that is “The Promise of Protons.” (if you read REF 1, I wonder if politics and NRG would let such a trial move forward - but that is a different post for a different moment.)
That was now 3 years ago and little has changed:
Since that publication, there was the esophageal trial as we have discussed and there are more randomized trials on the horizon. But realistically, all has been pretty quiet. Yes COVID, but that shouldn’t stop proton data. This years ASTRO really showed me very little new data. I personally think IMRT in prostate cancer is strengthening significantly with SIB approaches and options for hypofractionation.
COMPPARE: Prostate Cancer Trial
In a similar fashion, we have the COMPPARE trial where I think showing a clear benefit that most radiation oncologist will believe is very, very unlikely. The primary endpoint is bowel urgency EPIC scores and there is an early time frame at just 2 years. That timeline is consistent with toxicity profile and perhaps some difference can be shown on that measure. But at the same time, the cancer control metric is very poor in my view. It is a biochemical failure rate at 3 years with a loose definition of failure with a PSA of 2ng/ml above the nadir. Just for reference, PSAs will often decline for 3-4 yrs after radiation. To use this metric almost guarantees equivalence on the cancer control aspect.
And worse yet, the bar in the treatment of prostate cancer is moving. Rather than uniform dose across the whole gland, we are quickly heading towards boosting the visible lesion as was proven beneficial in the FLAME Randomized Phase III trial (REF 7) which boosted focal lesion showing and improvement in disease free survival with no change in toxicity. Realistically, the FLAME trial results mean that the only potential upside of this trial is a massive reduction in side effects (against an already low risk treatment) that might somehow leave an opportunity to argue for higher possible dosing with protons vs IMRT - but that leads to protons chasing, not leading that path forward and again, not a big win.
I dislike writing negative takes, but realistically, I don’t see an avenue for this to be a positive trial as written. If you know something I’m missing, please comment. But as I see it,
If you have $2, I’d bet the under on the COMPPARE trial.
MRI Linac has prospective data far quicker:
The first MRI Linac in the US started treating patients in 2014 (compared to Loma Linda with protons in 1990) and now, less than a decade later, there is prospective randomized data showing it created a likely benefit in prostate cancer (and I think it is even more useful in the treatment of upper abdomen disease). I don’t think it is a landmark trial that changes the standard of care, but realistically, it is impressive data for MRI Linac in the treatment of prostate cancer - one well done prospective trial is worth a pile of retrospective analysis.
In January 2023, the phase III randomized single center MIRAGE trial was published comparing SBRT with CT guidance vs MRI guidance (REF 8). MRI guidance showed a significant reduction toxicity - both Gr2 acute toxicity and patient reported outcomes. As always, there are caveats. Margins are different (2mm less in the MRI arm so essentially just tighter to prostate avoiding more normal tissue) and there is no cancer outcome data (which is important with margin adjustment) - it is an acute toxicity trial and the differences seem to dissipate by as early as the 3 months. But it deserves credit as well. It took just 156 patients to show a difference. So it appears possible to show improvements due to improvements in technology even in prostate cancer where results for men treated with radiation are, generally, excellent. Kudos on the design - unlike the lung cancer trial and esophageal trial, here the new technology was used attempt to reduce margins. Early proton work seemed to use the technology to push doses and we’ve seemed to move towards equivalent approaches somehow expecting protons to outperform. Here technology was used to reduce margins and, at least for acute toxicity, it appears to have worked.
So where does that leave Proton Therapy as we head into 2023:
So back to our head and neck cancer trial - there have been 440 patients randomized in the treatment of oral pharynx cancer using either protons and photons. The patient numbers are strong and it is THE opportunity for proton therapy to illustrate a real clinical improvement. It is an excellent patient population that really must show a significant improvement. The trial structure is strong and the outcome metrics are robust and should have ample opportunity to illustrate any difference between the two approaches.
In the upcoming weeks, I’ll review some of the prospective and retrospective data, and some of my work showing that I think it should be a positive trial and then, after the prior data discussion, I will broaden the scope a bit and discuss the ramifications of both an obviously positive and/or negative trial result.
Note: There are two other well designed trials happening in Europe - the TORPEdO trial in the UK (REF 9) and the DAHANCA 35 Trial in the Danish Head and Neck Cancer Group (REF 10), but these will likely be reported later as they are still accruing.
REFERENCES:
OPC Trial The Journey from Concept to Activation:
https://www.sciencedirect.com/science/article/pii/S1053429617301170?via%3DihubLung Cancer Randomized Trial:
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6008104/Esophageal Cancer Randomized Trial:
https://ascopubs.org/doi/10.1200/JCO.19.02503Retrospective Comparative Trial in Concurrent Chemoradiation Patients
https://jamanetwork.com/journals/jamaoncology/fullarticle/2757520Prostate Cancer COMPPARE Trial:
https://clinicaltrials.gov/ct2/show/NCT03561220FLAME Trial Focal boost with IMRT to Prostate lesions:
https://pubmed.ncbi.nlm.nih.gov/33471548/MRI MIRAGE Single Institution MRI Linac Trial:
https://jamanetwork.com/journals/jamaoncology/fullarticle/2800541TORPEdO: Phase III UK Oropharyngeal Trial
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC9702982/DAHANCA 35: Two parallel HN Trials from the Danish Group:
https://clinicaltrials.gov/ct2/show/NCT04607694
www.protons101.com, home of the original Protons 101 website.
Content for the Protons101 blog written by Mark Storey MD.Updated 2/7/2023 to reflect changes in RADCOMP expectations.