Protons are blowing Rad Onc's boat out the CMS water

This forum made possible through the generous support of SDN members, donors, and sponsors. Thank you.
Members do not see this ad.
ramsesthenice said:
Typically yes. Device manufacturers usually opt for the 510(K) equivalence pathway and typically get it. If they are truly first to market they may have to pay for studies to get coverage but that is the exceeding minority. Also, most devices are not approved by the FDA but rather cleared. Separate pathways with different bars. Can you tell who's wife is in regulatory? First pharma and now med device 🙂
Click to expand...

I used to work in medical affairs for a RT device manufacturer, and part of my job was regulatory. I guess I will have to use proper terminology next time rather than firing off the cell phone hip.

I always found it curious that condoms and linacs are both 510(k) class 2 devices.
 
scarbrtj said:
The four diagnoses of breast, prostate, lung, and bone mets comprise about 75% of all patients in radiation oncology, and the number one indication is breast. You want to do a study looking for proton benefits in early stage breast? That'd be the equivalent of trying to find gravitational waves in a rad onc clinic... without a LIGO. And what is "better" exactly; what kind of local control or grade 1/2 toxicity deltas cross over from "meh" to "better" at these cost scales. There are randomized trials where if I were to summarize the titles of the trials I'd (correctly) say "IMRT is better for breast cancer." And what'd that get from ASTRO? Stones. Big gnarly staghorn calculi. Or this guy. But protons... let's keep billing for 'em until we can show they're better.
Click to expand...


Alright... early stage breast is a bit of a stretch... N+ is another story.

Lung is certainly up in the air (puns!!)... especially since no one is really doing IMPT optimization that is robust for 4D.

Any and everything GI. (again with robust optimization +monte carlo dose calc).

Prostate is probably a no go, but it could still stand to be tested with modern IMPT

The key is that we need REAL trials. ...and if accrual is too difficult, we need well-done comparisons that allow us to make informed decisions.



It has been said *here (edit) 1,000 times that IO is much more expensive than anything in rad onc... so my point is that we should be trying to find more value in our own back yard
 
RickyScott said:
We had 20 + yrs to show a benefit for protons over IMRT. They are likely worse in common indications like prostate cancer as they can not be delivered without rectal balloons or spacers. Breast they are a total joke.
Click to expand...


So game over? IMRT is the best we are every going to do?

It is important to note...

Passive scatter protons (i.e. almost every treatment from 20-5 years ago) : IMPT with robust optimization and monte-carlo dose calculation

as

3DCRT : IMRT/VMAT
 
Lamount said:
Alright... early stage breast is a bit of a stretch... N+ is another story.

Lung is certainly up in the air (puns!!)... especially since no one is really doing IMPT optimization that is robust for 4D.

Any and everything GI. (again with robust optimization +monte carlo dose calc).

Prostate is probably a no go, but it could still stand to be tested with modern IMPT

The key is that we need REAL trials. ...and if accrual is too difficult, we need well-done comparisons that allow us to make informed decisions.



It has been said 1,000 times that IO is much more expensive than anything in rad onc... so my point is that we should be trying to find more value in our own back yard
Click to expand...
I agree with finding value in our yard but let’s not make our yard smaller unnecessarily (increasing residency positions and limiting indications for RT) before inserting exotic trees that will end up dying.
 
RadOncDoc21 said:
I agree with finding value in our yard but let’s not make our yard smaller unnecessarily (increasing residency positions and limiting indications for RT) before inserting exotic trees that will end up dying.
Click to expand...

Residency expansion is killing the job market

Lack of innovation (not an exotic tree) is killing the field.
 
Neuronix said:
Pharma has to pay for studies to get their products FDA approved. Device manufacturers do not.
Click to expand...

True... but they would sell more proton units if phase III data made protons SOC for a new disease site
 
Neuronix said:
I used to work in medical affairs for a RT device manufacturer, and part of my job was regulatory. I guess I will have to use proper terminology next time rather than firing off the cell phone hip.

I always found it curious that condoms and linacs are both 510(k) class 2 devices.
Click to expand...
Honestly bad condoms are probably at least 10x as bad of a public health problem than a bad linac.
 
Lamount said:
True... but they would sell more proton units if phase III data made protons SOC for a new disease site
Click to expand...
Not likely. Protons, MR linac, all of the shiny toys...most of the realized revenue comes from getting people to your center in the first place. I can’t imagine what a trial would have to show to drive people out in droves to buy a proton unit since they are relatively ubiquitous anyway. And like neuronix said, a bad trial could set them back. Not worth the risk.
 
RickyScott said:
Bad condoms are an issue.
amp.cnn.com

Vietnam police seize 345,000 used condoms that were sold as new | CNN

Police in Vietnam have confiscated an estimated 345,000 used condoms which had been cleaned and resold as new, state media reported.
amp.cnn.com amp.cnn.com
Click to expand...
No condoms are too... March 2020 saw shortages of everything, except rad Onc labor I'm guessing....

www.dailymail.co.uk

People buying up condoms to put them on FINGERS to avoid coronavirus

Thanh Thai, from Sydney, took to the Ryde's District Mum's Facebook page on Wednesday, to share a picture of empty shelves after a store sold out of the latex contraceptives.
www.dailymail.co.uk www.dailymail.co.uk
 
Lamount said:
Residency expansion is killing the job market

Lack of innovation (not an exotic tree) is killing the field.
Click to expand...
Well reasoned and good metaphors (exotic tree = protons... I think that's a metaphor). But just keep in mind there are lines of linkage (read CMS' APM manifesto eg) between the inevitability of APM and the rise of protons. If APM becomes a field killer or job market killer, the exotic tree could bear some blame.
 
Following @scarbtj's way of looking at volume, here's a histogram looking at productivity.

1614133635714.png

This only looks at number of unique beneficiaries for the "management of five fractions" code. So this does not include certain procedures like SBRT, and also because it looks at unique beneficiaries it shouldn't matter whether you are moderately hypofractionated or not. There might be some slight spillover between years, like some patients who begin Dec 2017 would be counted Jan 2018. There are about 4200 RO's in this analysis. It peaks at about 50 (i.e., one new CMS patient per week). The mean value is 70.

If you assume 40% of patients are represented in this dataset (as was quoted earlier, not sure if anyone knows what "truth" is) that implies 175 new starts per year for the average RO.

Eight providers are credited with >300 unique beneficiaries, and about 90 are between 200-300.
 
scarbrtj said:
Well reasoned and good metaphors (exotic tree = protons... I think that's a metaphor). But just keep in mind there are lines of linkage (read CMS' APM manifesto eg) between the inevitability of APM and the rise of protons. If APM becomes a field killer or job market killer, the exotic tree could bear some blame.
Click to expand...
I agree...

Protons seem shady... because they are. The indications are dubious, and that leads people to take advantage, and leaves others to thoughtlessly dismiss the entire idea.

it is admittedly difficult to test... because it’s a hell of a randomization, and it’s hard to find a clean endpoint. The key will be things like the TTB.



A simple indication that should be uncontroversial is: we should use protons when they allow us to meet an important constraint that can’t be met with photons. Fortunately, we have been better clarifying the relationship between dosimetry and toxicity, so these data are becoming increasingly helpful.
 
wandering star said:
Following @scarbtj's way of looking at volume, here's a histogram looking at productivity.

View attachment 331085
This only looks at number of unique beneficiaries for the "management of five fractions" code. So this does not include certain procedures like SBRT, and also because it looks at unique beneficiaries it shouldn't matter whether you are moderately hypofractionated or not. There might be some slight spillover between years, like some patients who begin Dec 2017 would be counted Jan 2018. There are about 4200 RO's in this analysis. It peaks at about 50 (i.e., one new CMS patient per week). The mean value is 70.

If you assume 40% of patients are represented in this dataset (as was quoted earlier, not sure if anyone knows what "truth" is) that implies 175 new starts per year for the average RO.

Eight providers are credited with >300 unique beneficiaries, and about 90 are between 200-300.
Click to expand...
This tracks. Nicely in fact. If we are at about 100 de novo new patients per year per rad onc now, and 33-50% of a rad onc's service are palliative/XRT for cancer survivors/distantly diagnosed (>1y ago), you would arrive at ~150-200 new starts overall on average per rad onc per year. To put another way, it can make sense that for 175 new starts per year 75 would be palliative and/or initially diagnosed greater than 1-2 years prior to starting the XRT.
 
evilbooyaa said:
Depends on if you can do daily adaptive treatments on it or not.
Click to expand...
It takes up to one hour to do the adaptive treatment including MR re-sim, MD contour, and Physics/RTT verification. then actual treatment At best you can get 10 treatments per day and that is if you are insanely busy. Consider how much technical revenue can be generated on a conventional liniac with 20-25 patients per day and you can see the MR Linacs are not worth it.
 
Gfunk6 said:
It takes up to one hour to do the adaptive treatment including MR re-sim, MD contour, and Physics/RTT verification. then actual treatment At best you can get 10 treatments per day and that is if you are insanely busy. Consider how much technical revenue can be generated on a conventional liniac with 20-25 patients per day and you can see the MR Linacs are not worth it.
Click to expand...

In all of my job searches nobody has ever offered me a cent of technical revenue or a clear pathway to technical revenue. Regular SBRT generates me no professional wRVUs for treatment delivery. Replans generate me professional wRVUs, and I'm paid $ per professional wRVU so I'm happy to adapt away.

PS: 60 minutes per adaptive case is optimistic, 60-90 is more realistic.
 
Gfunk6 said:
It takes up to one hour to do the adaptive treatment including MR re-sim, MD contour, and Physics/RTT verification. then actual treatment At best you can get 10 treatments per day and that is if you are insanely busy. Consider how much technical revenue can be generated on a conventional liniac with 20-25 patients per day and you can see the MR Linacs are not worth it.
Click to expand...

One hour? That's less than ideal.

I've heard Ethos can get adapted plans turned around within 15-30 minutes

That being said, if you didn't use the MRI Linac as a workhorse (and had a TrueBeam or something to be your workhorse) then maybe it would end up as a positive.
 
Neuronix said:
PS: 60 minutes per adaptive case is optimistic, 60-90 is more realistic.
Click to expand...
In all fairness, this quote was from Assuta Hostpial in Tel Aviv, Israel. In addition to being quite experienced with their MRI linac, it sounds like they have a fellowship-trained Radiologist in the department to improve the fidelity of adaptive planning. Furthermore, they say that their RTTs are experienced enough to draw OARs on most cases before physician review.

Regarding tech fees, though you may not be receiving the financial benefit, consider that even if you are not receiving the technical fees, someone is. That person will certainly take into account the ROI of an MR Linac which is poor IMO.
 
Shrug believe what you want to believe. I've been hands on with Viewray mridian almost every day for 4 years. Adaptives with breathhold are 60-90 minutes.

Last I heard the institution is happy with its ROI. For what it's good at, it's very good, and we're very happy with it clinically for a number of scenarios. Sure if someone wants to come say "hey come work for us and we'll double your pay to do SBRT with this other machine", I'd strongly consider that request. But alas, here I am stuck with this really cool MRI-linac making the same as I'd make elsewhere without one.

I'm still curious how well Ethos can visualize targets for adaptation. Varian's demos to me I found very unimpressive. We're getting one soon so time will tell!
 
Really interesting discussions re: MRgRT. If I have any questions looks like Neuronix is one of the guys to ask. I have this feeling that many bothans are currently dying attempting to cut these 60-90 minute replanning timeframes down drastically. (Re: Varian, they always start out "unimpressive" then all of a sudden come on strong outta nowhere. Very Microsoft-ish.) I also have the feeling that the world is priming rad onc to get ready to enter the Age of This Is Why We Can't Have Nice Things. Followed rapidly by the Age of No One Can Have Nice Things.
 
I honestly think people have unrealistic expectations. Our non-adaptive breathhold pancreas sbrts on trilogy were booked for an hour just 5-10 years ago. We had the world's largest pancreatic sbrt experience at the time. Add on adaptive and you want that for free or to even cut time? Quality treatments take time.
 
Neuronix said:
Shrug believe what you want to believe. I've been hands on with Viewray mridian almost every day for 4 years. Adaptives with breathhold are 60-90 minutes.
Click to expand...
I have been using a Unity for a couple of years now and my experience is a little better, but not by much. With plan adaption pancreatic SBRT is typically 45 min from door to door. I don't do breath hold or compression and rather use a T2 navigated triggered image for planning and adaption. If the images are subpar and I have to do a lot of contouring it can easily go closer to or over an hour. I don't know about you, but I only adapt the things that are actually important. I don't care 1 bit about the ground truth numbers for structures that are not right by the GTV or likely to be exceeded.

The easiest thing to do with SBRT on an MRI is far and away prostate. Thanks to the mesorectal fat, the background is white on T2 and all the organs of interest are clearly delineated and easily deformed with basic stock algorithms. A prostate SBRT goes something like this:

Patient set up: 5 min
2 min T2 acquisition and reconstruction: 5 min
Deformable structure registration: 2-3 min
Manual contour tweaking: 2-5 min (tops)
Plan optimization: 5 min
Delivery: 10 min (assuming about 1.5 min per beam)

Even with the easiest thing you can possibly do the absolute best case if everything goes right is somewhere around a half an hour. This is with MRI certified therapists and experienced physics support. When you move into areas like the pancreas normal structures get harder to see and there are more of them so the computer will have to work harder to get you a plan. May also need a 7 min scan to get good images. Literally every step might take more time.

If you treat a lot of prostate patients 10 hours a day I guess it might be possible to get in the neighborhood of 20 patients through per day but again this would be a very high end estimate.
 
wandering star said:
Following @scarbtj's way of looking at volume, here's a histogram looking at productivity.

View attachment 331085
This only looks at number of unique beneficiaries for the "management of five fractions" code. So this does not include certain procedures like SBRT, and also because it looks at unique beneficiaries it shouldn't matter whether you are moderately hypofractionated or not. There might be some slight spillover between years, like some patients who begin Dec 2017 would be counted Jan 2018. There are about 4200 RO's in this analysis. It peaks at about 50 (i.e., one new CMS patient per week). The mean value is 70.

If you assume 40% of patients are represented in this dataset (as was quoted earlier, not sure if anyone knows what "truth" is) that implies 175 new starts per year for the average RO.

Eight providers are credited with >300 unique beneficiaries, and about 90 are between 200-300.
Click to expand...
I was looking at this again. And did a double-take.

Do you all realize this guy just said there are 8 radiation oncologists in the U.S. that were responsible for starting >1 weeks of EBRT on >300 CMS patients in 2018. And ~100 responsible for >200 new starts?

In other words, less than 5% of practicing U.S. rad oncs are averaging starting >4 Medicare patients on (>1 week) beam PER WEEK in the US?
 
This seems low (even in today's market). These numbers may not count Medicare Advantage plans (I do not know - just postulating) which account for >1/3 of the Medicare population (with regional variation).
 
Neuronix said:
How big is your ITV/PTV? This is kind of surprising to hear for a pancreas.

I contour 3 cm ring around target as shown by Amsterdam group when using 50 Gy / 5 fractions (my usual dose for pancreas or oligomet).
Click to expand...

3-5 mm depending on what I can get away with.

The reason I don't use breath hold is not because I don't want to. Its because I can't. The gaiting functions on the Unity have not been cleared yet by the FDA and we have not found a practical way to do good gaiting using our existing systems in the tube. What it means is that I can do MR-guided SBRT for patients with minimal motion. The plus side is that generally means body body lesions or those with vascular encasement in which case localization is greatly simplified because it is very easy to align to vessels. It has gone well. I have not had any significant treatment toxicities and outcomes have been good. But I have to be more selective with who I can treat than you.
 
radiaterMike said:
This seems low (even in today's market). These numbers may not count Medicare Advantage plans (I do not know - just postulating) which account for >1/3 of the Medicare population (with regional variation).
Click to expand...
A highly motivated person can obtain this information by digging...

Medicare Advantage/Part D Contract and Enrollment Data | CMS

Important Notice: CMS reloaded the January 2025 files on February 27, 2025. The correct data files should be dated 1/21/2025 or later.The Medicare Advantage (MA) / Part D Contract and Enrollment Data section serves as a centralized repository for publicly available data on contracts and plans...
www.cms.gov www.cms.gov
 
theradiator said:
A highly motivated person can obtain this information by digging...

Medicare Advantage/Part D Contract and Enrollment Data | CMS

Important Notice: CMS reloaded the January 2025 files on February 27, 2025. The correct data files should be dated 1/21/2025 or later.The Medicare Advantage (MA) / Part D Contract and Enrollment Data section serves as a centralized repository for publicly available data on contracts and plans...
www.cms.gov www.cms.gov
Click to expand...
Someone calculates or quotes a number from published data and someone else says "This is just an illusion." This is not the full picture. These numbers are low, because my feelings tell me so. Pay no attention to the man behind the curtain.

I'm tired of all the digging. Every time somebody digs it's like "Whoa these vetted and published numbers, and multiplication/division as mathematical operators, can't be right!" In theory ASTRO should be digging and telling what's what. Good luck with that.
 
ramsesthenice said:
I have been using a Unity for a couple of years now and my experience is a little better, but not by much. With plan adaption pancreatic SBRT is typically 45 min from door to door. I don't do breath hold or compression and rather use a T2 navigated triggered image for planning and adaption. If the images are subpar and I have to do a lot of contouring it can easily go closer to or over an hour. I don't know about you, but I only adapt the things that are actually important. I don't care 1 bit about the ground truth numbers for structures that are not right by the GTV or likely to be exceeded.

The easiest thing to do with SBRT on an MRI is far and away prostate. Thanks to the mesorectal fat, the background is white on T2 and all the organs of interest are clearly delineated and easily deformed with basic stock algorithms. A prostate SBRT goes something like this:

Patient set up: 5 min
2 min T2 acquisition and reconstruction: 5 min
Deformable structure registration: 2-3 min
Manual contour tweaking: 2-5 min (tops)
Plan optimization: 5 min
Delivery: 10 min (assuming about 1.5 min per beam)

Even with the easiest thing you can possibly do the absolute best case if everything goes right is somewhere around a half an hour. This is with MRI certified therapists and experienced physics support. When you move into areas like the pancreas normal structures get harder to see and there are more of them so the computer will have to work harder to get you a plan. May also need a 7 min scan to get good images. Literally every step might take more time.

If you treat a lot of prostate patients 10 hours a day I guess it might be possible to get in the neighborhood of 20 patients through per day but again this would be a very high end estimate.
Click to expand...
Naive question as I don't have a lot of experience with MRI-adaptive planning - but is the adaptation worth it for the tweaks you make considering in those 14-18 mins there could be changes in rectal gas/bladder filling? Purely from a planning and treatment perspective as opposed to a replanning charge perspective.
 
DireMoose said:
Naive question as I don't have a lot of experience with MRI-adaptive planning - but is the adaptation worth it for the tweaks you make considering in those 14-18 mins there could be changes in rectal gas/bladder filling? Purely from a planning and treatment perspective as opposed to a replanning charge perspective.
Click to expand...
Usually MRI guidance allows for re-evaluation of this before treatment delivery. I'm not sure if all allow monitoring during the therapy or not.

Similarly with the Ethos adaptive platform which is CT based for which you can re-CBCT immediately prior to tx, although there is no 'real-time' tracking.
 
DireMoose said:
Naive question as I don't have a lot of experience with MRI-adaptive planning - but is the adaptation worth it for the tweaks you make considering in those 14-18 mins there could be changes in rectal gas/bladder filling? Purely from a planning and treatment perspective as opposed to a replanning charge perspective.
Click to expand...
Excellent question. The answer is that it depends. For small tweaks probably not; especially if you are using multiple fractions. A fair critique of slower treatments like MR linac or CK is that its great you can monitor for intrafraction motion but you are also more likely to experience significant intrafraction changes because the treatments are markedly longer. Once you get comfortable with the machines you can do a pretty good job of selecting patients who it is really useful in and who it may be more of a hassle than its worth.
 
DireMoose said:
Naive question as I don't have a lot of experience with MRI-adaptive planning - but is the adaptation worth it for the tweaks you make considering in those 14-18 mins there could be changes in rectal gas/bladder filling? Purely from a planning and treatment perspective as opposed to a replanning charge perspective.
Click to expand...

For prostate bladder filling and gas bubbles you can see during treatment on the cine images that are running during the radiation.

On Viewray you can adjust the target during treatment if there is a shift. You could pause and adjust the whole treatment if you had to, but I don't think I've ever done that outside of gross patient movement or poor tracking of the target structure.

There was a Cureus paper looking at changes in bowel in the abdomen pre and post treatment. It's very uncommon to have significant shifts in large/small bowel in the abdomen without the patient moving.
 
thecarbonionangle said:

Proton Therapy | Tampa Oncology and Proton

topdocsfl.com topdocsfl.com

Another proton center coming to Florida. This is expected to really fill an immediate need in underserved population. People really don’t want that low dose bath!
Click to expand...
Is it possible to tout protons without employing the hokiest, dosimetrically misleading, medically misleading graph. The graph is not bad if all you use is single beam X-ray treatments I guess.

l1fGGvj.png
 
IAmTheWallnerus said:
Is it possible to tout protons without employing the hokiest, dosimetrically misleading, medically misleading graph. The graph is not bad if all you use is single beam X-ray treatments I guess.

l1fGGvj.png
Click to expand...
Considering there are no values, I’m surprised they didn’t exaggerate it more. Nothing like a unit less graph to display objective data.
 
You must log in or register to reply here.

Similar threads

C
CMS Innovation
Replies
3
Views
712
TheWallnerus
TheWallnerus
uchihastan
Rad Onc vs. Heme/Onc
Replies
16
Views
2K
medgator
M
Gfunk6
Rad Onc? Burnt Out?
Replies
36
Views
4K
thecarbonionangle
thecarbonionangle
B
Rad Onc Procedures and RVUs
Replies
35
Views
5K
CurbYourExpectations
CurbYourExpectations
D
Saving Rad Onc
2
Replies
60
Views
7K
RickyScott
R
Top