SBRT dose constraints & TG101

[Gfunk6]

There has not been an update to SBRT dose constraints in years. It is well understood by experts that some dose constraints (e.g. spinal cord) are well validated and should be respected while others (e.g. great vessels) are essentially made-up and are routinely violated for target coverage.

Personally, I completely ignore the intermediate SBRT dose constraints as they make absolutely no sense to me.

From talking to colleagues, it seems like constraint violation is directly proportional to SBRT experience (e.g. more experience = more confidence in violations). I'm curious where you all stand and if there are newer published constraints that I may have missed.

---

Members don't see this ad.

 

[BobbyHeenan]

I agree I have struggled in this space as well, especially 3-5 fraction regimens.

This is sometimes helpful for three to five fraction regimens.

https://www.practicalradonc.org/article/S1879-8500(21)00105-3/abstract?dgcid=raven_jbs_etoc_email

UK consensus is another one, though at times I have viewed it as a bit too conservative.

UK Consensus on Normal Tissue Dose Constraints for Stereotactic Radiotherapy - PubMed

Six UK studies investigating stereotactic ablative radiotherapy (SABR) are currently open. Many of these involve the treatment of oligometastatic disease at different locations in the body. Members of all the trial management groups collaborated to generate a consensus document on appropriate...

pubmed.ncbi.nlm.nih.gov

 

[reddonkey]

Agree with the prior sentiments. I find the below publication of Timmerman's tables from a few years ago helpful as a conservative guide (and an interesting read about how all of these were sort of made up and became a reference):

https://www.redjournal.org/article/S0360-3016(21)02831-5/fulltext

 

[RadOncMegatron]

Agreed. I believe Timmerman also said a lot of this stuff is made up.

 

[Gfunk6]

Agree with the prior sentiments. I find the below publication of Timmerman's tables from a few years ago helpful as a conservative guide:

https://www.redjournal.org/article/S0360-3016(21)02831-5/fulltext

This is extremely interesting, thanks for sharing the link.

 

[radiation]

There has not been an update to SBRT dose constraints in years. It is well understood by experts that some dose constraints (e.g. spinal cord) are well validated and should be respected while others (e.g. great vessels) are essentially made-up and are routinely violated for target coverage.

Personally, I completely ignore the intermediate SBRT dose constraints as they make absolutely no sense to me.

From talking to colleagues, it seems like constraint violation is directly proportional to SBRT experience (e.g. more experience = more confidence in violations). I'm curious where you all stand and if there are newer published constraints that I may have missed.

https://www.redjournal.org/issue/S0360-3016(20)X0015-0

There was an update to quantec with more modern data for hypofractionated data called HyTEC

 

[Palex80]

If you can't meet constraints, just fractionate.

Oh... I forgot... In the US "stereotactic RT" is defined by number of fractions, not treatment technique.

 

[TheWallnerus]

There has not been an update to SBRT dose constraints in years. It is well understood by experts that some dose constraints (e.g. spinal cord) are well validated and should be respected while others (e.g. great vessels) are essentially made-up and are routinely violated for target coverage.

Personally, I completely ignore the intermediate SBRT dose constraints as they make absolutely no sense to me.

From talking to colleagues, it seems like constraint violation is directly proportional to SBRT experience (e.g. more experience = more confidence in violations). I'm curious where you all stand and if there are newer published constraints that I may have missed.

“Learn the rules like a pro so you can break them like an artist.”

- Pablo Picasso

 

[Neuronix]

I came also to post HyTEC. There are some issues with HyTEC though, such as the data simply doesn't exist for some of these organs at risk, so they don't report on them.

So much of TG-101 is made up. It's lifted straight from Timmerman's article in seminar in radiation oncology (https://www.sciencedirect.com/science/article/pii/S1053429608000301?via=ihub), and he has been very clear that there is varying levels of evidence. One time I compared TG-101 and Timmerman tables directly and there was one minor difference that I'm convinced is a typo.

Anyway, I tell the residents all the time that the transition from junior resident to senior resident and attending is knowing the evidence behind the constraints and the risk if you violate them. Being disease site specific means knowing you're going to break the rules, what the risk is when you do, and feeling comfortable to do it anyway.

My favorite TG-101 constraints are some of the 5 fraction volume constraints. Since when has 14.5 Gy in 5 fractions hurt a spinal cord, 16.5 Gy in 5 fractions hurt a trachea, or 12.5 Gy in 5 fractions hurt a duodenum? This is what happens when a bunch of physicists codify a review article...

 

[Gfunk6]

@Neuronix

What do you use as a constraint for 5 fractions spine with regard to the cord? TG101 says 30 Gy, HyTec says 28.8 Gy.

 

[Neuronix]

It's a little bit of devil is in the details. I do a high resolution MRI in treatment position, fuse it myself at the treatment level(s), and then define the cord proper without PRV. I then treat with 6D couch or with adaptive MRI-RT to ensure position. The numbers I use may not make sense if you're not using these same steps to define cord.

 

[Gfunk6]

It's a little bit of devil is in the details. I do a high resolution MRI in treatment position, fuse it myself at the treatment level(s), and then define the cord proper without PRV. I then treat with 6D couch or with adaptive MRI-RT to ensure position. The numbers I use may not make sense if you're not using these same steps to define cord.

I actually do the same - MRI fusion and treatment with 6D couch and spine tracking on CK. I generally add a mm or two on contoured cord to create a PRV. I limit V30 < 0.01 cc. CK ususal imaging interval is 60 seconds but I default to 20 seconds if lesions is within close proximity to the cord or for high-risk re-treatments.

 

[Neuronix]

Good. 30 Gy<0.03 cc off protocol. We have a protocol that goes up to 32 Gy.

 

[ramsesthenice]

I came also to post HyTEC. There are some issues with HyTEC though, such as the data simply doesn't exist for some of these organs at risk, so they don't report on them.

So much of TG-101 is made up. It's lifted straight from Timmerman's article in seminar in radiation oncology (https://www.sciencedirect.com/science/article/pii/S1053429608000301?via=ihub), and he has been very clear that there is varying levels of evidence. One time I compared TG-101 and Timmerman tables directly and there was one minor difference that I'm convinced is a typo.

Anyway, I tell the residents all the time that the transition from junior resident to senior resident and attending is knowing the evidence behind the constraints and the risk if you violate them. Being disease site specific means knowing you're going to break the rules, what the risk is when you do, and feeling comfortable to do it anyway.

My favorite TG-101 constraints are some of the 5 fraction volume constraints. Since when has 14.5 Gy in 5 fractions hurt a spinal cord, 16.5 Gy in 5 fractions hurt a trachea, or 12.5 Gy in 5 fractions hurt a duodenum? This is what happens when a bunch of physicists codify a review article...

The 5 fraction are my favorite. How have we done so much 20/5 without killing people LOL. I definitely use it as a training too with respect to when guidelines have literally no meaning. Everyone is different but I certainly feel like a few of my colleagues under treat to meet arbitrary constraints.

Slightly related, anyone else see Medtronic has added a 10 cGy/min limit to most of their pacers that they say is completely irrespective of dose? No one can show me any actual data and this feels like a solution looking for a problem to me. But some are respecting it like god and I just reviewed a 23 min delivery for an 8Gy/fx bone met SBRT plan with a mean dose of <1 Gy to the device. Sigh.

 

[evilbooyaa]

Timmerman's numbers are all made up. The volumetric constraints for spinal cord are stupid. I only care about point max, defined as 0.035cc.

If it wouldn't be horrible to the plan, I do favor just taking the entire cord on MRI (or on CT if post-op SBRT) to a point dose < 30Gy in 5fx. If there is significant epidural disease that needs those higher doses after say a separation surgery, then a CT myelogram may be more useful given difficulties with artifact on post-instrumentation MRIs.

 

[IonsAreOurFuture]

I came also to post HyTEC. There are some issues with HyTEC though, such as the data simply doesn't exist for some of these organs at risk, so they don't report on them.

So much of TG-101 is made up. It's lifted straight from Timmerman's article in seminar in radiation oncology (https://www.sciencedirect.com/science/article/pii/S1053429608000301?via=ihub), and he has been very clear that there is varying levels of evidence. One time I compared TG-101 and Timmerman tables directly and there was one minor difference that I'm convinced is a typo.

Anyway, I tell the residents all the time that the transition from junior resident to senior resident and attending is knowing the evidence behind the constraints and the risk if you violate them. Being disease site specific means knowing you're going to break the rules, what the risk is when you do, and feeling comfortable to do it anyway.

My favorite TG-101 constraints are some of the 5 fraction volume constraints. Since when has 14.5 Gy in 5 fractions hurt a spinal cord, 16.5 Gy in 5 fractions hurt a trachea, or 12.5 Gy in 5 fractions hurt a duodenum? This is what happens when a bunch of physicists codify a review article...

I agree, almost all structures can tolerate 20 Gy in 5 fractions to a point dose, if not the whole thing, so I typically had considered most constraints smaller than that to be "made up."

There are exceptions for larger volumes, backed by clinical data, like small bowel loops must be kept under 15 Gy to 120cc, which is part of the original QUANTEC analysis of patient outcomes.

Hippocampus starts measurably failing at 7.4 Gy to 40% of the volume.

Many constraints are getting extremely low, like 1 gray or less mean heart dose, 3 Gy or less mean LAD dose, and the volume of liver getting zero dose or less than 1 gray. My training originally consisted of Emami tables, for 1/3, 2/3, and 100% of an organ being treated. I sort of miss the simplicity, but at least we now have Clear Check and other ways to automate the DVH analysis.

 

[BobbyHeenan]

Many constraints are getting extremely low, like 1 gray or less mean heart dose, 3 Gy or less mean LAD dose, and the volume of liver getting zero dose or less than 1 gray. My training originally consisted of Emami tables, for 1/3, 2/3, and 100% of an organ being treated. I sort of miss the simplicity, but at least we now have Clear Check and other ways to automate the DVH analysis.

I had noted that as well.

With that said (I may be wrong here, I haven't reviewed updated data), the original NEJM Darby paper (best data is in the supplements) showed basically minimal absolute increase in cardiac events for a mean heart dose of 1 to 3 Gy. So where is this coming from?

I'd imagine a month of ozempic would take that number down too for most of my patients.

 

[evilbooyaa]

I had noted that as well.

With that said (I may be wrong here, I haven't reviewed updated data), the original NEJM Darby paper (best data is in the supplements) showed basically minimal absolute increase in cardiac events for a mean heart dose of 1 to 3 Gy. So where is this coming from?

I'd imagine a month of ozempic would take that number down too for most of my patients.

View attachment 385594

Proton advocates. That's who its coming from.

Nevermind that proton dosimetry on a computer screen for the past 50 years is not what the patient has actually been receiving with fixed RBE assumptions.

NOW, the supporters say, the protons are better with variable RBE calculation.

At least until there's that next unknown unknown that shows that, again, what the computer screen shows is not what the patient is getting.

 

[Palex80]

There are exceptions for larger volumes, backed by clinical data, like small bowel loops must be kept under 15 Gy to 120cc, which is part of the original QUANTEC analysis of patient outcomes.

Really?
15 Gy in how many fractions for 120cc?

Thinking of how we used to treat the entire pelvis or large chunks of the lumbar spine with ap/pa open fields on the simulator with 30 Gy in 10 fractions for bone mets, I imagine that constraint was seldom met.

 

[BobbyHeenan]

Proton advocates. That's who its coming from.

Nevermind that proton dosimetry on a computer screen for the past 50 years is not what the patient has actually been receiving with fixed RBE assumptions.

NOW, the supporters say, the protons are better with variable RBE calculation.

At least until there's that next unknown unknown that shows that, again, what the computer screen shows is not what the patient is getting.

Gotcha.

I'm not sure the cost and potential skin/rib/cosmetic (especially with expanders/reconstruction...have seen bad anecdotes and Harvard series suggest signal there) detriments would outweigh the photon risk for a drop in mean heart from say 3 to 1 Gy.

With good DIBH whole breast I don't typically see mean heart doses above 3 Gy anyway...usually 1ish with photons.

 

[evilbooyaa]

I agree, almost all structures can tolerate 20 Gy in 5 fractions to a point dose, if not the whole thing, so I typically had considered most constraints smaller than that to be "made up."

There are exceptions for larger volumes, backed by clinical data, like small bowel loops must be kept under 15 Gy to 120cc, which is part of the original QUANTEC analysis of patient outcomes.

Hippocampus starts measurably failing at 7.4 Gy to 40% of the volume.

Many constraints are getting extremely low, like 1 gray or less mean heart dose, 3 Gy or less mean LAD dose, and the volume of liver getting zero dose or less than 1 gray. My training originally consisted of Emami tables, for 1/3, 2/3, and 100% of an organ being treated. I sort of miss the simplicity, but at least we now have Clear Check and other ways to automate the DVH analysis.

Citations needed for all of the bolded

 

[Palex80]

...The volume of liver getting zero dose or less than 1 gray...

1 Gy to the whole liver is equivalent to how many Martinis?

 

[MidwestRadOnc]

Citations needed for all of the bolded

I am also wondering if I missed something in training and got through boards with sheer luck.

 

[Ray D. Ayshun]

I always figured some of the absurd sbrt constraints were taking into account rapid dose fall-off and akin to a PRV.

 

[BobbyHeenan]

1 Gy to the whole liver is equivalent to how many Martinis?

I was about to say I think I've had weekends worse on the liver than 1 Gy.