SRS Planning Question

[RadOncMegatron]

This is for LINAC bases stereotactic treatments.

When I was in residency I was drilled on normal brain volume V12 < 10 CC, CI, GI, Paddick Index, and other weird indexes...

When I look at newer protocols like NRG-BN013 (unless I missed something)

There is no brain V12 constraint and they only look at CI. This is what they say:

Dose Conformity:
The ratio of the prescription isodose volume to the target volume should be between 1.0 and 2.0. Preferably, the conformity index should be between 1 and 1.2. It is understood that this ratio may be difficult to achieve with some very small lesions. For lesions less than 5 mm in size, a ratio up to 3.0 is preferred but up to 5.0 is acceptable

I feel BN013 is right. If you have a tight CI and meet the other planning paramters then GI and other weird indexes do not really matter. I also like how they allow up to 5.0 for small lesions. Although, I still use the V12, maybe because it was drilled into me so hard and is easy to meet.

Does anyone else use the "barebones" constraints used in BN013 for LINAC based SRS?

---

Members don't see this ad.

 

[Palex80]

Never used them.

 

[Bequerel]

I went to PRO lecture and asked a couple of the speakers afterwards about constraints. I was confused about whether or not you subtract tumor volume when calculating V20 (3fx) v24 (5 fx) <20cc. I got conflicting answers about how to calculate and sometimes they can’t meet constraints. My basic takeaway was “keep it tight”.

 

[Palex80]

I went to PRO lecture and asked a couple of the speakers afterwards about constraints. I was confused about whether or not you subtract tumor volume when calculating V20 (3fx) v24 (5 fx) <20cc. I got conflicting answers about how to calculate and sometimes they can’t meet constraints. My basic takeaway was “keep it tight”.

I always subtract GTV when evaluating V10 / V12 / V18.

However, this was not done in uniform in many of the studies that evaluated the risk of radiation necrosis.
Have a look at this article:

Dose-Response Effect and Dose-Toxicity in Stereotactic Radiotherapy for Brain Metastases: A Review - PMC

Brain metastases are one of the most frequent complications for cancer patients. Stereotactic radiosurgery is considered a cornerstone treatment for patients with limited brain metastases and the ideal dose and fractionation schedule still remain ...

pmc.ncbi.nlm.nih.gov

When you look at Table 2 at the column "Definition of Predictive Factor", you will see that some authors defined that factor as Brain+Target, while others opted from Brain-Target, one even used Brain-PTV (!).

 

[Neuronix]

I feel BN013 is right. If you have a tight CI and meet the other planning paramters then GI and other weird indexes do not really matter. I also like how they allow up to 5.0 for small lesions. Although, I still use the V12, maybe because it was drilled into me so hard and is easy to meet.

Does anyone else use the "barebones" constraints used in BN013 for LINAC based SRS?

I think that the trials are just trying to be inclusive to increase accrual. I don't think that it should be taken to mean that CI up to 2.0 and infinite GI are acceptable in all cases when delivering quality radiosurgery.

Single fraction and three fraction constraints cannot be used in a randomized trial like this. There's a size based intervention assignment.

What I use for single fraction is V12Gy of brain minus GTV or just select for fractionated after a size threshold of something like 2.5 cm - 3.0 cm. If I can't meet the V12Gy, then I fractionate. Why 3 fraction vs. 5 fraction is up to interpretation. I personally just like to single or five fraction. Other things I've also considered are 5 fractions over 30 cc and 3 fractions below 30 cc (per Alliance A071801). You could also use V20 (3fx)<20cc and if not met go up to 5 fractions.

Once you go above a 5 fraction constraint, there's no consensus as to what to do. I just ignore the 5 fraction constraint and still treat 30 Gy / 5 fractions in such a case. 25 Gy / 5 fractions is too wimpy for most malignant situations, and what other choice do you have? Maybe some lengthier or other fractionation or WBRT are also acceptable options.

 

[Palex80]

I think that the trials are just trying to be inclusive to increase accrual.

DEI for dosimetrists!

Sorry, I had to troll!

 

[Neuronix]

DEI for dosimetrists!

I loled.

Still, there is a legitimate concern about how to handle low resource environments that don't have the equipment necessary to perform SRS to the same standards as high quality centers. Certainly a trial like BN013 is probably including them, assuming that some of these centers are participating in clinical trials. Looking at quality metrics vs. outcomes post-hoc can make for interesting analyses as well.

 

[Palex80]

I loled.

Still, there is a legitimate concern about how to handle low resource environments that don't have the equipment necessary to perform SRS to the same standards as high quality centers. Certainly a trial like BN013 is probably including them, assuming that some of these centers are participating in clinical trials. Looking at quality metrics vs. outcomes post-hoc can make for interesting analyses as well.

 

[Neuronix]

View attachment 394256

The worst is when Paddick himself calls you.

Yes, this actually happened to me.

 

[evilbooyaa]

I care about V12 for single fx. If it's a small lesion I want it < 3cc. Medium lesion, < 5cc. 'Large' lesion, < 10cc. If > 10cc and is otherwise not a crap plan, it needs fractionated. So, the issue likely boils down to, if you allow V12 < 10cc you're going to get people planning **** SRS for like a 5mm lesion. Needs to have some thought into it. And then it makes DCA look better than IMRS even though IMRS can at times be a better way to spare brain tissue by allowing higher dose into the non-brain structures a tumor might be abutting/close to.

There is no reason to prioritize conformality over an actual legitimate DVH besides dogma. A lesion along the interior aspect of the skull can have less conformity but allow better brain sparing than just treating a concentric circle.

 

[emt409]

V12 is a useful surrogate for a single lesion risk, but it really falls apart quickly for multi-mets. IMO, it's really a collinear surrogate for metrics that actually biologically plausibly drive necrosis risk like V18-21.

If you get an SRS certification, they will check your plans for all those other metrics, which are important surrogates of general quality. On an individual patient basis, does the R50 or paddick GI matter that much? No, but if you do 500 radiosurgeries a year, the quality of your plans will manifest in your outcomes.

 

[Palex80]

V12 is a useful surrogate for a single lesion risk, but it really falls apart quickly for multi-mets.

I usually evaluate each V12 for each metastasis alone.

This is tricky if you are using one plan for multiple metastases, but still doable.
If you are using individual plans/isocenters for each metastasis, this is easily doable.

 

[Neuronix]

the quality of your plans will manifest in your outcomes.

Reference?

 

[medgator]

Reference?

Another version of "protons are better, just look at this plan on this screen"

 

[BobbyHeenan]

Someone correct me if I'm wrong....

V12 was originally identified as important for AVM gamma knife necrosis risk. Then the same metric (just V12, did NOT subtract GTV) was verified for brain mets as important.

You need to eval the V12 around each lesion, not for the whole brain/plan.

Irradiated volume as a predictor of brain radionecrosis after linear accelerator stereotactic radiosurgery - PubMed

Analysis of patient and treatment variables revealed V8 Gy-V16 Gy to be the best predictors for RN using linear accelerator-based single-fraction SRS for brain metastases. We propose that patients with V10 Gy >10.5 cm(3) or V12 Gy >7.9 cm(3) be considered for hypofractionated rather than singl …

pubmed.ncbi.nlm.nih.gov

It all comes down to it's just a surrogate for size. I'm sure there are papers/groups out there that use V12 minus GTV as well.

I have no strict cut off but I start getting a little bit nervous when V12 gets around 6 cc. With that said, not all V12's are the same, and as someone mentioned a peripheral GTV V12 may have a good amount of dose in skull or skin. I start thinking about fractionation though at 2 cm/V12 > 6 cc.

The italian group (not randomized, but single vs 27 Gy in 3) suggested V18 (they defined as V18 minus GTV) as a metric for a 3 fraction regimen to look out for.

 

[evilbooyaa]

I usually evaluate each V12 for each metastasis alone.

This is tricky if you are using one plan for multiple metastases, but still doable.
If you are using individual plans/isocenters for each metastasis, this is easily doable.

Why do you consider it tricky if using one plan for multiple metastases?

Brain - GTVAll - V12 < 5cc x # of lesions?

 

[BobbyHeenan]

Why do you consider it tricky if using one plan for multiple metastases?

Brain - GTVAll - V12 < 5cc x # of lesions?

I like to just look at each lesion, but to do that you often have to create a sub volume of just normal brain around that lesion (and doesn’t include the V12 of some far off lesion).

in your formula above you may get a V12 minus brain that is 10 cc but if you’re treating 6 mets that’s not too bad.

It’s all about the V12 at each lesion, not as a sum (or at least that’s how the papers did it).

Edit - never mind I misunderstood. Your formula would be a quick way to look at it. Not as “detailed” as looking at sub volumes around each lesion but I get it.

 

[Palex80]

Why do you consider it tricky if using one plan for multiple metastases?

Brain - GTVAll - V12 < 5cc x # of lesions?

Sorry, I may sound stupid, but I do not really get the equation... Does that actually tell me which metastasis may be causing an issue?

What I normally do is:
I create a Brain volume for each of the metastases by adding a margin to the GTV of each metastases and the subtract the GTV of the corresponding metastasis.
So, if I am treating 3 mets I end up with
Brain1 = (GTV Brain Met 1 + 3 cm margin [cropped outside brain]) - GTV Brain Met 1
Brain2 = (GTV Brain Met 2 + 3 cm margin [cropped outside brain]) - GTV Brain Met 2
Brain3 = (GTV Brain Met 3 + 3 cm margin [cropped outside brain]) - GTV Brain Met 3

I then evaluate the V12 / V10 for each of these Brain structures. So, I am able to tell which brain met may be problematic.

I mean it's "tricky" in the sense that it is cumbersome to manually do that.

 

[emt409]

Someone correct me if I'm wrong....

V12 was originally identified as important for AVM gamma knife necrosis risk. Then the same metric (just V12, did NOT subtract GTV) was verified for brain mets as important.

You need to eval the V12 around each lesion, not for the whole brain/plan.

Irradiated volume as a predictor of brain radionecrosis after linear accelerator stereotactic radiosurgery - PubMed

Analysis of patient and treatment variables revealed V8 Gy-V16 Gy to be the best predictors for RN using linear accelerator-based single-fraction SRS for brain metastases. We propose that patients with V10 Gy >10.5 cm(3) or V12 Gy >7.9 cm(3) be considered for hypofractionated rather than singl …

pubmed.ncbi.nlm.nih.gov

It all comes down to it's just a surrogate for size. I'm sure there are papers/groups out there that use V12 minus GTV as well.

I have no strict cut off but I start getting a little bit nervous when V12 gets around 6 cc. With that said, not all V12's are the same, and as someone mentioned a peripheral GTV V12 may have a good amount of dose in skull or skin. I start thinking about fractionation though at 2 cm/V12 > 6 cc.

The italian group (not randomized, but single vs 27 Gy in 3) suggested V18 (they defined as V18 minus GTV) as a metric for a 3 fraction regimen to look out for.

Correct

 

[evilbooyaa]

Sorry, I may sound stupid, but I do not really get the equation... Does that actually tell me which metastasis may be causing an issue?

What I normally do is:
I create a Brain volume for each of the metastases by adding a margin to the GTV of each metastases and the subtract the GTV of the corresponding metastasis.
So, if I am treating 3 mets I end up with
Brain1 = (GTV Brain Met 1 + 3 cm margin [cropped outside brain]) - GTV Brain Met 1
Brain2 = (GTV Brain Met 2 + 3 cm margin [cropped outside brain]) - GTV Brain Met 2
Brain3 = (GTV Brain Met 3 + 3 cm margin [cropped outside brain]) - GTV Brain Met 3

I then evaluate the V12 / V10 for each of these Brain structures. So, I am able to tell which brain met may be problematic.

I mean it's "tricky" in the sense that it is cumbersome to manually do that.

Sure, I suppose if there is a huge difference in size then can be an issue. But for most 5-10mm mets, there is unlikely to be 'one' that is driving a difference. I don't love doing single iso multi lesion all across the brain with single fraction for this exact reason.