Repeat SRS to brainstem?

This forum made possible through the generous support of SDN members, donors, and sponsors. Thank you.
R

RadOnc2013

Full Member
10+ Year Member
Joined
Nov 10, 2011
Messages
40
Reaction score
3
A patient has a hx of early-stage NSCLC s/p lobectomy and chemo, and currently is NED elsewhere in the body. She has a solitary brain met originally measuring less than 1 cm and involving the thalamus extending into the brainstem, treated with SRS to 14 Gy. The lesion initially shrank but now has grown again and measures nearly 2 cm. She has no other sites of disease in her body. She is awake and alert, but has worsening neurologic symptoms (weakness and numbness on her left side mainly - she needs assistance with ambulation and this is worsening - she also just started developing double vision). She can't take steroids - she nearly died the last time she took them due to an adverse reaction. She is not a candidate for any other local therapy. The Neuro-Radiologist is certain that this is failure of initial SRS and progression of disease, this is NOT radiation necrosis. She is not a candidate for any other systemic therapy. Would you offer re-SRS? If so, how would you do it?

I just saw this patient today. Maybe "SRS" of 8 Gy x 1 as palliation to try to stop the worsening of the patient's neurologic symptoms? I highly doubt I can cure her. I also consented her for the chance of death from treatment-related complications.
Members don't see this ad.
 
You haven't given us any chronology.
 
Her lobectomy and chemo were in 2015.
Prior SRS was in 2016.
She had some of the same neuro symptoms before the prior SRS, but they have been getting worse since then.
Anything else?
 
Members don't see this ad :)
If this is her only site of disease, why not fractionate to tumor + margin of 40-60 Gy at 2 Gy a fraction? We know she can tolerate at least 40 Gy/2GY as this is a whole brain dose and SRS historically was combined with whole brain. I would think this is safer and offer a much higher bed, probably a lot more than the original 14Gy.
 
Last edited:
  • Like
Reactions: 1 users
I'd probably forgive 7 Gy of that 14 Gy. Thus, it becomes a BED exercise where I'd allow an accumulated BEDGy3 equiv to 54/30. It's all a bit guesswork isn't it. But I've never really had problems (and your patient already has problems). E.g.,

7/1 = 7*(1+7/3) = 23 BED Gy3
54/30 = 54*(1+1.8/3) = 86.4 BED Gy3

So you have about 60 Gy3 (86.4 minus 23) to work with. Assume 5 Gy fractions.

60 = 5*x*(1+5/3) = x equals 4.5 fractions. So give 4-5 fractions of 5 Gy. I like 5Gy per day vs something less per day just because of the clinical situation, not prolonging tx unnecessarily, etc.
 
RadOnc2013 said:
Her lobectomy and chemo were in 2015.
Prior SRS was in 2016.
She had some of the same neuro symptoms before the prior SRS, but they have been getting worse since then.
Anything else?
Click to expand...

Also, consider Keytruda if not done already (unlikely)
 
Consider a possibility of pseudoprogression/"pseudoradionecrosis" if the patient has already taken a checkpoint inhibitor. I've elected to do nothing on a recent patient like this, and the lesion got better.
 
I agree with scarbrtj. You can probably retreat her with 5 x 5 Gy.
I would have probably given her 8 x 3.5 Gy or 7 x 4 Gy, but that's just my gut feeling.

I would NOT retreat with 1 x 8 Gy. You will probably end up with the a worse situation in 3-4 months when she reprogresses and you then have very little options left.

If there's doubt about true progression do consider getting some other imaging. Amino-acid PET is fancy and can help. Although given the fact that this lady got "only" 14 Gy SRS, I presume that it's a local failure. Failure rate after SRS with such a dose is high. It's one of the reasons I dislike doing single-shot SRS in the brainstem and rather opt for fractionated treatments.
 
Thank you guys. This is really very helpful. I will probably be posting other questions in the future, as I am in a solo practice and I want to make sure I am doing the best that I can for the patients I see.
 
RadOnc2013 said:
Thank you guys. This is really very helpful. I will probably be posting other questions in the future, as I am in a solo practice and I want to make sure I am doing the best that I can for the patients I see.
Click to expand...
Can we charge you? Just kidding...

One small (not very ethical) point: Even if your patient with the repeat SRS/FSRT develops a necrosis after retreatment, dies and someone from the family decides to sue you: Since with the dose you are left, you are no longer able to durably cure this tumor, the chance that someone will be able to actually prove that she died from a radionecrosis and not from the recurrent tumor is slim. There is still going to be vital tumor there and it's tough to tell if that or the necrosis caused her death.
 
Hm have a citation for when/how to give the avastin? Her neuro symptoms are rapidly getting worse so I am telling my physicist to plan really fast ha.
 
I am curious to see what others think, but as an aside, some locations within the brain stem are more critical than others. I have seen docs freak out because a cerebellar lesion is next to posterior tract, which carries proprioception? This sounds like it is midbrain? At a minimum, 300 x 10 or 3750/15 is safe - historically we didn’t count the stereo dose to whole brain, maybe discount 2 gy.
 
Members don't see this ad :)
RadOnc2013 said:
Hm have a citation for when/how to give the avastin? Her neuro symptoms are rapidly getting worse so I am telling my physicist to plan really fast ha.
Click to expand...
Here's one source, though there's many more depending on context, but it gives an idea of dosing and time to response, which can be pretty rapid.
Randomized Double-Blind Placebo-Controlled Trial of Bevacizumab Therapy for Radiation Necrosis of the Central Nervous System - ScienceDirect
Obviously no NSCLC patients, but I've heard "it's definitely progression" a lot, and when biopsy was done, it was RN. That was typically in the setting of single fractions of 18 Gy or more. This offers a non-steroidal therapy that might help while you do the planning process, etc, and if she improves clinically, might keep you from retreating the brainstem.
 
I highly doubt 14 Gy to a less than 1 cm met caused necrosis, and it very likely did not control the disease.
 
  • Like
Reactions: 1 user
Agree with Ray D. Ayshun and others. We would treat this lady with 25Gy/5fx, and strongly consider bevacizumab. If her platelets and kidneys are ok, it is reasonable to consider high dose COX2 inhibitor in lieu of steroid.
 
Ray D. Ayshun said:
I don't disagree, but
Accelerator-based stereotactic radiosurgery for brainstem metastases. - PubMed - NCBI
12% local failure at 1yr and 5% complications at 2. I'm wondering if avastin might help now or help prevent rn in the future if this is retreated.
Click to expand...
I also agree with Avastin since it is active in lung cancer (recent data from ASCO actually show synergy with immunotherapy) When I see results like 12% failure at one year the first thought that goes through my mind - vast majority with a brainstem met who are alive one year later probably are alive because they derived a really significant benefit from a systemic therapy, which in turn gave local control benefit. 14 Gy in one fraction is not enough radiation on its own to give 90% local control.
 
After a lit review on first line SRS for brainstem mets, I came to the conclusion a few years ago that 15 Gy in 1 fraction is the right dose for brainstem mets up to 2 cm in size. Here's a good paper on the topic:

Gamma knife radiosurgery for brainstem metastases: the UCSF experience. - PubMed - NCBI

Any less on dose and the mets just fail on you. You might be able to do more (maybe 16 Gy or more for tiny mets), but that's definitely pushing it. I remember seeing that one of the Alliance trials also had 15 Gy in 1 fraction for brainstem mets and also the Gammaknife textbook by Sheehan and Lunsford also recommends 15 Gy in 1 fraction in the primary setting (both say up to 2 cm met).

Avastin is being used a lot for palliation of brain mets and radiation necrosis with limited data. Hard to get a lot of insurers to pay for it. Also patients often have contraindications. But it's still a reasonable thought as something to try.

In the absence of much data for brainstem, I just extrapolate spine data to the brainstem. There is a lot of data on re-irradiation of spinal cord.

For example:

Proposal of human spinal cord reirradiation dose based on collection of data from 40 patients. - PubMed - NCBI

These guys recommend up to BED 150 Gy2. 14 Gy in 1 fraction is 112 Gy2. An additional 20 in 10 fractions would be BED 40Gy2 so you'd be at 152Gy2 if you did that. You could argue to reduce to 18 Gy in 10 fractions. I'm going to ignore hot spot considerations in this setting. There are many other papers in spinal cord re-irradiation with BEDs on the range of 125-135 being "safe" at least 6 months after primary radiation. Once you get over a combined BED 135Gy2 you're always going to be risking toxicity in the cord or brainstem. There's still a debate how much risk there is over 135Gy2 and the mitigating effects of increasing time interval between primary and secondary radiotherapy.

This also jives with the palliative data for re-irradiation of brainstem DIPG. There are a number of papers on this topic as well, such as: Survival benefit for patients with diffuse intrinsic pontine glioma (DIPG) undergoing re-irradiation at first progression: A matched-cohort analysis on behalf of the SIOP-E-HGG/DIPG working group - ScienceDirect. In that setting they've typically given 54 Gy in 1.8 Gy/fx to a BED of 102.6Gy2 in the first line. They then follow up with 10 fractions to 18 Gy up to 30 Gy. These are the same ballpark BEDs as this 14 Gy in 1 fraction, ignoring any radiosurgery effect not accounted for in the linear-quadratic model.

So what I would do if I wanted to reirradiate in this setting is 20 Gy in 10 fx IMRT/IGRT and cross my fingers. Some people like 1.2 Gy BID fractionations. I find the data unconvincing, the patient inconvenience high, and the BEDs difficult to calculate, so I don't do it. I personally think 5 Gy x 5 fractions is too aggressive here to the brainstem on top of a prior 14 Gy in 1 fraction. I'd do it elsewhere in the brain potentially and consent for RT necrosis, but not in the brainstem.
 
Radiosurgery and whole brain therapy in the treatment of brainstem metastases:
https://s3.amazonaws.com/academia.edu.documents/44965068/Radiosurgery_and_whole_brain_therapy_in_20160421-4448-17vft7u.pdf?AWSAccessKeyId=AKIAIWOWYYGZ2Y53UL3A&Expires=1529072117&Signature=7fVb8ljtHh3s+FTrrjjIWw6DSvw=&response-content-disposition=inline; filename=Radiosurgery_and_whole_brain_therapy_in.pdf


150+ pts with whole brain followed by stereo to brainstem:
https://www.hindawi.com/journals/isrn/2013/652895/tab1/


Why would you use a dose lower than that for whole brain because you are reducing the volume/much longer interval between treatments - I would think the opposite would be true? When I trained 10 years ago it was very common for almost all stereo to be combined with whole brain. This was not considered "re irradiation," but if you separate them by a year it is?

Anyone who has been active in stereo for 10 years, will have had some experience delivering 12-16 Gy stereo to the brainstem following whole brain failure, but somehow doing it in reverse with a lower volume/long time interval raises issues.

Whole brain radiation therapy with or without stereotactic radiosurgery boost for patients with one to three brain metastases: phase III results of... - PubMed - NCBI
 
Last edited:
Because it's in the brainstem. If it was elsewhere in the brain, no problem.

I see that you edited your post. Ok I'll edit mine too.

Not sure what to make of that table. But the paper you referenced used a mean GK dose of about 11 Gy. The people I know who stereo brainstem mets after whole brain typically use 12 Gy or 5 Gy x 5 fractions (equivalent BED2).

11 Gy SRS and 30 Gy in 10 gives a combined BED of 146.5 Gy2 which again is in my typical comfort zone. As you go higher, you'll push the risk.

Still, maybe it's ok to do 30 Gy in 10 in the current setting of prior 14 Gy in 1 fraction. It depends on your risk tolerance. You certainly have a nice time interval there on your side.
 
Last edited:
I dont want be flippant but the bed calculations can make you nervous, so consider ignoring them given the widespread empirical experience combining whole brain with stereo
 
I don't take it as flippant. This is how I approach brainstem re-irradiation, but it's just one possible model that could be erroneous.

I'm still not sure I would do more than a focal 30 Gy in 10. However, maybe I'm being too conservative and I will review your data further and I'm open to further discussion.
 
Neuronix said:
11 Gy, 1 fraction, alpha/beta 2 = BED 71.5 Gy2
30 Gy, 10 fractions, alpha/beta 2 = BED 75 Gy2
75+71.5 = 146.5

All calculations courtesy of Android RBApp.
Click to expand...
I thought so. My take:

Why would you choose an alpha/beta of 2 for CNS. It's generally accepted to be 3-ish, especially given that there seems to be substantial time-based "forgiveness" of previous XRT dose in at least parts of the CNS. If you set it too low (unnecessarily), you risk underdosing the cells that have alpha/betas in the 10 range (ie NSCLC cells in the brainstem). You already have to underdose as is given previous XRT. (Disclaimer: it's true that if 3 is "wrong," and 2 is "right," you risk much higher late effects for a given "safe" fractionation--but if you worry less about late effects here, like I do, then alpha/beta=3 is reasonable and 2 isn't).
 
Neuronix said:
I don't take it as flippant. This is how I approach brainstem re-irradiation, but it's just one possible model that could be erroneous.

I'm still not sure I would do more than a focal 30 Gy in 10. However, maybe I'm being too conservative and I will review your data further and I'm open to further discussion.
Click to expand...

Another point... if the a/b of CNS is actually 2, then the BEDGy2 of 54/30 is about 103 Gy2. So by suggesting an accumulated dose of ~146.5 BED-Gy2 as "safe," you also suggest that....

146.5=1.8*x*(1+1.8/2) = x = 43

... 43 fractions of 1.8 Gy (about 77 Gy in 1.8 Gy fractions) to the brainstem is "safe" and I say NO WAY.

EDIT: which is to say, if CNS a/b=2, 11/1 plus 30/10 would be 50% more late effect probability than 54/30 and thus, super-dangerous. Reductio ad absurdum.
 
Last edited:
scarbrtj said:
Another point... if the a/b of CNS is actually 2, then the BEDGy2 of 54/30 is about 103 Gy2. So by suggesting an accumulated dose of ~146.5 BED-Gy2 as "safe," you also suggest that....

146.5=1.8*x*(1+1.8/2) = x = 43

... 43 fractions of 1.8 Gy (about 77 Gy in 1.8 Gy fractions) to the brainstem is "safe" and I say NO WAY.

EDIT: which is to say, if CNS a/b=2, 11/1 plus 30/10 would be 50% more late effect probability than 54/30 and thus, super-dangerous. Reductio ad absurdum.
Click to expand...

You are taking me far out of context here. BED 125-135Gy2 is the combined dose for re-irradiation with an interval of at least 6 months. That is stated in the paper I referenced and is also used in a number of other spinal papers. If you like I can find you those references by Rades, Sehgal, etc.

Also, most papers on CNS re-irradiation use an alpha/beta of 2. RT necrosis is considered a late effect.
 
Neuronix said:
You are taking me far out of context here. BED 125-135Gy2 is the combined dose for re-irradiation with an interval of at least 6 months. That is stated in the paper I referenced and is also used in a number of other spinal papers. If you like I can find you those references by Rades, Sehgal, etc.

Also, most papers on CNS re-irradiation use an alpha/beta of 2. RT necrosis is considered a late effect.
Click to expand...
Never seen a/b=2 for CNS; but clearly I don't like it (it just doesn't seem to fit clinically). Can't be for sure, but I assume one reason they chose a/b=2 was to make 50/25 equal 100 BED-Gy2 (says so in paper). Anything in a combined-course fashion, eg GK + WBRT, which has a BED-Gy2 greater than 100 BED-Gy2 has a higher late effect risk than 50/25. Using 11/1 GK and 30/10 WBRT = 146.5 BED-Gy2. Which implies an increased 46.5% late side effect incidence over and above 50/25. Do I believe that? Do you?

If the a/b=3 for CNS, 11/1 + 30/10 = 111 BED-Gy3. And 50/25 = 83. Which would be a ~33% increased risk of late effects.

If a/b=4, it works out to ~25% increased risk of late effects.

If one believes 11/1 plus 30/10 has no more increased late effect risk than 50/25...
11*(1+11/x) + 30*(1+3/x) = 50*(1+2/x) = x = ~12
... the alpha/beta of CNS is closer to 12, which implies that CNS repairs sublethal damage (SLD) much faster than if its a/b equals 2 or 3. There is a lot of data for that IMHO; a/b=2 means SLD repair occurs over VERY long time periods (relatively). Again, if you choose a/b=2, it makes repeat dosing calculations all the more conservative (in terms of risking late effects) than assuming a/b=3.
 
Neuronix said:
I don't follow your logic about 46.5% late side effect incidence over and above 50/25.

Radiation Oncology/Toxicity/Spine - Wikibooks, open books for an open world

Not a single paper referenced here uses alpha/beta = 3.
Click to expand...

If a/b=2

50/25 = 100 BEDGy-2

11Gy gamma knife then 30/10 WBRT = 146.5 BED-Gy2

(146.5-100)/100 = 46.5% higher BED-Gy2 for 11/1 plus 30/10 vs 50/25 (with attendant late side effect risk increase). The BED calc's are *only* meaningful in relation to another iso-alpha-beta BED calc's, after all.
 
That's not the way to interpret the data. There is a combined BED2 cut point found in the patient experience that the authors label as "safe" compared to a cut point above which they observed toxicity.

See also: Reirradiation human spinal cord tolerance for stereotactic body radiotherapy. - PubMed - NCBI

"Nieder et al. suggested a cumulative total BED of 135.5 Gy2 (nBED = 68 Gy2/2) as safe, provided that the interval between courses is ≥6 months and neither course exceeds 98 Gy2 (nBED = 49 Gy2/2) (23, 24). Our cumulative total BED is slightly greater at 140 Gy2 (nBED = 70 Gy2/2);"

This reference has its own caveats (the SBRT is more conservative than TG-101, the objectives are for thecal sac, and it suggests no more 9 Gy single fraction after conventional fractionation). Again it is for spinal cord (not brainstem). The point of my posts was, once you start going above these levels, you start to run into radiation myelitis risk in the cord. Whether that translates into cranial neuropathy risk in the brainstem is not well defined, but in the absence of good data it is what I use to estimate the risk to the brainstem. It could be overly conservative or could apply to medulla but not pons. I have also seen people use brainstem surface vs. brainstem core.
 
Neuronix said:
hat's not the way to interpret the data.
Click to expand...
I'm not interpreting data per se. I'm giving you the BED-Gy2 for 50/25 vs a combined SRS/WBRT of 11Gy and 30/10, and comparing the BED's. The BED of the latter is 46.5% greater than the BED of the former. Disregard the paper (or time or anything else) for that simple comparison.
Neuronix said:
The point of my posts was, once you start going above these levels, you start to run into radiation myelitis risk in the cord.
Click to expand...
Which "level"? They're suggesting 50/25 equals a "100 level." They then say a 40% higher level is safe. That would suggest 70/35 ("nBED = 70 Gy2/2") in one course is safe. There's some data to say that level may not be safe. So clearly, only 140 BED-Gy2 is safe to the cord over a long period of time (ie with an interdigitated tx break, ie two separate courses). To me, that's prima facie evidence the CNS alpha/beta is higher than 2: the CNS clearly repairs quite a bit of SLD in 6 mos.
Neuronix said:
Our cumulative total BED is slightly greater at 140 Gy2
Click to expand...
Is it really? We don't go above 50/25 (100 Gy2) in one sitting to the cord. We have learned that the hard way. These guys suggest you can go 40% higher than 50/25. Yes, yes... breaks, separate courses, >6 mos interval, etc. So in reality there is forgiveness, and the BED is not "really" 140 lifetime, IMHO. It's semantics I admit:
1) You can give a cumulative 140 BED-Gy2 to the CNS safely with a long break between courses.
vs
2) After a long break, the BED of a given course to the CNS decreases by 50% or more.
 
Neuronix said:
Not a single paper referenced here uses alpha/beta = 3.
Click to expand...
TBH this is pretty good to know; I hadn't really considered different alpha/betas for CNS since residency. In spirit of further honesty, I have used a/b=3 for so many years I'll stick with it. You made me crack out my Hall. Which I have the 5th edition, and in Table 22.1 he lists a/b of 1.7-4.9 for CNS. And then in the problem examples at end of chapter he says, "the alpha/beta is assumed to be 3 Gy for late-responding tissues and 10 Gy for early-responding tissues. The reader, of course, may substitute other values if they seem to be more appropriate." The difference of 2 vs 3 is not insignificant :)
 
Ya’ll lost me at hello.
 
  • Like
Reactions: 1 users
RadOncDoc21 said:
Ya’ll lost me at hello.
Click to expand...
It's wonky. BUT, I've always found it helpful to float freely--with some confidence--between different fraction sizes at will/per need, sliding around the late/early/tumor effects along the way.
 
Out of curiosity, what are people using as their initial (not re-irradiation) dose/fractionation for brainstem mets? How many doing SRS, vs fractionated SRT?
 
Radiator20 said:
Out of curiosity, what are people using as their initial (not re-irradiation) dose/fractionation for brainstem mets? How many doing SRS, vs fractionated SRT?
Click to expand...
I do SFRT only. There are a few retrospective (and thus possibly biased!) data, suggesting SFRT may be better.
I usually give 6 x 5 Gy to the 95%-PTV-encompassing isodose.

Edit: Typo.
 
Last edited:
I usually give 6 Gy x 5 and limit max point dose to 31. 2 in any overlap between PTV and brainstem and try to give 30 GY to less than 05% of brainstem volume- GTV. Sometimes this means prescribing 95% dose to 95% volume or reducing dose to 5.5 x 5 if you are expanding ptv by 2mm (which I typically do, as these lesions are often small )
 
Just wanted to say, thanks for the interesting back and forth between Neuronix and Scarbtj. I'd likely go with A/B=2 for any CNS stuff, personally.

I don't plan on single fractionating brainstem mets pretty much ever, even in the upfront setting.

I'm late to the discussion, but a fractionated approach of 5Gy x 5 or just doing 30Gy in 10 will give the patient relatively durable palliation without overt risk of brainstem toxicity, IMO.
 
Yes I do 15 Gy in a single fraction for brainstem mets (Gammaknife, no margin). Edit: up to 2 cm. I haven't run into this, but if it is larger I'd probably fractionate.

That got me thinking... If they come back and need whole brain at least 6 months in the future, sure I give 30 Gy in 10 fractions.

So I probably overthought the earlier case when I wrote 20 Gy in 10 fractions. I still wouldn't go higher than 30 Gy in 10 though. Maybe I'm a wuss.
 
  • Like
Reactions: 1 user
Neuronix said:
Yes I do 15 Gy in a single fraction for brainstem mets (Gammaknife, no margin).

That got me thinking... If they come back and need whole brain at least 6 months in the future, sure I give 30 Gy in 10 fractions.

So I probably overthought the earlier case when I wrote 20 Gy in 10 fractions. I still wouldn't go higher than 30 Gy in 10 though. Maybe I'm a wuss.
Click to expand...

I do the same.

Anything around 1.5-2.0 cm or larger I take to the linac to fractionate - extrapolating from the non-randomized italian data (9 Gy X 3 or 5-6 Gy X 5). For single fraction I give ~15 Gy on the GK.

Cleveland clinic has a nice paper/experience with single fraction for brainstem mets, albeit with some poor survival and in an era without immunotherapy.

Stereotactic radiosurgery for single brainstem metastases: the cleveland clinic experience. - PubMed - NCBI
 
  • Like
Reactions: 1 user
evilbooyaa said:
Just wanted to say, thanks for the interesting back and forth between Neuronix and Scarbtj. I'd likely go with A/B=2 for any CNS stuff, personally.
Click to expand...

Thank ya very much. Not to beat the dead horse, but using the eq's from my original post above, you arrive at 25 Gy/5 fx's as "safe" (4.5 fractions actually, but I rounded up) here with a/b=3 and just 20Gy/4 fx's if you go with a/b=2. Again, the lower you set your a/b, the more "tumor conservative" your retreat doses must become. As everyone here is choosing 25/5 for re-treat everyone is actually saying CNS a/b equals 3 instead of 2 :)
 
I'll be a bit contrarian and say I'd go 6 Gy x 5.
You can alpha/beta all you want, but if it's truly the sole site of disease, try to control the tumor.
You can treat necrosis with Avastin/steroids; you won't have another chance at tumor control.
Thalamus/midbrain interface is different than pons. Less prone to necrosis and less devastating if it manifests. Equally fatal when tumor recurs.
 
You must log in or register to reply here.
Next unread thread

Similar threads

SBRTreble
Zap-X and SRS
Replies
8
Views
583
Gfunk6
Gfunk6
R
Re-SRS in a patient who has had WBRT?
Replies
35
Views
1K
NotMattSpraker
N
Kroll2013
re-Irradiation of brainstem lesion after previous SRS
Replies
25
Views
2K
radoncopotamus
R
R
SRS and refractory depression
Replies
15
Views
1K
RickyScott
R
M
Linac based repeat srs dose
Replies
13
Views
1K
Ray D. Ayshun
R
Top