Best dose/fractionation to Koos3 AN

[Kroll2013]

Advertisement - Members don't see this ad

A woman in her early 30s, previously healthy, presented with progressive right-sided hearing decrease and persistent tinnitus over several months.
No facial numbness, no facial palsy, no imbalance, and no other cranial nerve deficits were reported. Performance status normal.
Audiogram (08/06/2025):
 Right ear: Sensorineural hearing loss, PTA ≈ 32 dB HL, AI ≈ 59%.
Speech discrimination 100% at 60 dB, suggesting serviceable hearing (AAO-HNS Class B).
 Left ear: Essentially normal thresholds, PTA ≈ 15 dB HL.
She retains usable hearing in the affected ear, an important prognostic factor when selecting RT vs surgery.

MRI Brain with contrast (T1Gd, T2, CISS sequences) dated 11/11/2025.
Right vestibular schwannoma, Koos grade III
o Enhancing CPA mass extending into the IAC.
o Volume: ~4 cc (well within SRS/FSRT suitability range).
o Displaces but does not severely compress the brainstem.
o No cystic degeneration.
o Clear visualization of 7th and 8th nerve entry zone (last two images).
o No hydrocephalus.

we discussed with her observation (since she does not have previous serial MRIs). she insisted on doing RT.

what would be the best dose/fractionation regimen, knowing that the distance cochlea-tumor is <2mm ?
- 25 Gy in 5 fr at 80%
- 18 Gy in 3 fr at 80%
- 54 Gy in 28 fr
- other

 

[BobbyHeenan]

THese are challenging and I don't think anyone knows for sure the best approach here. In the US, some centers would treat this with single fraction gamma knife (probably UPMC and UVA), others would use 3-5 fractions. All of the options you have presented in my opinion are perfectly reasonable within standard of care.

I have a very busy CNS practice (have access to both gamma knife and linac radiosurgery platforms). Initially I treated most of these with 12.5 - 13 Gy X 1 single fraction, but now I tend to treat these with 25-26 Gy in 5. I had one patient get a cranial neuropathy after single fraction, so for larger ones I tend to fractionate. That is just anectdotal though.

I bet @Neuronix has some good ideas.

 

[Ray D. Ayshun]

THese are challenging and I don't think anyone knows for sure the best approach here. In the US, some centers would treat this with single fraction gamma knife (probably UPMC and UVA), others would use 3-5 fractions. All of the options you have presented in my opinion are perfectly reasonable within standard of care.

I have a very busy CNS practice (have access to both gamma knife and linac radiosurgery platforms). Initially I treated most of these with 12.5 - 13 Gy X 1 single fraction, but now I tend to treat these with 25-26 Gy in 5. I had one patient get a cranial neuropathy after single fraction, so for larger ones I tend to fractionate. That is just anectdotal though.

I bet @Neuronix has some good ideas.

How do you prescribe, to 80% line as mentioned, or a more homogenous plan?

 

[BobbyHeenan]

How do you prescribe, to 80% line as mentioned, or a more homogenous plan?

Usually hot plans on any platform...GK usually 50-75% isodose lines, linac 80 ish, sometimes hotter.

If long course standard frac (I reserve for very large tumors who can't have surgery, more mass effect and/or edema) I use a homogenous plan.

A hotter plan often allows for more rapid dose fall off/better sparing of cochlea or even adjacent cranial nerves. I usually contour associated nerves to be aware of dosing, though if a tumor is running through a nerve or touching the tumor, only so much you can spare it.

 

[Neuronix]

THese are challenging and I don't think anyone knows for sure the best approach here. In the US, some centers would treat this with single fraction gamma knife (probably UPMC and UVA), others would use 3-5 fractions. All of the options you have presented in my opinion are perfectly reasonable within standard of care.

I have a very busy CNS practice (have access to both gamma knife and linac radiosurgery platforms). Initially I treated most of these with 12.5 - 13 Gy X 1 single fraction, but now I tend to treat these with 25-26 Gy in 5. I had one patient get a cranial neuropathy after single fraction, so for larger ones I tend to fractionate. That is just anectdotal though.

I bet @Neuronix has some good ideas.

Thanks for mentioning me.

This is very likely to keep growing at that size and that age. You could get a serial MRI to prove it, but odds are very high that it's growing. It will almost certainly need management over the next few years. So I think it's reasonable to treat now or have a plan for future treatment.

Where I've practiced, these would mostly be resected based on the patient's age and fears of secondary malignancy. In addition, that will be hard to salvage in the event of an RT failure, and the patient has a long life expectancy in which it can fail. Trying to preserve the serviceable hearing could be one reason to wait to operate.

Before I got thrown out of my last job, I treated one or two of these a week. I have used every SRS dose-fraction and both GK and Linac due to my own evolving preferences and working with many different surgeons. I've mostly had good responses, various hearing outcomes, and rare complications with all dose-fractionations to the point where I feel like it's a bit of a wash. The data is well summarized here: Stereotactic Radiosurgery for Vestibular Schwannomas: Tumor Control Probability Analyses and Recommended Reporting Standards - PubMed .

One investigator I know has some published animal data that 18 Gy / 3 fractions preserved the cochlea better than 12 Gy / 1 fraction, so I tended to prefer fractionation based on that. But, I'm still unclear whether that's really true in humans. There is an NRG concept in development by Tony Wang and the Columbia group to test 1 vs. 5 fraction for vestibular schwannoma, which I hope gets off the ground.

How do you prescribe, to 80% line as mentioned, or a more homogenous plan?

On GK I typically do 50% IDL with manual planning, but consider a higher IDL at times. With the new optimizer sometimes it can give you good plans with higher IDL so I don't force it to 50% IDL. We did a paper in a different context that also looked at IDL and a lot of other parameters including CI/GI (there are many others who have looked at this as well), and it's not really that straightforward to just say 50% is the right IDL to make the best plan unless you believe that the hot spot dose matters in some way for tumor control. I mean 50% IDL is fine, but if the autoplanner pops out with a random 64% or whatever, that's also fine with me.

On MLC-based linac SRS I strongly consider setting Dmax 120% given all the 80% IDL data from cone linac and CK. Though if I do 12 Gy in 1 fraction on that system, do I just set Dmax 200% since that's allowed on GK? Then why not do that for 18 Gy / 3 fractions or 25 Gy / 5 fractions since fractionated should be gentler? Anyway, it's a thought experiment really, and I'm curious to hear if anyone has any good reasons for hot spot limits beyond just historical reasons. I have heard an argument for dose rate effects.

One thing to consider is that constraining the hot spot in the optimizer sometimes can drive up the CI/GI with tiny targets. Something the size we're discussing will likely have a Dmax in the 110-115% range from an SRS optimizer in my experience.

@Kroll2013 BTW, what is the PTV margin you have in your contours? It looks pretty generous to me.

 

[Kroll2013]

Thanks for mentioning me.

This is very likely to keep growing at that size and that age. You could get a serial MRI to prove it, but odds are very high that it's growing. It will almost certainly need management over the next few years. So I think it's reasonable to treat now or have a plan for future treatment.

Where I've practiced, these would mostly be resected based on the patient's age and fears of secondary malignancy. In addition, that will be hard to salvage in the event of an RT failure, and the patient has a long life expectancy in which it can fail. Trying to preserve the serviceable hearing could be one reason to wait to operate.

Before I got thrown out of my last job, I treated one or two of these a week. I have used every SRS dose-fraction and both GK and Linac due to my own evolving preferences and working with many different surgeons. I've mostly had good responses, various hearing outcomes, and rare complications with all dose-fractionations to the point where I feel like it's a bit of a wash. The data is well summarized here: Stereotactic Radiosurgery for Vestibular Schwannomas: Tumor Control Probability Analyses and Recommended Reporting Standards - PubMed .

One investigator I know has some published animal data that 18 Gy / 3 fractions preserved the cochlea better than 12 Gy / 1 fraction, so I tended to prefer fractionation based on that. But, I'm still unclear whether that's really true in humans. There is an NRG concept in development by Tony Wang and the Columbia group to test 1 vs. 5 fraction for vestibular schwannoma, which I hope gets off the ground.



On GK I typically do 50% IDL with manual planning, but consider a higher IDL at times. With the new optimizer sometimes it can give you good plans with higher IDL so I don't force it to 50% IDL. We did a paper in a different context that also looked at IDL and a lot of other parameters including CI/GI (there are many others who have looked at this as well), and it's not really that straightforward to just say 50% is the right IDL to make the best plan unless you believe that the hot spot dose matters in some way for tumor control. I mean 50% IDL is fine, but if the autoplanner pops out with a random 64% or whatever, that's also fine with me.

On MLC-based linac SRS I strongly consider setting Dmax 120% given all the 80% IDL data from cone linac and CK. Though if I do 12 Gy in 1 fraction on that system, do I just set Dmax 200% since that's allowed on GK? Then why not do that for 18 Gy / 3 fractions or 25 Gy / 5 fractions since fractionated should be gentler? Anyway, it's a thought experiment really, and I'm curious to hear if anyone has any good reasons for hot spot limits beyond just historical reasons. I have heard an argument for dose rate effects.

One thing to consider is that constraining the hot spot in the optimizer sometimes can drive up the CI/GI with tiny targets. Something the size we're discussing will likely have a Dmax in the 110-115% range from an SRS optimizer in my experience.

@Kroll2013 BTW, what is the PTV margin you have in your contours? It looks pretty generous to me.

ty a lot for your reply. you are definitely right for the PTV- will stick to 1 mm - o mm near cochlea. reducing the mean dose to cochlea below 4-5 GY will be very challenging though.

 

[Neuronix]

ty a lot for your reply. you are definitely right for the PTV- will stick to 1 mm - o mm near cochlea. reducing the mean dose to cochlea below 4-5 GY will be very challenging though.

Yes this is how I used to do it on linac SRS, but more recently I've gone to GTV=PTV in a case like this. Obviously, you need a strong linac SRS setup for that and opinions do vary about marginless frameless linac radiosurgery.

The 4-5 Gy is a single fraction constraint I would assume since I'm not sure you'll even get that in one fraction. These cases are tough for cochlear sparing. My general rule of thumb is you can get to 1/2 of prescription dose in 1.5-2 mm if you push hard on one OAR.

What GTV or PTV coverage minimum do you use for the target? I see different opinions on how much you can cheat on coverage on a case like this to bring down cochlear dose. On the other hand, the correlation between cochlear dose and hearing outcome in the published studies is not the strongest data.

 

[Kroll2013]

Yes this is how I used to do it on linac SRS, but more recently I've gone to GTV=PTV in a case like this. Obviously, you need a strong linac SRS setup for that and opinions do vary about marginless frameless linac radiosurgery.

The 4-5 Gy is a single fraction constraint I would assume since I'm not sure you'll even get that in one fraction. These cases are tough for cochlear sparing. My general rule of thumb is you can get to 1/2 of prescription dose in 1.5-2 mm if you push hard on one OAR.

What GTV or PTV coverage minimum do you use for the target? I see different opinions on how much you can cheat on coverage on a case like this to bring down cochlear dose. On the other hand, the correlation between cochlear dose and hearing outcome in the published studies is not the strongest data.

exactly why i am requesting your help for this case , because no data on how low we can get GTV coverage without compromising tumor control and what are the relevant constraints on cochleas for 5 fractions, to preserve remaining hearing.

 

[Neuronix]

exactly why i am requesting your help for this case , because no data on how low we can get GTV coverage without compromising tumor control and what are the relevant constraints on cochleas for 5 fractions, to preserve remaining hearing.

I've heard anywhere from 95%-99% coverage is acceptable. Apparently, the series from Europe often use 95-98% coverage at prescription dose with no difference in control rates. I've been using 99-100% coverage of GTV at prescription dose, though sometimes I'll cheat to 98.x%.

Getting into the weeds a bit, even with a seemingly perfect fusion you can have slight differences in the GTV between the T1 post-contrast MP-RAGE and a high-resolution balanced steady state free precession (aka on Siemens scanners TRUE-FISP or TRUFI) in the IAC protocols, where the MP-RAGE looks larger. You can also window differently to change the change contours a bit. So what exactly is 95% or 99%? There are too many little variables across institutions and individuals sometimes to account for these nuances.

I don't think we have fractionated data for hearing preservation. My understanding from a few years ago was that nobody has found a direct relationship between fractionated SRS dose and hearing preservation for vestibular schwannoma. Though, I always joke there's a retrospective series for anything, so maybe I'm just unaware.

 

[evilbooyaa]

First off, I would treat and not observe. She is symptomatic, and a few mm growth is going to make life much more difficult down the line. I understand she's young at 30 but if she wants to try to preserve her hearing, would do RT and accept risk of secondary malignancy. If she's open to losing her hearing, then push for surgeyr to avoid RT in a 30 year old.

In terms of how to RT - I would mostly defer to Neuronix who has treated more of these in a month than I probably have in my life, but in my non-expertise, I would do 25/5 to this, allow Hot spots up to 120%, and prioritize brainstem/cochlear constraints as feasible, while shooting for 95/100 coverage at 25Gy. Tighten margins to 0-1mm at the cochlear interface with tumor. I'd need to see what the plan looked like in terms of how comfortable I felt with a 5Fx plan. Brainstem is fine going up to point dose 30Gy in 5fx.

I would never 1 fx this due to brain stem CC constraint. I would probably never 3Fx this due to cochlea constraint.

Additionally, I would consider it very reasonable to recommend conventional fractionation in this patient with some amount of serviceable hearing in the ear. If I could not meet 95/100 coverage at 25/5 while keeping D0.035cc of Cochlea < 25Gy and patient was amenable to 25 treatments, that's what I would do.