Whole brain RT and neuro-cognitive outcomes

[Gfunk6]

For your patients that undergo WBRT, do any of you perform any interventions to reduce the risk of neuro-cognitive sequeale?

Based on the recent results of RTOG 0614 there appears to be a benefit of using memantine (Namenda). However, there are some statistical problems with the data as mentioned in the article.

I also haven't seen preliminary results for hippocampal-sparing WBRT from RTOG 0933. However, I do use the "hippocampal avoidance zone" as an optional dose constraint in my younger patients (age < 50) receiving partial brain XRT.

Any other studies people are aware of?

---

Members don't see this ad.

 

[Pointless]

Are people in practice seeing a lot of neurocognitive sequelae following WBRT? From my limited experience, it just doesn't seem to be a big issue for the patients that we treat, probably owing to the fact that most of the patients who receive WBRT have life expectancies in the 4-6 month range... thoughts?

 

[seper]

I see fair amount of cognitive decline in survivors. I don't do hippocampal sparing and insurances don't pay for prophylactic Namenda.

 

[Gfunk6]

Are people in practice seeing a lot of neurocognitive sequelae following WBRT? From my limited experience, it just doesn't seem to be a big issue for the patients that we treat, probably owing to the fact that most of the patients who receive WBRT have life expectancies in the 4-6 month range... thoughts?

Breast caner patients can live for a few years, even with brain mets. In my clinic, this group of patients tend to live long enough to have neurologic issues. For other common histologies (especially lung), you are right that they frequently do not live long enough to suffer neuro-cognitive decline from tx.

 

[Palex80]

Catchy techniques and expensive medication are an option.
On the other hand, patients who we think may live long enough to experience their long term neuro-cognitive impairment from WBRT, could benefit from more fractionated schemes.
You can try 36/2 for example instead of 30/3.

I sometimes prescribe 30/2 followed by a stereotactic boost to the resection cavity for patients with solitary resected mets.
I've also had two patients, which I treated with 30/1 b.i.d. for SCLC very-LD PCI.
They are both doing well at 2 years follow-up without any sequlae and still recurrence free. Both of them still have their desk jobs.

 

[Gfunk6]

On the other hand, patients who we think may live long enough to experience their long term neuro-cognitive impairment from WBRT, could benefit from more fractionated schemes.
You can try 36/2 for example instead of 30/3.

I've heard this argument before but is there ANY data to back it up? Part of the problem is that trials to determine whole-brain RT dose were done during an era with poor metrics of neuro-cognition (e.g. MMSE). Also, if you increase the dose of whole brain RT you can increase the risk of necrosis with salvage SRS. Finally, if you take the leap of faith that neural stem cell eradication is the mechanism of neuro-cognitive decline from WBRT then fractionation will not help. The RTOG 0933 dose constraints for the "hippocampal avoidance zone" are D100 < 9 Gy and max dose < 16 Gy.

 

[TarHeelRadOnc]

I see fair amount of cognitive decline in survivors. I don't do hippocampal sparing and insurances don't pay for prophylactic Namenda.

Have used memantine for 2 patients (both ER positive breast ca) receiving WBRT. Private insurance has paid for drug in both cases. Must be regional variation.

 

[Palex80]

I've heard this argument before but is there ANY data to back it up?

The only data I am aware of come from CNS-lymphoma RTOG trials on hyperfractionated treatment. You can also speculate that some of the data for medulloblastoma treatment also point into that direction.

Part of the problem is that trials to determine whole-brain RT dose were done during an era with poor metrics of neuro-cognition (e.g. MMSE).

Absolutely. Plus you have a huge drop-out rate from patients, who die before they can undergo repeat MMSE. So you statistics suck...

Also, if you increase the dose of whole brain RT you can increase the risk of necrosis with salvage SRS.

I have never seen brain necrosis because of salvage SRS after WBRT. Pardon me, but I have never seen it.
I really don't think that salvage SRS after 36/2 will lead to more necrosis than after 30/3.

Finally, if you take the leap of faith that neural stem cell eradication is the mechanism of neuro-cognitive decline from WBRT then fractionation will not help. The RTOG 0933 dose constraints for the "hippocampal avoidance zone" are D100 < 9 Gy and max dose < 16 Gy.

I agree. On the other hand we do regularly see leukoencephalopathy in other areas outside of the hippocampus after WBRT with a decline in MMSE. You are not going to stop these kind of side effects if you spare out the hippocampus.

I find hippocampus sparing fascinating, but I have heard of some very controversial data on the dose constraints you have to meet, to actually spare it. Some people claim, that as little as 2 Gy to the hippocampus will kill of these stem cells (PMID: 12874001), so if you aint using protons (IMPT) you can probably forget about reaching that kind of a constraint even with your most sophisticated IMRT. The RTOG trial will hopefully tell us more about what dose really matters.

Until then, I find it ok to have the patient come an extra 1.5 weeks for treatment, if she/he is in good shape.

 

[Gfunk6]

For memantine dosing, RTOG used 20 mg/daily. Do you Rx 20 mg po daily or 10 mg po bid?

 

[Ursus Martimus]

patients who we think may live long enough to experience their long term neuro-cognitive impairment from WBRT, could benefit from more fractionated schemes.
You can try 36/2 for example instead of 30/3.

I sometimes prescribe 30/2 followed by a stereotactic boost to the resection cavity for patients with solitary resected mets.
I've also had two patients, which I treated with 30/1 b.i.d. for SCLC very-LD PCI.
.

Is there is stitch of evidence for any of the above? Or are you just quoting from the J. of Anecdotal Med?

 

[Palex80]

Is there is stitch of evidence for any of the above? Or are you just quoting from the J. of Anecdotal Med?

No need to be sarcastic.

There is no randomized trial looking into the effects of fractionation on long term toxicity.

However, I think that most radiation oncologists would agree, that the fraction size is the most important factor in predicting long term toxicity of radiation therapy in general, if you deliver the same level of dose.

There is some data coming from lymphoma patients treated with WBRT pointing out that hyperfractionated therapy can be beneficial in terms of neurocognitive outcomes.
http://www.ncbi.nlm.nih.gov/pubmed/16193393
Medulloblastoma data also point out, that hearing loss may remain at the same level, despite higher total dose due to hyperfractionation.
http://www.ncbi.nlm.nih.gov/pubmed/22851561

I understand your concerns however.

 

[deleted4401]

Do you do 30 Gy/10 for every patient, UM?

The post op trial by Patchell used 1.8 Gy fractions, so there is a stitch of evidence for lower fractionation, in fact, it's in a seminal paper that PGY-2s can quote. People use any range of fractions and it's not unreasonable to drop the fx dose for brain mets patients.

I use GPPA, and use 6 months as a cut off for changing from 30/10 to 35/14, but 40/20 is used occasionally. Caveat is for breast CA. They seem to do better so 35/14 for most of them. And, even if GPPA predicts poor outcome, if under 50, I'll consider it. Even with that scheme, 75% get 30 in 10.

S

 

[deleted4401]

So, with the caveat that I have to send the patient's out for SRS, my bias is for patients with 1-3 mets, unless they are at death's door, to send them for SRS. I almost never give combination therapy. Coming from a training that regarded WBRT on the same footing as cyanide, I prefer SRS because its fast - gets them to chemo quicker, probably less toxic, and doesn't cost as much more than WBRT then it did in the past (especially now). Hopefully, we will join the late 20th century soon.

The people with 4 or more, that's who I use the GPPA for.

 

[medgator]

The people with 4 or more, that's who I use the GPPA for.

GPPA? I guess I'm a bit behind the times. I know of RPA classification but I googled wikibooks to find out what you were talking about.

http://en.wikibooks.org/wiki/Radiat...erview#Graded_Prognostic_Assessment_.28GPA.29

GPA vs RPA per wikibooks:

http://www.ncbi.nlm.nih.gov/pubmed/17931798?dopt=Abstract

The new index, the GPA, is as prognostic as the RPA and more prognostic than the other indices. The GPA is the least subjective, most quantitative and easiest to use of the four indices. Future clinical trials should compare the GPA with the RPA to prospectively validate these findings.

 

[TarHeelRadOnc]

Do you do 30 Gy/10 for every patient, UM?

The post op trial by Patchell used 1.8 Gy fractions, so there is a stitch of evidence for lower fractionation, in fact, it's in a seminal paper that PGY-2s can quote. People use any range of fractions and it's not unreasonable to drop the fx dose for brain mets patients.

I use GPPA, and use 6 months as a cut off for changing from 30/10 to 35/14, but 40/20 is used occasionally. Caveat is for breast CA. They seem to do better so 35/14 for most of them. And, even if GPPA predicts poor outcome, if under 50, I'll consider it. Even with that scheme, 75% get 30 in 10.

S

Patchell study also treated to 50.4Gy, and was for patients with solitary brain mets and predominantly RPA I classification... ie not the typical patient in whom we are considering WBRT

The lymphoma data referenced by Palex was after high dose MTX. The medullo data is minimally applicable to adult patients with metastatic disease.

I guess what I'm trying to say is that there aren't data suggesting improved neurocognitive outcomes with 35/14 or 40/20 compared to 30/10. I do 30/10 in nearly all WBRT patients. For patients getting combination (SRS&WBRT) or getting salvage WBRT after prior SRS I will use 2.5Gy/fx in an attempt to limit RT necrosis risk. For rare patients with comparatively favorably prognosis/natural hx (ie ER+ Breast, limitted extracranial dz, long interval between primary tx and metastatic progression) I will do 35/14, but as Simul pointed out these "favorable px" whole brain cases represent a small minority. Also agree with a "whole brain sparing" approach (ie SRS alone) for patients with 1-3 mets and good prognosis.

I am prescribing Namenda in ~1/3 of my WBRT patients (again for those with comparatively favorable px). I agree hat there are statistical problems with 0614, but figure that we have better data for Namenda than we do for adjusting the WBRT fractionation in these patients.

 

[medgator]

Patchell study also treated to 50.4Gy, and was for patients with solitary brain mets and predominantly RPA I classification... ie not the typical patient in whom we are considering WBRT

The lymphoma data referenced by Palex was after high dose MTX. The medullo data is minimally applicable to adult patients with metastatic disease.

I guess what I'm trying to say is that there aren't data suggesting improved neurocognitive outcomes with 35/14 or 40/20 compared to 30/10. I do 30/10 in nearly all WBRT patients. For patients getting combination (SRS&WBRT) or getting salvage WBRT after prior SRS I will use 2.5Gy/fx in an attempt to limit RT necrosis risk.

.

Sometimes we just don't have data for things and we have use our clinical judgement and understanding of the disease course and underlying radiation biology.

I use 2.5 Gy/fx in general because that's how the rtog did it in 95-0? when incorporating srs. And many patients may end up with srs down the road.

If someone has a pretty borderline PS, I'll do 30/10 or even the occasional 20/5

 

[TarHeelRadOnc]

Sometimes we just don't have data for things and we have use our clinical judgement and understanding of the disease course and underlying radiation biology.

I use 2.5 Gy/fx in general because that's how the rtog did it in 95-0? when incorporating srs. And many patients may end up with srs down the road.

If someone has a pretty borderline PS, I'll do 30/10 or even the occasional 20/5

I get the art vs science of radiation oncology discussion as it relates to fractionation. But I think that argument can go both ways.

Multiple trials have compared different WBRT fractionation schedules and found no benefit in terms of survival outcomes for one regimen over another. In terms of neurocognitive outcomes, the strongest radiobiologic predictor of late toxicity is total dose not dose per fraction. The BED2 of 30/10 (using a generic alpha/beta of 3 for late effects) is 36Gy, whereas the BED2 of 35/14 is 38.5Gy and for the RTOG dose of 37.5/15 its 41.25Gy. Using almost any radiobiologic model, the estimated risk of late (neurocognitive) toxicity would be higher for the later regimens than for the former.

To get back to a bigger picture issue, I think it is interesting on this thread that some are discounting the results of RTOG 0614 - a trial that is largely considered to be positive by the principle investigators and the RTOG- due to relatively minor statistical concerns and are simultaneously arguing for more protracted WBRT fractionation, in the complete absence of data demonstrating a benefit. I'm not arguing that all WBRT patients should get Namenda- or 30/10 for that matter; but for the comparatively good px patients (with estimated survival >6 mo) the data is much stronger for concurrent & adjuvant Namenda than for a 4 week course of whole brain.

 

[seper]

"In terms of neurocognitive outcomes, the strongest radiobiologic predictor of late toxicity is total dose not dose per fraction"

What is your statement based on? I honestly must have missed the paper that you are referring to.

 

[TarHeelRadOnc]

"In terms of neurocognitive outcomes, the strongest radiobiologic predictor of late toxicity is total dose not dose per fraction"

What is your statement based on? I honestly must have missed the paper that you are referring to.

No paper.

Dose definitely matters. POG 8631 showed ~10 point IQ difference in long term survivors of medulloblastoma with 23.4Gy vs 36Gy. Similarly, RTOG 0212 showed neurocognitive deficits in 60% of those randomized to 25/10 versus 80-90% in those randomized to 36/18 or 36/24 BID. By contrast, no study has shown an improvement in neurocognitive outcomes with regimens using 2-2.5Gy/fx compared to standard 30/10. In fact in cross-study comparisons, the incidence of neurocognitive decline on MMSE was more common among patients treated on the WBRT alone arm of RTOG 0118 (37.5/15) compared to the 30/10 arm of RTOG 9104.... Certainly not proof, but hypothesis generating.

My statement was based more on simple radiation biology. The linear-quadratic model, and resultant BED equation, were designed to compare different fractionation regimens in terms of tumor control and late toxicity. Regardless of whether you use an alpha/beta of 2 or 3 for late responding tissues, changes in total dose are weighted more than changes in dose per fraction. For example, using an alpha/beta of 3, the BED of 30/10 is equivalent to 31.8/12. These biologically equivalent regimens are 12% different in dose/fraction but only 6% different in total dose. In other words, for each ~1% change in total dose, you need a ~2% opposite direction change in fractional dose to maintain biological equivalence... ie adjustments in total dose have a greater predicted influence on late toxicity (and tumor control) than similar proportional changes in dose per fraction (except perhaps for tumor control in the unique settings of melanoma and possibly prostate cancer). That is why the BED2 for late effects of 30/10 is lower than that for 35-37.5/2.5 or 40/2 (see prior post). The BED for tumor control (using generic alpha/beta of 10) is also lower for 30/10, but this has not translated into an improvement in survival with the more protracted regimens.

 

[RadRadRad]

"In terms of neurocognitive outcomes, the strongest radiobiologic predictor of late toxicity is total dose not dose per fraction"

What is your statement based on? I honestly must have missed the paper that you are referring to.

As Tarheel mentions, there is real data (not radiobiological theories) showing worse neuro toxicity from higher dose RT with smaller daily fractions than lower dose with larger daily fractions. PMID: 20800380

 

[Palex80]

One can also quote the negative EORTC and RTOG trials on Astro°II dose-finding, with higher side-effect rates in the high-dose arms.

However, one has to be careful, when comparing different fractionation schedules for WBRT. Patients treated with 2-2.5 Gy/d tend to get higher BEDs for the tumor as well.
Most people will do 20 x 2 Gy instead of 10 x 3 Gy, if they are concerned about late neurocognitive sequlae. However 10 x 3 Gy is about 18 x 2 Gy in terms of BED for tumor and/or CNS.
Thus, reducing the dose per fraction often leads to people prescribing higher total BEDs as well.

Why is that?
One reason could be, that people don't use BED-calculators. The other reason is selection bias. Patients who get lower doses per fraction are usually the ones with better prognosis. The better the prognosis, the higher the tendency of physicians to prescribe more aggressive treatment.

 

[medgator]

As Tarheel mentions, there is real data (not radiobiological theories) showing worse neuro toxicity from higher dose RT with smaller daily fractions than lower dose with larger daily fractions. PMID: 20800380

25/10 vs 36/18. I think that's why many do 30/15 (including myself). Both 25/10 and 30/15 are recognized PCI regimens in the NCCN.

 

[seper]

You are right, there no data to support your statement. Both total dose and dose per fraction matter, and we don't know which parameter is the most important.

No paper.

Dose definitely matters. POG 8631 showed ~10 point IQ difference in long term survivors of medulloblastoma with 23.4Gy vs 36Gy. Similarly, RTOG 0212 showed neurocognitive deficits in 60% of those randomized to 25/10 versus 80-90% in those randomized to 36/18 or 36/24 BID. By contrast, no study has shown an improvement in neurocognitive outcomes with regimens using 2-2.5Gy/fx compared to standard 30/10. In fact in cross-study comparisons, the incidence of neurocognitive decline on MMSE was more common among patients treated on the WBRT alone arm of RTOG 0118 (37.5/15) compared to the 30/10 arm of RTOG 9104.... Certainly not proof, but hypothesis generating.

My statement was based more on simple radiation biology. The linear-quadratic model, and resultant BED equation, were designed to compare different fractionation regimens in terms of tumor control and late toxicity. Regardless of whether you use an alpha/beta of 2 or 3 for late responding tissues, changes in total dose are weighted more than changes in dose per fraction. For example, using an alpha/beta of 3, the BED of 30/10 is equivalent to 31.8/12. These biologically equivalent regimens are 12% different in dose/fraction but only 6% different in total dose. In other words, for each ~1% change in total dose, you need a ~2% opposite direction change in fractional dose to maintain biological equivalence... ie adjustments in total dose have a greater predicted influence on late toxicity (and tumor control) than similar proportional changes in dose per fraction (except perhaps for tumor control in the unique settings of melanoma and possibly prostate cancer). That is why the BED2 for late effects of 30/10 is lower than that for 35-37.5/2.5 or 40/2 (see prior post). The BED for tumor control (using generic alpha/beta of 10) is also lower for 30/10, but this has not translated into an improvement in survival with the more protracted regimens.

 

[TarHeelRadOnc]

You are right, there no data to support your statement. Both total dose and dose per fraction matter, and we don't know which parameter is the most important.

I don't think that you read the above posts correctly. There are actually data (derived from a randomized clinical trial; not anecdotal experience and conjecture) suggesting that lower dose, mildly hypofractionated regimens are associated with less neurocognitive toxicity than longer conventionally fractionated regimens. By contrast, there are not data showing a survival benefit of the higher dose regimens.

We can debate all you want, but the arguments in favor of shorter course treatments for these patients are stronger than the arguments for longer fractionation. I don't have as much of a problem with 35/14 as I do with 40/20. Why would you monopolize a patient for a 4 week course of palliative whole brain if there is absolutely no benefit (and a theoretical detriment)? I'm hesitant to say it, but the fee-for-service reimbursement model is the only rationale I can come up with to use this regimen in routine clinical practice.

 

[Palex80]

I'm hesitant to say it, but the fee-for-service reimbursement model is the only rationale I can come up with to use this regimen in routine clinical practice.

[YOUTUBE]http://www.youtube.com/watch?v=qMxX-QOV9tI[/YOUTUBE]

 

[RadRadRad]

I think 30/15 is a very reasonable PCI regimen for small cell.
I don't see this regimen used very frequently outside of small cell PCI.

As Palex mentions there might be a variety of reasons to opt for lower dose per fraction. The reason I commonly hear is: this patient has a longer life expectancy than average and MD wants to reduce the chance of late neurocognitive effects. However in these same patient we want to maximize local control which typically means taking the total dose higher (think 2nd patchell trial where they went to 50.4 in 28 fractions). Yet the above mentioned study shows that you cannot achieve both of these goals, as higher total dose in lower fraction size leads to increased neurotoxicity.