PCI for SCLC is DEAD!

Started by Palex80
Physicians and veterinarians: Turn your expertise into action! Explore CDC’s Epidemiology Elective Program and Epidemic Intelligence Service today.
This forum made possible through the generous support of SDN members, donors, and sponsors. Thank you.
Get help with your application

Use all the free resources available to you from SDN: articles, guides, expert advising, forums discussions, and school research.

Browse articles Compare schools My application plan Read Interview Feedback
Practice for interviews Ask an Advisor
Advertisement - Members don't see this ad
Ok, let me kick off the discussion then.

I saw (most of) the slides, and this is one slide, I don't quite understand.
1789313648709.png



So, the primary endpoint is CFFS meaning you have to show a decline in cognition (based on tests: Hopkins Verbal Learning Test–Revised (HVLT-R), Controlled Oral Word Association (COWA) Test, Trail Making Test (TMT Parts A and B) compared to your baseline test or die. Everyone starts at 100%, the endpoint is measuring decline, not your test result as an absolute mark.

Could someone explain to me, why at 3 months both groups drop by >50% compared to baseline?

Randomisation was done within 16 weeks of the last chemo cycle, so these patients had their baseline test while on chemo, or within a few months upon completion of chemo.

It's not death events, it's neurocogition events.
Less than 10% were dead at 3 months. And progression also doesn't seem that bad.
1789314096286.png


And it's likely also not brain mets. At 3 months the rates are still not that high.
1789313957724.png


I fully understand that PCI will produce neurocognitive decline at 3 months, but why is the surveillance curve also dropping that hard at 3 months? One would imagine that patients who are done with chemo, will actually have stable (if not even better) neurocognition. But a 50% drop?

Just a thought...
 
Palex80 said:
Ok, let me kick off the discussion then.

I saw (most of) the slides, and this is one slide, I don't quite understand.
View attachment 424494


So, the primary endpoint is CFFS meaning you have to show a decline in cognition (based on tests: Hopkins Verbal Learning Test–Revised (HVLT-R), Controlled Oral Word Association (COWA) Test, Trail Making Test (TMT Parts A and B) compared to your baseline test or die. Everyone starts at 100%, the endpoint is measuring decline, not your test result as an absolute mark.

Could someone explain to me, why at 3 months both groups drop by >50% compared to baseline?

Randomisation was done within 16 weeks of the last chemo cycle, so these patients had their baseline test while on chemo, or within a few months upon completion of chemo.

It's not death events, it's neurocogition events.
Less than 10% were dead at 3 months. And progression also doesn't seem that bad.
View attachment 424496

And it's likely also not brain mets. At 3 months the rates are still not that high.
View attachment 424495

I fully understand that PCI will produce neurocognitive decline at 3 months, but why is the surveillance curve also dropping that hard at 3 months? One would imagine that patients who are done with chemo, will actually have stable (if not even better) neurocognition. But a 50% drop?

Just a thought...
Click to expand...

lot of deaths in both arms, which counts for this endpoint also
 
Palex80 said:
Could someone explain to me, why at 3 months both groups drop by >50% compared to baseline?
Click to expand...
I agree as above it's not death making the big contribution to this early drop.

I don't know of course, but I suspect this is just what the process of being treated for SCLCa does. The other hypothesis is that the testing in bunk and they are assigning cognitive decline wrongly...I think this is unlikely.

In words, the graph is telling us... SCLC patients experience measurable cognitive decline or death roughly 50% of the time within 3 months of completion of chemotherapy when treated with chemotherapy, thoracic chemoradiotherapy +/- IO without administration of PCI, which itself seems to confer a further 50% relative risk of decline. This new baseline further degrades on average over time.

The testing is sensitive enough to pick up decline associated with non-CNS targeted therapy and from diagnosis itself. This is notable and indicative of the cognitive cost of chemotherapy or just cancer diagnosis IMO.

I wouldn't get too granular with the data as a clinician. Both chemo and brain radiation make you dumber, PCI never makes you smarter, PCI reduces intracranial progression risk but at a cost of decreased cognition beyond expected from chemo and without a PFS (or anticipated OS) benefit.

From my personal experience...I am frankly embarrassed that I offered WBRT (PCI or in different metastatic settings) to so many patients earlier in my career. My perception is that it clearly degrades most patients substantially and I have used it only as a last result for years. I think some young patients with Her2 positive bCa refractory to targeted therapy have probably benefited in terms of OS.

It is possible IMO that the decrease in performance status associated with WBRT negatively interacts with adjuvant IO or causes more hospitalizations leading to more sensitive discovery of systemic failure. PCI clearly reduces brain failure by about 15%, but this is not manifested as PFS (which is interesting).

I can't imagine anyone justifying PCI based on this data.
 
Their neurocognitive testing must be so good to pick up damage caused by 25 Gy / 10 ( plus some of which was HA). Hard to notice this clinically.
 
Palex80 said:
Could someone explain to me, why at 3 months both groups drop by >50% compared to baseline?
Click to expand...
It’s death and or cognitive decline from treatment and or subacute limbic encephalopathy (as a paraneoplastic thing) and or brain mets and or PCI and or SRS. May have missed a few and ors.
communitydoc13 said:
I don't know of course, but I suspect this is just what the process of being treated for SCLCa does
Click to expand...
I tend to agree. And or just having SCLC. I probably would experience some/any cognitive decline from going through thoracic chemoRT.
FletchersGarb said:
Their neurocognitive testing must be so good to pick up damage caused by 25 Gy / 10 ( plus some of which was HA). Hard to notice this clinically
Click to expand...
“Any” cognitive decline will be sensitive measure and “so good” to pick up 25/10 versus no PCI. Probably.
 
Last edited:
FletchersGarb said:
Their neurocognitive testing must be so good to pick up damage caused by 25 Gy / 10 ( plus some of which was HA). Hard to notice this clinically.
Click to expand...

I don't know about that in my experience. 25/10 after coming off 6 cycles of carbo/etoposide is another hit. I have noted it clinically at least in terms of KPS or QoL decline, though I'm obviously not diving deeply into their neurocognitive function.

Sure, no major dementia but if you feel fatigued and don't have your pep I'd guess that bleeds into a neurocognitive test.
 
communitydoc13 said:
I agree as above it's not death making the big contribution to this early drop.

I don't know of course, but I suspect this is just what the process of being treated for SCLCa does. The other hypothesis is that the testing in bunk and they are assigning cognitive decline wrongly...I think this is unlikely.

In words, the graph is telling us... SCLC patients experience measurable cognitive decline or death roughly 50% of the time within 3 months of completion of chemotherapy when treated with chemotherapy, thoracic chemoradiotherapy +/- IO without administration of PCI, which itself seems to confer a further 50% relative risk of decline. This new baseline further degrades on average over time.

The testing is sensitive enough to pick up decline associated with non-CNS targeted therapy and from diagnosis itself. This is notable and indicative of the cognitive cost of chemotherapy or just cancer diagnosis IMO.

I wouldn't get too granular with the data as a clinician. Both chemo and brain radiation make you dumber, PCI never makes you smarter, PCI reduces intracranial progression risk but at a cost of decreased cognition beyond expected from chemo and without a PFS (or anticipated OS) benefit.

From my personal experience...I am frankly embarrassed that I offered WBRT (PCI or in different metastatic settings) to so many patients earlier in my career. My perception is that it clearly degrades most patients substantially and I have used it only as a last result for years. I think some young patients with Her2 positive bCa refractory to targeted therapy have probably benefited in terms of OS.

It is possible IMO that the decrease in performance status associated with WBRT negatively interacts with adjuvant IO or causes more hospitalizations leading to more sensitive discovery of systemic failure. PCI clearly reduces brain failure by about 15%, but this is not manifested as PFS (which is interesting).

I can't imagine anyone justifying PCI based on this data.
Click to expand...

I had basically given it up for extensive stage.

I now feel comfortable giving it up in limited stage...though I really wasn't pushing it much at all in the last 5 years.

I did enroll on the hippocampal avoidance PCI RTOG trial...but obviously that isn't going to be a thing anymore.
 
BobbyHeenan said:
I don't know about that in my experience. 25/10 after coming off 6 cycles of carbo/etoposide is another hit
Click to expand...
Oh I concur. However I was simply responding to his question if there is good cognitive testing to pick up the cognitive effects of 25/10. Evidently per Maverick there is.

Man I remember the era when Anna Gregor and Drew Turrisi were using 36/20 for PCI because her data was suggesting that was the superior schedule.
BobbyHeenan said:
I did enroll on the hippocampal avoidance PCI RTOG trial...but obviously that isn't going to be a thing anymore
Click to expand...
Well sure it is. No PCI is the ultimate HA 😉

hahaha
 
Advertisement - Members don't see this ad
thesauce said:
I don’t think it’s dead:

www.cancernetwork.com

Prophylactic Cranial Irradiation Improves Survival in Limited-Stage SCLC | CancerNetwork

In a real-world CASCADE database analysis, prophylactic cranial irradiation was associated with a median overall survival of 29.8 months in limited-stage SCLC.
www.cancernetwork.com www.cancernetwork.com
Click to expand...
"My words fly up. My thoughts remain below. Words without thoughts never to heaven go." - Hamlet III iii

It doesn't really matter what you, me, or all rad onc thinks. It matters what med onc thinks. So get back to me on that in a year or so.
 
Palex80 said:
Ok, let me kick off the discussion then.

I saw (most of) the slides, and this is one slide, I don't quite understand.
View attachment 424494


So, the primary endpoint is CFFS meaning you have to show a decline in cognition (based on tests: Hopkins Verbal Learning Test–Revised (HVLT-R), Controlled Oral Word Association (COWA) Test, Trail Making Test (TMT Parts A and B) compared to your baseline test or die. Everyone starts at 100%, the endpoint is measuring decline, not your test result as an absolute mark.

Could someone explain to me, why at 3 months both groups drop by >50% compared to baseline?

Randomisation was done within 16 weeks of the last chemo cycle, so these patients had their baseline test while on chemo, or within a few months upon completion of chemo.

It's not death events, it's neurocogition events.
Less than 10% were dead at 3 months. And progression also doesn't seem that bad.
View attachment 424496

And it's likely also not brain mets. At 3 months the rates are still not that high.
View attachment 424495

I fully understand that PCI will produce neurocognitive decline at 3 months, but why is the surveillance curve also dropping that hard at 3 months? One would imagine that patients who are done with chemo, will actually have stable (if not even better) neurocognition. But a 50% drop?

Just a thought...
Click to expand...
Look at the patient numbers at each time point. Eg in the Cum Inc of Brain mets at 12 mo have 64 and 66. In the neurocognitive testing it is 14 and 5. Likely patients just didn't show for the neurocog testing and were excluded from this analysis.
 
thesauce said:
I don’t think it’s dead:

www.cancernetwork.com

Prophylactic Cranial Irradiation Improves Survival in Limited-Stage SCLC | CancerNetwork

In a real-world CASCADE database analysis, prophylactic cranial irradiation was associated with a median overall survival of 29.8 months in limited-stage SCLC.
www.cancernetwork.com www.cancernetwork.com
Click to expand...
The best retrospective analysis cannot account for all the hidden variables that make some patients get an intervention while others do not. Experiments matter.


temujim said:
Look at the patient numbers at each time point. Eg in the Cum Inc of Brain mets at 12 mo have 64 and 66. In the neurocognitive testing it is 14 and 5. Likely patients just didn't show for the neurocog testing and were excluded from this analysis.
Click to expand...
I think this represents prior events/censoring. By 12 months, there are very few remaining at risk patients evaluable regarding cognitive decline or death endpoint.

Cognitive decline happened earlier and more frequently than brain mets.
 
thesauce said:
I don’t think it’s dead:

www.cancernetwork.com

Prophylactic Cranial Irradiation Improves Survival in Limited-Stage SCLC | CancerNetwork

In a real-world CASCADE database analysis, prophylactic cranial irradiation was associated with a median overall survival of 29.8 months in limited-stage SCLC.
www.cancernetwork.com www.cancernetwork.com
Click to expand...
Retrospective data cannot and should not trump a randomized controlled trial

A real world study is just a fancy made up term for retrospective data.

Additionally - I just cannot take seriously a paperthat STILL INCLUDES SCLC PATIENTS NOT RECEIVING BRAIN MRI AT DIAGNOSIS PUBLISHED in 2026
In patients with data on staging brain imaging, those receiving PCI were less frequently staged with brain MRI (60.9% versus 73.5%, p = 0.014)
Click to expand...
1789749998747.png

35% of patients did not undergo MRI brain staging in this retrospective analysis.

Yes, they did a subgroup analysis of those who DID get MRI and still saw a OS benefit. But, it's still retrospective data and does not refute a clinical trial.

I have been discussing MRI surveillance for LS-SCLC pts for years. This reinforces my practice, and thus is good, for me. If you've routinely been doing PCI, really consider whether it's actually benefiting patients, as long as you get them a freaking MRI to stage them at diagnosis.
 
evilbooyaa said:
Retrospective data cannot and should not trump a randomized controlled trial

A real world study is just a fancy made up term for retrospective data.

Additionally - I just cannot take seriously a paperthat STILL INCLUDES SCLC PATIENTS NOT RECEIVING BRAIN MRI AT DIAGNOSIS PUBLISHED in 2026

View attachment 424701
35% of patients did not undergo MRI brain staging in this retrospective analysis.

Yes, they did a subgroup analysis of those who DID get MRI and still saw a OS benefit. But, it's still retrospective data and does not refute a clinical trial.

I have been discussing MRI surveillance for LS-SCLC pts for years. This reinforces my practice, and thus is good, for me. If you've routinely been doing PCI, really consider whether it's actually benefiting patients, as long as you get them a freaking MRI to stage them at diagnosis.
Click to expand...
Pci was therapeutic in many cases hence the OS benefit. That being said I feel like the ones that have refused it with me and don't comply with MR surveillance invariably end up in the hospital with neuro sx
 
You must log in or register to reply here.

Similar threads

TheWallnerus
Evicore to declare benign radiotherapy DEAD
Replies
45
Views
4K
grenz
G
Y
  • Poll Poll
AI is Evil
Replies
28
Views
3K
temujim
T
TheWallnerus
Skincure is suing CMS
Replies
9
Views
2K
NotMattSpraker
N
D
Is this just the beginning?
Replies
14
Views
2K
WildRivers
W
Beerus
Is this the Varian/Elekta killer?
Replies
39
Views
4K
drowsy12
D