PCI in SCLC

[scarbrtj]

---

Members don't see this ad.

Our opinions seem to differ on this, but to me there is a lot more to 'service' than just OS. Prevention of hospitalization, surgery, morbidity, symptomology, etc. I can't read that article you posted - got a pubmed link?

Abolishing Mammography Screening Programs? A View from the Swiss Medical Board

 

[medgator]

I give PCI to all LS-SCLC, and all CR/PR patients in ES-SCLC, and I give chest RT to ES-SCLC patients who have PR/CR as well

That's pretty much my practice these days as well. A few years ago, I wasn't doing consolidation chest RT, but after seeing that patient, I realized in vivo that it was something I should be doing (considering that was about the time that the Slotman study for TRT in ES-SCLC that you linked to was published).

 

[Palex80]

I'm sorry, I don't understand that point.
You say that if you pick up 1-2 lesions BEFORE the patients develops a dozen more, you are saving the patient from WBRT?
I doubt that. Surely you can treat those 1-2 lesions, but the other dozen is still going to pop up.

What do I tell my 4+ year es sclc survivor who had a CR after initial EP, PCI and then went on to chemoradiation for salvage for isolated mediastinal recurrence 2 years ago?

Pretty sure the pci eradicated micromets intracranially as well as the carbo/VP-16 did extra-cranially

n=1 is not evidence based medicine.
Show him the charts of the 200 ES-SCLC you treated and died within a year. He will be grateful he is alive.

 

[medgator]

I'm sorry, I don't understand that point.
You say that if you pick up 1-2 lesions BEFORE the patients develops a dozen more, you are saving the patient from WBRT?
I doubt that. Surely you can treat those 1-2 lesions, but the other dozen is still going to pop up.
So unless you are planning to doing a dozen SRS over the course of several month

n=1 is not evidence based medicine.
Show him the charts of the 200 ES-SCLC you treated and died within a year. He will be grateful.

That still does not prove your absolution that es-sclc cant be effectively "cured" with a combination of chemo therapy and brain irradiation. Vp16 gets into the brain and between that and xrt, it may be just enough to sterilize micromets in the brain, while the platinum and vp16 do the same, extracranially in a small subgroup of patients

Sent from my SAMSUNG-SM-N910A using SDN mobile

 

[Palex80]

That still does not prove your absolution that es-sclc cant be effectively "cured" with a combination of chemo therapy and brain irradiation. Vp16 gets into the brain and between that and xrt, it may be just enough to sterilize micromets in the brain, while the platinum and vp16 do the same, extracranially in a small subgroup of patients

Sent from my SAMSUNG-SM-N910A using SDN mobile

Theoretically, yes. There are about 1-5% of ED-SCLC patients that seem to be "cured" when you look at the charts of published studies 5 years down the road.
That's what we have as solid evidence.
I do however think that we should drive our treatment decisions based on what's the best treatment for 95-99% of the patients and not 1-5% of them.

Otherwise we should be probably be irradiating glioblastomas with a BED well above 100 Gy. There are some Phase II protocols frok Colorado and Japan showing long term survival with approaches like these for a few selected patients.

We should also be offering PCI to all triple negative locally advanced breast cancer patients. I am sure many of you have seen a triple negative breast cancer patient with aggressive local and systemic treatment, in complete remission and then developing an isolated brain failure with multiple mets or meningeal spread and dying of this.
We may have cured this patient with PCI up-front. What is the NNT to achieve that? Probably hundreds...

 

[evilbooyaa]

I'm sorry, I don't understand that point.
You say that if you pick up 1-2 lesions BEFORE the patients develops a dozen more, you are saving the patient from WBRT?
I doubt that. Surely you can treat those 1-2 lesions, but the other dozen is still going to pop up.


n=1 is not evidence based medicine.
Show him the charts of the 200 ES-SCLC you treated and died within a year. He will be grateful he is alive.

Assuming the bolded is in reference to me.

If you see 1 to 3 (or whatever your number is) brain mets (say on initial screening scan), do you treat them with SRS or do you just observe them until they become symptomatic, cause the patient might need whole brain in the future anyways?

There are patients who undergo SRS (either once, or a few times over the years), and never require whole brain. Obviously there are some that eventually do require whole brain, but that number is not 100%. If they had uncontrolled intracranial disease just sitting in their brain would they be more likely to have those lesions grow and seed the rest of the brain?

The Yale data, both single and multi-institutional, shows us that if you treat with EGFR-TKI (which is more than observation) those patients still live shorter than if you do SRS up front.
With 22 months follow-up - Of 100 patients treated with SRS first, followed by TKI, 25% required repeat treatment with SRS, and 20% of patients required whole brain (either alone or in conjunction with SRS).

So yes, this data is a bit biased because EGFR-TKI does have some activity in the brain, but only 20% of these patients that got SRS first needed WBRT, with median f/u of almost 2 years. If you're still skeptical I'll try and find data outside of the EGFR-TKI data.

 

[medgator]

Nsclc...

DEFINE_ME

 

[FrostyHammer]

Nsclc...

DEFINE_ME

Can't access whole paper, but seems like similar issues as POP RT we discussed in the other forum with small sample sizes. What is up with that gigantic OS difference?

 

[medgator]

Can't access whole paper, but seems like similar issues as POP RT we discussed in the other forum with small sample sizes. What is up with that gigantic OS difference?

Yeah seemed kinda odd, but even if a fraction of it is real, definitely something to think about it

 

[Lamount]

"It is important to also consider that patients in our clinical setting do not generally have access to third generation TKIs for the treatment of EGFR-mutated and ALK-rearranged NSCLC. Therefore, considering the efficacy of these newer agents in CNS disease, the results presented must be further appraised with data acquired from larger studies, which include patients treated with osimertinib, alectinib among others, which have proven to decrease the development of BM in the early and advanced setting."

This is a major consideration. I don't really think you can do anything with these data for EGFR/Alk patients who are naive to Osi/Alectanib

 

[medgator]

"It is important to also consider that patients in our clinical setting do not generally have access to third generation TKIs for the treatment of EGFR-mutated and ALK-rearranged NSCLC. Therefore, considering the efficacy of these newer agents in CNS disease, the results presented must be further appraised with data acquired from larger studies, which include patients treated with osimertinib, alectinib among others, which have proven to decrease the development of BM in the early and advanced setting."

This is a major consideration. I don't really think you can do anything with these data for EGFR/Alk patients who are naive to Osi/Alectanib

What percentage of your practice is egfr/alk+ to begin with? Not much for me

 

[Lamount]

What percentage of your practice is egfr/alk+ to begin with? Not much for me

Maybe 10% of my practice.

...but this study was specifically looking at "high risk NSCLC" -i.e. a little more than 80% (66/81) of the patients on study were EGFR/ALK positive.

Only 13 non-mutant patients were included on the trial... as another indication for enrollment besides mutation was CEA elevation (which is news to me as a risk factor)... thus I would argue that these data don't apply to most of my patients...

...and its use in mutation driven lung cancer is questionable given the lack of third generation TKIs -which have excellent CNS penetration.

 

[TheWallnerus]

Can't access whole paper, but seems like similar issues as POP RT we discussed in the other forum with small sample sizes. What is up with that gigantic OS difference?

Yeah seemed kinda odd, but even if a fraction of it is real, definitely something to think about it

Definitely **think** about it. Thinking never hurt anybody. Don't alter SOC on account of it. Somebody here today on SDN said something very smart & catchy: "not too good to be true, just too good to be replicated." We have walked down this PCI road a million times in NSCLC it seems thus far and the street sign says Via Dolorosa. To this study's authors I say: cool story bros.

 

[thecarbonionangle]

Control arm had less actionable mutations than experimental arm. Old tkis which dont penetrate CNS as well. Old treatment with no immunotherapy. Small numbers not there to explain large difference. We already know from RTOG and another study it reduces cranial progression.

Not much to see here folks

 

[evilbooyaa]

Control arm had less actionable mutations than experimental arm. Old tkis which dont penetrate CNS as well. Old treatment with no immunotherapy. Small numbers not there to explain large difference. We already know from RTOG and another study it reduces cranial progression.

Not much to see here folks

This was what I came here to say. Imbalance in actionable mutations (95% in expermiental arm had EGFR or ALK mutation vs only 75% in control arm). Not surprised that EGFR/ALK patients treated with any generation TKI had better OS than the alternative. Trying to 'control' for this on multivariate analysis is admirable, but not equivalent.

When they looked at subsets that only had actionable mutations, P value went above 0.05:
"Nonetheless, the OS benefit was not significant, though numerically longer, when assessing only patients with oncodriver mutations (EGFRm and ALKr). In this case OS was 64.5 months (95%CI: 21.3–107.7) vs. 28.4 (95%CI: 22.9–33.8) months, among patients in the experimental and observation arms, respectively (HR: 0.54 [95%CI: 0.27–1.09]; p=0.089) (Figure 3B)"

I do wonder if NSCLC patients should get regular screening MRIs, as what proportion of these patients showed up with large symptomatic BMs that ended up actually affecting OS. I was used to screening brain MRI 1x/year for otherwise metastatic NSCLC patients from my med-oncs - anyone doing this routinely in their stage IV (or even post-CRT stage III) patients?

 

[TheWallnerus]

And this is what not doing PCI after local therapy for SCLC gets you

 

[Palex80]

Sorry to resurrect this topic. I was looking at the MAVERICK trial and the question popped up:

Why do we need a randomized trial for this question as a first step?

I fully understand that randomized trials are the best way to prove or disprove something, but the experimental arm has never been properly examined in a non-randomized manner. The experimental arm (omission of PCI) is therapy deescalation, s.o.c. is to deliver PCI in LD-SCLC.
There are however a number of questions that are still open, for instance what is the rate of isolated brain recurrence in patients who do not receive PCI and have PR/CR at after CRT for LD-SCLC? How many of these patients reccur with one lesion vs. multiple lesions? What is the optimal regime to observe those patients with MRI of the brain (MRI every 3 or 6 months or wait for symptoms?).

We have done single-arm deescalation trials before, think of the single arm trials testing no adjuvant RT in DCIS of the breast, for instance this one.


Wouldn't it have been wiser to first conduct a Phase-II single-arm trial testing MRI-surveillance instead of PCI in LD-SCLC with PR/CR after CRT?
We know that 25-35% of all LD-SCLC are disease-free and alive at 5 years with Turrisi-fractionation full CRT + PCI (depends on which graph you look - Turrisi vs. CONVERT and there's certainly the effect of stage migration that made CONVERT results look better).

Now let's do that regime, but simply drop the PCI.

Imagine such a trial would show a 20+% rate of isolated brain recurrence. I couldn't find data for isolated brain recurrence following PCI for LD-SCLC, but I presume it's low, maybe even <5%.

Woulnd't that be enough to "bury" a randomized trial trying to prove non-inferiority of deferring PCI? I don't think anyone believes you can still "rescue" those excessive 15+% patients recurring with brain mets using SRS +/- WBRT. You may be able to salvage some, but in the end half of those patients who didn't get PCI and recurred in the brain are going to die because of those brain mets, which basically means the test regime would result in at least 7% worse survival than s.o.c.

I am hearing some of you saying it's tough to randomize patients in MAVERICK, so perhaps this could have been an interesting approach.
It would definetely need less patients and accrual would be easier (since many patients, as it appears, do not want PCI anyhow!). "Oh, you do not want PCI? That's great, we have just the right trial for you! I can guarantee you, you are not getting PCI."

And sure, if the rate of isolated brain recurrence would be very low (<10%) without PCI, then you don't even need the randomized trial. You can make the point that PCI shouldn't be done anymore, since few patients reccur anyhow and by salvaging them, the number-needed-to-treat with PCI is just to high to justify it.

The only "problem" would arise if the rate of isolated brain recurrence without PCI was somewhere around 15%, you would then probably need a randomized trial to prove non-inferiority.

I fear the MAVERICK trial may be rushing things a bit...

 

[Palex80]

I do wonder if NSCLC patients should get regular screening MRIs, as what proportion of these patients showed up with large symptomatic BMs that ended up actually affecting OS. I was used to screening brain MRI 1x/year for otherwise metastatic NSCLC patients from my med-oncs - anyone doing this routinely in their stage IV (or even post-CRT stage III) patients?

Good points. We don't screen so far; but you are absolutely correct, it may be worth doing it!

 

[Radonc90]

It is now 2021...

Anyone here does frequent MRIs?
- At Diagnosis
- Post Chemo-RT: how often? Thin cuts 1-mm vs standard 3-mm cuts?

The idea is to spare the pt from WBRT, if possible...

 

[BobbyHeenan]

It is now 2021...

Anyone here does frequent MRIs?
- At Diagnosis
- Post Chemo-RT: how often? Thin cuts 1-mm vs standard 3-mm cuts?

The idea is to spare the pt from WBRT, if possible...

I struggle with this too - how often to get MRIs in stage 3 nsclc. Obviously up front, but if asymptomatic when to do again?

I typically push for re staging brain any time they progress extracranially.

 

[evilbooyaa]

@Palex80 I get what you're saying, but the issue is that there will be too much disagreement on what is an 'acceptable' number or even 'acceptable' endpoint to a phase II trial. And there will be a subset of Rad Oncs who say "there's no phase III data, so I will still PCI' if you just run a phase II. Rad Oncs as a field have to be bashed in the head with ph III trials (see all the criticism of SABR-COMET for being a randomized ph II) for pretty much anything that is not peds.

Maybe this will be like CSI dose reduction in average risk medullo, but I more imagine it to go the way of the Takahashi trial.

I do worry that the MAVERICK trial is asking too many questions - I don't know the utility of re-doing Takashahi (for ES-SCLC patients), but I suppose they want to validate results.

I was looking to see if anything had been published, and found this retrospective series:

The value of prophylactic cranial irradiation in limited-stage small cell lung cancer: should it always be recommended? - PMC

Prophylactic cranial irradiation (PCI) is a standard treatment for limited-stage small cell lung cancer (LS-SCLC) showing a response to initial treatment, but many patients do not receive PCI due to comorbidities or refusal. This study aims to ...

www.ncbi.nlm.nih.gov

The issue is that there was no MRI surveillance in either group. That is where most proponents of no-PCI see the failure in every single study that has espoused benefits of PCI.

 

[Radonc90]

I just saw this study Scandinavian dose escalation trial: although the 45 Gy BID (À la Turrisi et al) is not as good as 60 Gy BID in terms of median survival but at 4 yrs/48 months it looks the same to me...

Just a deadly disease, no matter what the early outcome is (median survival), as time goes on survival keeps dropping down at 4-5 years...

PS: responders offered PCI by the way...

 

[FrostyHammer]

I just saw this study Scandinavian dose escalation trial: although the 45 Gy BID (À la Turrisi et al) is not as good as 60 Gy BID in terms of median survival but at 4 yrs/48 months it looks the same to me...

Just a deadly disease, no matter what the early outcome is (median survival), as time goes on survival keeps dropping down at 4-5 years...

PS: responders offered PCI by the way...

Not sure this will be relevant when (not 'if') immunotherapy for LS SCLC is proven

 

[Palex80]

I just saw this study Scandinavian dose escalation trial: although the 45 Gy BID (À la Turrisi et al) is not as good as 60 Gy BID in terms of median survival but at 4 yrs/48 months it looks the same to me...

Just a deadly disease, no matter what the early outcome is (median survival), as time goes on survival keeps dropping down at 4-5 years...

PS: responders offered PCI by the way...

It is a small trial and it is a Phase II.
Despite that, there are differences observed and the trial does support the concept of isotoxic dose escalation. This has been also attempted in stage III NSCLC. Isotoxic dose escalation means you simply use the same contraints for both dose levels and try to get as much dose as possible in the tumor in the dose escalated arm. It can be beneficial and may be the approach, bearing in mind that non-isotoxic dose escalation as already tested in stage III NSCLC, did not work.

Does the Norwegian trial mean we should be treating all LD-SCLC with dose escalation? I do not know.
But bearing in mind that more than a third of the progressions in LD-SCLC do occur locally only even with 45 Gy bid, it is something to consider.
66 Gy delivered daily are also biologically "more" than 45 Gy bid in terms of BED, but CONVERT failed to show any benefit with 66 Gy delivered daily.
So one could hypothesize that the way to go forward is accelerated radiation therapy looking at isotoxicity. CONVERT also showed that 66 Gy delivered daily were more toxic than 45 Gy bid, supporting this idea.

The whole question of how immunotherapy may make this data irrelevant is also open. Immunotherapy does appear to be active in SCLC, but it's not like it's dramatically improving results in ES-SCLC and is certainly no s.o.c. yet in LD-SCLC. Trials are running and results are awaited.