NEJM case report of central airway necrosis after 50 Gy/5 Fx

[RadSki]

http://www.nejm.org/doi/full/10.1056/NEJMc1203770

Curious what the preliminary safety profile is from RTOG 0813.

---

Members don't see this ad.

 

[Gfunk6]

When studying for oral boards, the latest word I heard from RTOG 0813 was that they have escalated to 12 Gy x 5 which means that 11.5 Gy x 5 was deemed safe.

I don't really see why we should use case studies for evidence when we have good retrospective data and an on-going single-arm, prospective dose-escalation trial. I mean you can make a case study of very poor breast cosmesis after only 50 Gy or brain necrosis after 15 Gy x 1 of SRS, but what would that prove when there are literally decades of randomized data? When your numerator and denominator are both one, I would not call this robust data.

 

[deleted4401]

Penn hates SRS/SBRT. GFunk right. This is a safe, validated treatment. Lobectomies can't be performed on everybody. Definitive EBRT has terrible results with poor LC and poor OS. Gotta be careful with everything we do, but SBRT is a a great treatment.

EDIT: also, remember NEJM hates RT! They'll publish a case report of this nature rather than something that shows RT works.

 

[seper]

Disagree. I don't consider 10 Gy X 5 for central lesions to be safe (not enough data). I have been treating central lesions to 9Gy X 5 qd and this letter validates my concerns.

 

[medgator]

There's always that 1-5% risk of normal tissue toxicity, even when we follow tissue tolerance/constraints. Hence the idea of the TD 5/5, 50/5 concepts. It's the risk/benefit ratio we weigh for each patient to obtain maximal tumor control at the price of the small risk for toxicity. That's why we never quote a 0% risk of long-term toxicity. Gfunk alluded to this above.

Agree with the rest of the posts on this thread. I'll take retrospective data and an ongoing prospective study any day of the week over some case report in the NEJM

 

[Scrybe9]

Penn hates SRS/SBRT.

First off, this is an absurd statement.

While the protocol does not specify the number of patients per dose level it is likely on the order of 10-20, so adding an additional n of 1 to that is meaningful. This is a single-arm trial of safety and efficacy. The whole point is to identify these issues before they are widely adopted so that our entire community can practice in a manner that improves patient care with minimal harm. If that means pointing out a potentially dangerous practice that is being widely adopted without extensive safety data, so be it.

Think about the analogy from therapeutic drugs that make their way through phase II/III only to be found to be unsafe when released to a mass audience. Early adoption of technology serves to drive our field, but never underestimate the harm that we can cause with the beam.

 

[deleted4401]

Fine. It's an absurd statement. I've been known to make those. I retract it.

But, this is the NEJM. To bring a toxicity that is seen and expected in 5% or less cases and report it like that doesn't seem to add value. There are trials specifically looking at this question. If 3 out of the first 6 patients had this toxicity, sure, I get the point of a case report in the world's leading journal. But, when it comes to radiation, all they want to report is toxicity, inefficacy, and high cost of treatment. Sigh ... I bet within one week, Walt Bogdanich from the New York Times will write about this, and then discuss all of our snake-oil techniques, and why rad-oncs are just money-grubbing technicians.

 

[Gfunk6]

While the protocol does not specify the number of patients per dose level it is likely on the order of 10-20, so adding an additional n of 1 to that is meaningful.

No, it's not. In fact, it's irrelevant. Those 10-20 patients were carefully selected, screened and met rigorous inclusion/exclusion criteria with meticulous QA hardware requirements BEFORE they were treated.

The Penn case report is just a random patient culled from nowhere. Would she have even met inclusion criteria on RTOG 0813? Did they follow RTOG 0813 dose constraints precisely? The paper says they did, but who knows. Did they actually treat the peripheral lesion and central lesion simultaneously?

 

[Scrybe9]

No, it's not. In fact, it's irrelevant. Those 10-20 patients were carefully selected, screened and met rigorous inclusion/exclusion criteria with meticulous QA hardware requirements BEFORE they were treated.

From the article: "The patient was treated with SBRT in accordance with a protocol for a registration study that allows for long-term surveillance of adverse events; the protocol was approved by an institutional review board. Dose, fractionation, technique, and constraints were established and applied in accordance with published standards (citing RTOG 0813)"

So what you are saying is that you have more faith in an RTOG protocol being applied at a wide variety of local institutions than a seemingly identical protocol at a single institution, which also happens to be an RTOG member?

I'm a little surprised how shocked people are at the presentation of a toxicity, which is well-established with previous fractionation schedules, which was now seen at a lower threshold. What would you have them do, not report something they believe to be clinically important?

 

[seper]

We can accept 5% rate of serious (grade 3) toxicity - TD5/5 concept.
We should not accept even 1% rate of grade 5 toxicity (death). That's what this case report is about.

Also, now since it is published, SBRT community should consider itself informed of risks.

 

[medgator]

I'm a little surprised how shocked people are at the presentation of a toxicity, which is well-established with previous fractionation schedules, which was now seen at a lower threshold. What would you have them do, not report something they believe to be clinically important?

Did they report this to RTOG? Or did they decide to have their 5 mins of fame by creating a case report and publishing it in the NEJM? Did they try to tease out the etiology of the hemoptysis as this patient had multiple mediastinal recurrences/mets along with the airway injury. It seems like there are too many possible factors here, and a short case report really isn't doing this any justice

 

[medgator]

Did they follow RTOG 0813 dose constraints precisely? The paper says they did, but who knows. Did they actually treat the peripheral lesion and central lesion simultaneously?

They did treat both simultaneously. There's a picture of the dose color wash of both lesions at the link (figure 1). Interesting how they gloss over this. Also this patient had multiple mediastinal recurrences/mets along with the airway injury. They don't really go into this issue much in the paper, yet I (along with most I would assume) would think this to be an important issue.

 

[deleted4401]

Dude, this patient would have never been on trial. She had two ipsilateral tumors - one in periphery and one central. It's T3 disease. Only specific T3 disease was treated on protocols, and it was not patients with 2 tumors in one lung. Plus, the patient had progressive disease. Think maybe that had something to do, as well, with her demise? So, I officially don't think this is reasonable. If you're going to treat two primaries at once, (i.e. locally advanced disease) do it at your peril.

Anyway, without treatment for stage I-II disease, median survival ranges from 8 months to 13 months.
With definitive RT, it's 1 year and 8 months.
On RTOG 0236, the median survival was 48 months, even with the toxicity!
There are risks of treatment and you have to go through it with the patient, but seriously, what else would you offer a patient? Of course we should have caution, of course we should notify patients of risks of treatment. But it would be unethical to treat a patient with early stage lung cancer any other way,

S

 

[Gfunk6]

They did treat both simultaneously. There's a picture of the dose color wash of both lesions at the link (figure 1). Interesting how they gloss over this. Also this patient had multiple mediastinal recurrences/mets along with the airway injury. They don't really go into this issue much in the paper, yet I (along with most I would assume) would think this to be an important issue.

Thus, this patient would not have qualified for RTOG 0813. In the community and in academics, we frequently use dose constraints and treatment guidelines for our own patients off protocol. This is in no way, shape or form even remotely equivalent to actually having the patient on a clinical trial.

So what you are saying is that you have more faith in an RTOG protocol being applied at a wide variety of local institutions than a seemingly identical protocol at a single institution, which also happens to be an RTOG member?

Are you really asking me if I have more faith in a multi-center, prospective trial from our nation's Radiation Oncology cooperative group or a single institution case report? If so, I'll go with the RTOG.

What would you have them do, not report something they believe to be clinically important?

People can and do publish whatever the heck they want. Our job is to look at data with a critical eye and proceed accordingly.

 

[Scrybe9]

There are risks of treatment and you have to go through it with the patient, but seriously, what else would you offer a patient? Of course we should have caution, of course we should notify patients of risks of treatment. But it would be unethical to treat a patient with early stage lung cancer any other way, S

People can and do publish whatever the heck they want. Our job is to look at data with a critical eye and proceed accordingly.

I agree with both of you on these points. I have no issue with treating with SBRT and believe it to be the lesser of several evils for patients with an otherwise fatal disease. Not knowing the actual details of this case it's hard to argue the finer points either way. I will say, however, that it is nice to see a debated topic on here instead of just as a information source for applying students.

 

[medgator]

Of course we should have caution, of course we should notify patients of risks of treatment. But it would be unethical to treat a patient with early stage lung cancer any other way,

S

Not sure I'd go that far and say that.

Obviously SBRT is preferred in medically-inoperable lung CA, but postage stamp with 70/35 or, ideally, hypofractionation is still reasonable in situations where SBRT may not be available or feasible (i.e. patient can't tolerate 30-40 min treatment for 3-4 sessions, Facility only has Cyberknife SRS and patient can't have fiducials placed, etc.). That's why NCCN recommends "definitive RT or SABR" (lol, I can't get over that picture of Timmerman in jedi gear) in early-stage medically-inoperable patients, and allows for conventional or hypofractionated 3DCRT at places without a formal SABR/SBRT program (stated in the RT section of the NCCN guidelines).

We have a CK where I'm at right now, and I had a patient who couldn't get fiducials placed (Hx of multiple iatrogenic PTX's in the past) and I ended up doing a hypofractionated regimen we used a lot in residency for non-SBRT cases ---> 70 Gy in 28 Fx

 

[Seldon1985]

Actually, looks Figure 1 caption states that there was a right lower lobe mass AND a right upper lobe mass - making this a T4 (separate tumor nodule(s) in a different ipsilateral lobe).

 

[deleted4401]

I've never done 70 Gy/28 Fx. Just looked over the article. Looks like a really good scheme. I'd still prefer Cyber/SBRT, though, just because of so much accumulated data from Japan, here, and other places.
-S

 

[medgator]

I've never done 70 Gy/28 Fx. Just looked over the article. Looks like a really good scheme. I'd still prefer Cyber/SBRT, though, just because of so much accumulated data from Japan, here, and other places.
-S

We used it a lot before we got into SBRT (and still used it for central tumors that were not amenable to 3-4 Fx Tx when I graduated). The control rates were pretty good, and we often went up to the 20 Fx scheme (70 at 3.5/fx) as I trained with the PI. I'm not doing that in the community though

My preference is still SBRT, of course, but I offer accelerated 3DCRT when it's not feasible to do it.

 

[RadSki]

Thanks for all the comments. I agree that we shouldn't change practice based on n=1. It's important to be aware of what how referring docs perceive our specialty. This case report got coverage in the ASCO daily news digest and other news sources, so med oncs, surgeons, and patients will hear about this. Unfortunately, this is the second case report or similar publication I've seen in the NEJM of a radiation toxicity since I started residency (see http://www.nejm.org/doi/full/10.1056/NEJMicm0810650). We should start submitting case reports of patients treated successfully for epilepsy with SRS, chemoRT for bladder preservation, anal cancer spared APR, etc. Would be nice to see some good radiotherapy press in the NEJM.

Hopefully we get some good data from 0813 that will allow us to move into a larger phase III trials to better delineate the control and toxicity rates for central tumors. -RS

 

[SchwettieBalls]

We can accept 5% rate of serious (grade 3) toxicity - TD5/5 concept.
We should not accept even 1% rate of grade 5 toxicity (death). That's what this case report is about.

This comment is ridiculous. We frequently subject patients with potentially lethal cancer to a risk of fatal treatment related toxicity on the order of 1-5% (essentially any concurrent chemo-rads regimen). If you are not willing to do this as a radiation oncologist you might as well not even come to work.

Would NEJM publish a case report of a patient with stage I lung cancer who underwent lobectomy and died of a postoperative pulmonary embolism? No. If they did, would (or should) thoracic surgeons stop operating on patients with stage I lung cancer? No.

 

[seper]

So you consent your stage I lung SBRT patients for treatment-related death? Patients should run away from your cancer program and do it fast.

This comment is ridiculous. We frequently subject patients with potentially lethal cancer to a risk of fatal treatment related toxicity on the order of 1-5% (essentially any concurrent chemo-rads regimen). If you are not willing to do this as a radiation oncologist you might as well not even come to work.

Would NEJM publish a case report of a patient with stage I lung cancer who underwent lobectomy and died of a postoperative pulmonary embolism? No. If they did, would (or should) thoracic surgeons stop operating on patients with stage I lung cancer? No.

 

[deleted4401]

Absolutely! Grade 5 toxicity has been seen and you have to mention it. I think not mentioning seems somewhat misleading. Things happen. These are sick patients and even though they may die of other reasons, any potentially curative treatment of NSCLC has a small, but potential risk of grade 5 toxicity. Anyone claiming otherwise - "patients should run away from your cancer program and do it fast"
-S

 

[seper]

You are correctly alluding to the difference between decline in pulmonary function, which is expected and to which SBRT may or may not be contributing, and an obvious fatal adverse event (such as bronchial toxicity described in the NEJM or radiation pneumonitis), which we are expected to avoid.

Absolutely! Grade 5 toxicity has been seen and you have to mention it. I think not mentioning seems somewhat misleading. Things happen. These are sick patients and even though they may die of other reasons, any potentially curative treatment of NSCLC has a small, but potential risk of grade 5 toxicity. Anyone claiming otherwise - "patients should run away from your cancer program and do it fast"
-S

 

[Palex80]

So you consent your stage I lung SBRT patients for treatment-related death? Patients should run away from your cancer program and do it fast.

What are you talking about?

Of course you need to obtain informed consent for treatment-related death if you are doing SBRT.
Pneumonitis can be deadly and treating close to the main airway or trachea can cause uncontrollable bleeding or mediastinitis in case of necrosis.

 

[deleted4401]

Look, fact is, people aren't going to back down after making some statements. Above it said: "So you consent your stage I lung SBRT patients for treatment-related death?" and I said, of course, and then the argument changed about what constitues treatment-related death. Grade 5 is grade 5 is grade 5. I'm not mad-dogging you, Seper, but there is no good evidence to not do it other than anecdote and case report. If you don't want to do it, it's fine, but going to less than 100 BED isn't supported in evidence, so at that point maybe reconsider doing it at all and sending to an academic center (that's what I do, since I don't have the modality). The Japanese data: 1.1%-3 >grade 2 pulmonary toxicity in many hundreds of patients utilizing 50 Gy/5 fx for central and peripheral tumors and they've done it in many hundreds of patients. The American data is similar. The Danish data is similar. The Swedish data is similar. The Pittsburgh data is similar. Everyone is showing the same thing: local control in the 90% range, grade 3 toxicity of 5% or less, and survival in the 3-4 year range. If you want to not do it or treat it at an ineffectual lower dose, that's fine, but I don't think it's based in any strong evidence to tell others to not do it.

People want to defend the NEJM publishing a case report of a T4N0M0 tumor being treated off-protocol and suffering a consequence, that's fine. They might want to mention that SBRT has led to a tripling of survival in patients that were nearly incurable before. The New York Times wants to talk about 3 mistakes a a sloppy hospital made, while forgetting to mention of radiation has cured and palliated and improved the lives of millions of people. As always, most radoncs will just sit there and take it because we get paid well and the job is enjoyable. But amongst ourselves, we oughtta be able to call a spade a spade. This is just more propaganda. Unfortunately, it's being produced by our own. They should be called out for mistreating a locally advanced patient, not supported for publishing nonsense.

And UPenn may be considere a "great institution", but the rad-onc department has not been an early adopter of new technology, nor had they contributed much to the literature in the early era of IMRT and SRS. For being a "top-notch place", they hardly published anything at the beginning, while Beaumont, UPMC, Stanford, MDACC, MSKCC were publishing paper after paper of outcomes data. Now, they are coming up with this nonsense. Give me a break.

 

[medgator]

What are you talking about?

Of course you need to obtain informed consent for treatment-related death if you are doing SBRT.
Pneumonitis can be deadly and treating close to the main airway or trachea can cause uncontrollable bleeding or mediastinitis in case of necrosis.

Yep. No different than a surgeon consenting for treatment-related death/bleeding etc. Of course the risk is remote, but still

 

[RadoncRoulette]

I don't have a major problem with the article as a whole. The benefit of a case report is that knowledge can sometimes be gained from rare events. I hope that NEJM readers are savvy enough to understand this point and not equate this to evidence based medicine. To me, it really underscores the importance of informed consent. I don't think it was wrong to treat the patient in such a fashion assuming there was a discussion of the known and unknown risks (including death).

In my opinion, this is a much more disturbing article:
http://www.practicalradonc.org/article/S1879-8500(12)00043-4/fulltext
Gastric perforation following stereotactic body radiation therapy of hepatic metastasis from colon cancer

Take a look. Mostly surgeon authors it appears. Fortunately it will be in PRO and not the NEJM so we don't have to read about it an any ASCO emails. In this situation there are clear oversights from the treating physicians it appears for a treatment with less evidence.

"The stomach and bowel were not contoured as part of the treatment plan and neither intravenous (IV) nor oral contrast was used to define the stomach during simulation"

No contours= No DVH= No constraints that were used!

 

[Gfunk6]

"The stomach and bowel were not contoured as part of the treatment plan and neither intravenous (IV) nor oral contrast was used to define the stomach during simulation"

No contours= No DVH= No constraints that were used!

Wow.

Well, you have to give them credit for having the cajones to publish this major error on their part. Generally, when surgeons run SBRT/SRS programs the outcomes tend to be much worse for patients. Rad Onc residency is four years for a reason.

 

[seper]

I'm genuinely surprised. Let me understand this right - I'm trying to learn:

- your consent form for SBRT has the word "death" printed on it in late side effects section, and
- when you consent SBRT patient, you tell them something like "radiation can cause damage to lungs and other organs that may lead to your becoming very sick and possibly dying"?

In my current practice and in residency training, both busy SBRT centers on East Coast, we don't expect treatment related mortality from a standard SBRT treatment. Any grade 5 toxicity would trigger an internal investigation.

Look, fact is, people aren't going to back down after making some statements. Above it said: "So you consent your stage I lung SBRT patients for treatment-related death?" and I said, of course, and then the argument changed about what constitues treatment-related death. Grade 5 is grade 5 is grade 5. I'm not mad-dogging you, Seper, but there is no good evidence to not do it other than anecdote and case report. If you don't want to do it, it's fine, but going to less than 100 BED isn't supported in evidence, so at that point maybe reconsider doing it at all and sending to an academic center (that's what I do, since I don't have the modality). The Japanese data: 1.1%-3 >grade 2 pulmonary toxicity in many hundreds of patients utilizing 50 Gy/5 fx for central and peripheral tumors and they've done it in many hundreds of patients. The American data is similar. The Danish data is similar. The Swedish data is similar. The Pittsburgh data is similar. Everyone is showing the same thing: local control in the 90% range, grade 3 toxicity of 5% or less, and survival in the 3-4 year range. If you want to not do it or treat it at an ineffectual lower dose, that's fine, but I don't think it's based in any strong evidence to tell others to not do it.

People want to defend the NEJM publishing a case report of a T4N0M0 tumor being treated off-protocol and suffering a consequence, that's fine. They might want to mention that SBRT has led to a tripling of survival in patients that were nearly incurable before. The New York Times wants to talk about 3 mistakes a a sloppy hospital made, while forgetting to mention of radiation has cured and palliated and improved the lives of millions of people. As always, most radoncs will just sit there and take it because we get paid well and the job is enjoyable. But amongst ourselves, we oughtta be able to call a spade a spade. This is just more propaganda. Unfortunately, it's being produced by our own. They should be called out for mistreating a locally advanced patient, not supported for publishing nonsense.

And UPenn may be considere a "great institution", but the rad-onc department has not been an early adopter of new technology, nor had they contributed much to the literature in the early era of IMRT and SRS. For being a "top-notch place", they hardly published anything at the beginning, while Beaumont, UPMC, Stanford, MDACC, MSKCC were publishing paper after paper of outcomes data. Now, they are coming up with this nonsense. Give me a break.

 

[rymd]

Digression: Do Medical oncologists consent a patient with the word "death" when they start chemotherapy? (Personally- I don't recall them doing that with patients...)

I'm genuinely surprised. Let me understand this right - I'm trying to learn:

- your consent form for SBRT has the word "death" printed on it in late side effects section, and
- when you consent SBRT patient, you tell them something like "radiation can cause damage to lungs and other organs that may lead to your becoming very sick and possibly dying"?

In my current practice and in residency training, both busy SBRT centers on East Coast, we don't expect treatment related mortality from a standard SBRT treatment. Any grade 5 toxicity would trigger an internal investigation.

 

[seper]

Sorry, didn't read the whole thing - you don't have SBRT.

I'm genuinely surprised. Let me understand this right - I'm trying to learn:

- your consent form for SBRT has the word "death" printed on it in late side effects section, and
- when you consent SBRT patient, you tell them something like "radiation can cause damage to lungs and other organs that may lead to your becoming very sick and possibly dying"?

In my current practice and in residency training, both busy SBRT centers on East Coast, we don't expect treatment related mortality from a standard SBRT treatment. Any grade 5 toxicity would trigger an internal investigation.

 

[deleted4401]

I don't do SBRT b/c we don't have it, but my lung cancer consent says specifically as a Rare side effect: "Fatal inflammation of the lung (pneumonitis)." I'm fairly sure UPMC's consent form said it, as well. If/when we start doing SBRT for lung in our little ol' country clinic, you'd better believe it will be on there. Anything can happen. Even if it is unrelated to the RT, if the time course correlates, we'll be blamed. It was reported in the literature before the case report, and so it is a significant concern, but obviously very rare.

As an aside, sorry to be inflammatory. There is certainly differing opinions on this, and I apologize for being jerky.
-S

 

[Palex80]

Digression: Do Medical oncologists consent a patient with the word "death" when they start chemotherapy? (Personally- I don't recall them doing that with patients...)

In Europe this is common practice, especially for the more aggressive chemotherapy regiments.

 

[medgator]

Digression: Do Medical oncologists consent a patient with the word "death" when they start chemotherapy? (Personally- I don't recall them doing that with patients...)

In Europe this is common practice, especially for the more aggressive chemotherapy regiments.

Yup. Some of the lymphoma/conditioning regimens for BMT can be intense and lethal in some cases

 

[Mandelin Rain]

A definitive surgery in this woman would have required right bilobectomy or pneumonectomy. She had significant comorbid disease. Her 30 day peri-op mortality was likely 5-10%, if not higher.

They have the option to turn such a patient down. We don't.