RT dose reduction in HPV positive oropharyngeal cancer

[scarbrtj]

I disagree, we can certainly get caught inside our own bubble, but I think with dose de-escalated IMRT in head and neck cancer there is a fine line where we are risking loco-regional recurrences to minimize side effects. Old school doses and margins were used for a reason, be it empirical or evidence based. Lets say we could do the trial to find the exact minimum dose required to cure a specific set of trial patients, would there be any margin of safety dose wise for real life if patients miss treatments, or due to holidays? I think one of the biggest errors that we are making in Rad Onc today is #1 over-imaging for setup (kV daily, fluoro, or multiple CBCT's daily if needed to correct for large shift) which may lead to more 2nd malignancies 15 years from now, and then #2 assuming that what you see on your CT sim is exactly where the tumor is going to be to the mm, on treatment day. CTV to PTV expansions also help to account for some tumor growth between sim and treatment start. Are we using margins so tight that patient's can fly through treatment with little to no side effects (and less time effort needed by us managing them on treatment), but then are at increased risk of failure? Dont get me wrong I am all for de-escalation trials, but let the results play out.

Oh c'mon now. You're worried about a tenth of a centigray from a kV image whilst the patient is getting anywhere from 40-200 cGy across the entire H&N region from daily IMRT? That's like getting worried about the blowback from a mouse fart during a hurricane. Surely... that is not the biggest error in rad onc.

"Old school doses and margins were used for a reason, be it empirical or evidence based." That was completely tossed out with the advent of the IMRT era and patients have done better.

"Are we using margins so tight that patient's can fly through treatment with little to no side effects (and less time effort needed by us managing them on treatment), but then are at increased risk of failure?" My good friend Nick Kuritzky had the lead abstract at ASTRO in 2008 looking at this. But on another note--who's "flying through" H&N chemoRT by dropping PTV expansion margins by a mm or few??? Don't sweat these margins too much in H&N IMRT; and by that I mean to imply that I don't think anybody has the market cornered on margin sanctimony, whether he be CC Wang or me.

---

Members don't see this ad.

 

[scarbrtj]

And sometimes you get another chance to cure a h&n patient if you mess up the first time.

Oh wait, no you don't....

Well don't scare the kids. Sure, sometimes you get another chance. And then you use really teeny margins, and slightly less dose. How about that? Although let's not start calling every local recurrence a "mess up" by the rad onc!

 

[medgator]

Well don't scare the kids. Sure, sometimes you get another chance. And then you use really teeny margins, and slightly less dose. How about that? Although let's not start calling every local recurrence a "mess up" by the rad onc!

"Reirradiation-related morbidity was significant"

But yes I would call using premature data to de-escalate dose and then subsequently having a failure a "boo boo" by the rad onc. I can't remember the last oropharynx failure I had using tried and tested techniques, hpv+ or otherwise...

 

[scarbrtj]

"Reirradiation-related morbidity was significant"

But yes I would call using premature data to de-escalate dose and then subsequently having a failure a "boo boo" by the rad onc. I can't remember the last oropharynx failure I had using tried and tested techniques, hpv+ or otherwise...

So flashback to the early 2000's/late 1990's. "Tried and tested" for H&N is usually opposed laterals, AP s'clav, some posterior electron strips for a few fractions, then a re-sim, a cone down. It involves set daily doses of 1.8-2 Gy almost always to larger and then smaller fields. And thus it had been for decades.

I have a new technique: H&N IMRT. It is untested, unverified, certainly minimal to no long-term data. It involves multiple beams, altered fractionation (e.g. 56 Gy in 35 fx to a large field across the entirety of the ~7 week treatment). It involves parotid-to-nearby-tissue dose gradients heretofore unseen. It involves entirely new thought processes by the treating physician. Basically, a total paradigm shift in treating patients. (It involves changing multiple variables in the treatment chain, not just one variable like dose de-escalation, and is such a change you really can't compare it to anything that's come before.) Will you use it or impugn others who start using it because they're doing so with premature data or, worse, no data?

Again... I'm only saying that dose de-escalation for HPV+ HN seems very feasible, has good data, makes biological sense, etc. Trying to say otherwise at this point is a little on the fear-mongering side. I remember when Ed Halperin used to fear-monger about IMRT. I guess he eventually started using it in various disease sites even though he never got his randomized, long-term-followup 3D vs IMRT trial. Much less a cobalt vs linear accelerator randomized trial. You don't wanna check your ethics at the door and use every wild idea published in every journal, but you don't wanna check your brain at the door and ignore reassuringly consistently promising trends either. There's a happy medium.

 

[medgator]

So flashback to the early 2000's/late 1990's. "Tried and tested" for H&N is usually opposed laterals, AP s'clav, some posterior electron strips for a few fractions, then a re-sim, a cone down. It involves set daily doses of 1.8-2 Gy almost always to larger and then smaller fields. And thus it had been for decades.

I have a new technique: H&N IMRT. It is untested, unverified, certainly minimal to no long-term data. It involves multiple beams, altered fractionation (e.g. 56 Gy in 35 fx to a large field across the entirety of the ~7 week treatment). It involves parotid-to-nearby-tissue dose gradients heretofore unseen. It involves entirely new thought processes by the treating physician. Basically, a total paradigm shift in treating patients. (It involves changing multiple variables in the treatment chain, not just one variable like dose de-escalation, and is such a change you really can't compare it to anything that's come before.) Will you use it or impugn others who start using it because they're doing so with premature data or, worse, no data?

Again... I'm only saying that dose de-escalation for HPV+ HN seems very feasible, has good data, makes biological sense, etc. Trying to say otherwise at this point is a little on the fear-mongering side. I remember when Ed Halperin used to fear-monger about IMRT. I guess he eventually started using it in various disease sites even though he never got his randomized, long-term-followup 3D vs IMRT trial. Much less a cobalt vs linear accelerator randomized trial. You don't wanna check your ethics at the door and use every wild idea published in every journal, but you don't wanna check your brain at the door and ignore reassuringly consistently promising trends either. There's a happy medium.

You're making the jump down in dose from 70 to 54-60 analagous to the 3D --> IMRT transition, which had nothing to do with dose at all.

The analogy doesn't make sense to me. A technique =/= a dose and never will. We know 70 Gy with chemo works well....

Should we just accept that we'll have some patients who have morbidity/mortality from disease recurrences because we didn't feel like waiting for the data to mature long enough to show that changing the fundamental Rx of what we do in H&N CA is OK? Do we consider a 1-year F/U NED to be adequate enough? Maybe you do, but I sure don't.

 

[scarbrtj]

You're making the jump down in dose from 70 to 54-60 analagous to the 3D --> IMRT transition, which had nothing to do with dose at all.

You may not recall, as I don't know how old you are, but the first IMRT H&N regimens altered fractionation schemes a lot. Even to this day, the low-risk ENI areas are commonly treated over a 7 week period with less daily dose as compared to "old days" whereas in the 3D era they were treated over a 5 week period usually at 2 Gy per day. (With my IMRT approach however I still use the "old way" of ~50 Gy/25fx to ENI areas, but it took me time to evolve to that.) At Emory our first IMRT H&N regimens circa 2000 were 28 fraction affairs, 2.4 Gy/day to gross dz, 1.8 Gy/day elective. So where you say it had "nothing to do with dose at all"... it absolutely did. Not to mention initial IMRT treatments took 30 minutes or more, which effectively changed the dose rate vs "old ways."

BEDs with time correction:
50/25
50*(1+2/10)-17.5 = 42.5Gy10

56/35
56*(1+1.6/10)-24.5 = 40.5Gy10

(So adopting IMRT inherently involved some modest amount of dose reduction, and LC rates did not decrease in ENI regions)

 

[medgator]

You may not recall, as I don't know how old you are, but the first IMRT H&N regimens altered fractionation schemes a lot. Even to this day, the low-risk ENI areas are commonly treated over a 7 week period with less daily dose as compared to "old days" whereas in the 3D era they were treated over a 5 week period usually at 2 Gy per day. (With my IMRT approach however I still use the "old way" of ~50 Gy/25fx to ENI areas, but it took me time to evolve to that.) At Emory our first IMRT H&N regimens circa 2000 were 28 fraction affairs, 2.4 Gy/day to gross dz, 1.8 Gy/day elective. So where you say it had "nothing to do with dose at all"... it absolutely did. Not to mention initial IMRT treatments took 30 minutes or more, which effectively changed the dose rate vs "old ways."

BEDs with time correction:
50/25
50*(1+2/10)-17.5 = 42.5Gy10

56/35
56*(1+1.6/10)-24.5 = 40.5Gy10

(So adopting IMRT inherently involved some modest amount of dose reduction, and LC rates did not decrease in ENI regions)

Again the doses you're talking about are elective low dose and high dose regions. What you gave 70 Gy, aka the gross primary tumor and involved lymph nodes, never changed* Not then, not now....Until people want to start getting cute and adopting data before it's ready for prime time









*-if anything, you're giving more now in many cases if you go >2 Gy/day to the primary tumor/lymph nodes

 

[seper]

You guys are probably talking about different p16 positive HNSCC populations. Young never smokers indeed do very well, and de-escalation will eventually pan out. HNSCC patients in my practice are pretty old, with up 100-pack year Hx of smoking, and concurrent alcoholism. Some of their tumors are indeed p16 positive, but I disregard those results. Lots of local relapses after 70 Gy despite acceleration or concurrent cisplatin.

 

[OTN]

You guys are probably talking about different p16 positive HNSCC populations. Young never smokers indeed do very well, and de-escalation will eventually pan out. HNSCC patients in my practice are pretty old, with up 100-pack year Hx of smoking, and concurrent alcoholism. Some of their tumors are indeed p16 positive, but I disregard those results. Lots of local relapses after 70 Gy despite acceleration or concurrent cisplatin.

Mayo de-escalation trials after TORS are using smoking history AND p16 positivity together for this exact reason. Even if p16+ patients with a smoking history don't do nearly as well.

 

[radiation]

You're making the jump down in dose from 70 to 54-60 analagous to the 3D --> IMRT transition, which had nothing to do with dose at all.

The analogy doesn't make sense to me. A technique =/= a dose and never will. We know 70 Gy with chemo works well....

Should we just accept that we'll have some patients who have morbidity/mortality from disease recurrences because we didn't feel like waiting for the data to mature long enough to show that changing the fundamental Rx of what we do in H&N CA is OK? Do we consider a 1-year F/U NED to be adequate enough? Maybe you do, but I sure don't.

One thing that I am not aware has been studied in detail in H&N is how do account for the fact that most modern doses are prescribed to an PTV, whereas in the older days it was to an isocenter. For example, in the Peters MDACC trial (which most of modern dosing is based on) the radiation was delivered with "parallel opposed fields of 6oCo y-rays matched to anterior supraclavicular fields. Treatment was delivered at a rate of 180 cGy per day to the central axis." In analogous fashion, the original SBRT lung data from Indiana was prescribed 60 Gy to the isocenter and it was subsequently found that it was closer to 54 Gy to a PTV. I believe this is also an issue in the old prostate dose-escalation studies. I suspect that the delivered elective nodal dose was probably lower than listed in many of these older trials.

 

[Phantom1]

With my IMRT approach however I still use the "old way" of ~50 Gy/25fx to ENI areas, but it took me time to evolve to that.) At Emory our first IMRT H&N regimens circa 2000 were 28 fraction affairs, 2.4 Gy/day to gross dz, 1.8 Gy/day elective. So where you say it had "nothing to do with dose at all"... it absolutely did. Not to mention initial IMRT treatments took 30 minutes or more, which effectively changed the dose rate vs "old ways."

Just wondering, you doing 2Gy fractions to ENI areas, so you doing sequential cowndowns then for your H&N plans? I think we are all in agreement that dose de-escalation should be tested. All I'm saying is that I don't dose de-escalate off trial currently.

 

[scarbrtj]

Just wondering, you doing 2Gy fractions to ENI areas, so you doing sequential cowndowns then for your H&N plans? I think we are all in agreement that dose de-escalation should be tested. All I'm saying is that I don't dose de-escalate off trial currently.

Yes I do an initial plan to 50 Gy, the only rule being I keep the cord PRV at <35-37 Gy. By the 50 Gy/25fx point, patient has lost weight, tumor has shrunk, etc. So you need a new scan and plan anyways. You then treat gross disease alone and my most common Rx is usually 21 Gy/14 fx over 7 days b.i.d. (here you keep the cord to <10 Gy); total tx time of 32 days. Sum everything up and the dosimetry turns out nicely in the Eclipse platform where you can even use the initial scan/dose as a base dose plan for 2nd plan optimization. This 2nd scan gross tx volume is always much smaller than the 50 Gy volume(s) of course so tolerance is OK or even better (I do over-contour a bit on the 2nd scan but if there is lots of shrinkage I do mostly treat the shrunken volume and not try to treat a pre-tx volume). The thing I don't like about 56 Gy/35 fx or its ilk is that you "bathe" the entire HN in dose over 7 weeks whereas in the "old days" you did have that cone down at the end, with less volume treated, and I kind of got tired of the side effects of 7 weeks of constant large volume HN XRT via IMRT given on one single plan/scan. I have been treating HN this way for 10+ years. Intellectually I like the approach because it 1) accounts for changes in anatomy over the course of tx (and IMRT plans are sensitive to anatomy changes), 2) offer modest acceleration from 35 to 32 tx days, and 3) somewhat recapitulates the old ideas of the past (in terms of treatment pacing and volumes and daily delivered fraction doses and also in terms of the volumes) and is a mild rip-off of concomitant boost technique (which did see some H&N anatomy getting daily doses of 3.3 Gy; for me I'm limiting to 3 Gy obviously).

 

[Palex80]

Thank you, scarbtj, very nice approach.