UK IMPORT LOW: 15 partial breast IMRTs are good

[scarbrtj]

---

Members do not see this ad.

Good 'ol @Palex80 mentioned UK IMPORT LOW last year. Another one I overlooked. Maybe you did too; so here it is.

 

[seper]

Good study, true. However, I'm pretty sure these were just small tangents, not multi-angle IMRT plans.

 

[scarbrtj]

Good study, true. However, I'm pretty sure these were just small tangents, not multi-angle IMRT plans.

"just small tangents" can be "multi-angle IMRT plans," n'est ce pas? "Multi" >1. Let's concentrate on the IMRT part for the public, and Evicore's, sake 🙂

 

[seper]

"The aim of test arm 2 is to deliver a uniform dose of 40Gy to the ‘cylinder’ containing the partial breast PTV using small tangential fields only" - supplemental appendix. I'm so glad now we have a reasonable PBI regimen to replace RTOG 38.5/10 fx bid madness.

 

[evilbooyaa]

Yes, it's "IMRT" in the sense of infinite field in field. Tangential field IMRT. Not VMAT/Tomo/Multiple angle IMRT.

Interesting data on outcomes for EBPBI. This is a more reasonable EBPBI dosing IMO than 38.5/10, which never seemed to make sense compared to the brachytherapy dose of 34.5/10.

 

[Burt Radnolds]

Thought provoking UK breast study

Old news bro.

 

[Fpg1245]

Do you really think you’ll ever be able to truly bill this as IMRT? I can’t see any payor taking this seriously.

 

[radiaterMike]

In the US you must use inverse planning to bill IMRT. Forward planned ‘IMRT’ was used in many of the early breast IMRT (and early U Michigan head and neck) studies. Nevertheless if you are looking for data to support IMRT you can cite this paper and the others. Ultimately insurers will want more than a few references though - usually requiring a clinically meaningful reduction of heart or lung dose or demonstration of an exceptional circumstance (large chest wall separation, post mastectomy implants, need for IMN coverage, prior thoracic radiation ...)

 

[scarbrtj]

Do you really think you’ll ever be able to truly bill this as IMRT? I can’t see any payor taking this seriously.

In the US you must use inverse planning to bill IMRT. Forward planned ‘IMRT’ was used in many of the early breast IMRT (and early U Michigan head and neck) studies. Nevertheless if you are looking for data to support IMRT you can cite this paper and the others. Ultimately insurers will want more than a few references though - usually requiring a clinically meaningful reduction of heart or lung dose or demonstration of an exceptional circumstance (large chest wall separation, post mastectomy implants, need for IMN coverage, prior thoracic radiation ...)

Non-inverse planned IMRT is too fidgety and time-consuming. Inverse planned IMRT is much quicker, gives better dosimetric results. (If we ever wanted to be scientifically pure, we would say true inverse optimization does not actually exist; we have never been known for our scientific purity though.) You can inverse plan mini-tangent breast in about 5 minutes or less. Insurance has usually paid IMRT for partial breast, although fraction allowance has varied. See p. 64 of the Textbook of Radiotherapy.

 

[medgator]

Non-inverse planned IMRT is too fidgety and time-consuming. Inverse planned IMRT is much quicker, gives better dosimetric results. (If we ever wanted to be scientifically pure, we would say true inverse optimization does not actually exist; we have never been known for our scientific purity though.) You can inverse plan mini-tangent breast in about 5 minutes or less. Insurance has usually paid IMRT for partial breast, although fraction allowance has varied. See p. 64 of the Textbook of Radiotherapy.

Inverse planned imrt can also increase contralateral breast dose if you're not careful. When running ecomp vs inverse planned imrt, the ecomp plans with a scv field have ended up looking the same or better in many cases

 

[scarbrtj]

Inverse planned imrt can also increase contralateral breast dose if you're not careful. When running ecomp vs inverse planned imrt, the ecomp plans with a scv field have ended up looking the same or better in many cases

Not true & zero worry in the case of mini-tangs/2 field. That is to say, there is nothing inherent in an inverse planned approach vs "ecomp" (I have a Varian/Eclipse-specific gestalt for "ecomp") that increases dose with one technique vs the other. Planner's beam choice and arrangement will always be the main determinant of contra breast dose. (I could show specific cases where you can achieve higher MUs per beam in an "ecomp" tangent approach vs inverse planned tangent approach, and a resultant higher "outside field scatter" depending on your a priori fluence homogeneity chosen... suffice it to say, there are few good & reliable generalizations that can be made about inverse vs ecomp IMRT.) But keep in mind: "ecomp" is inverse planning, btw. Semantics matter, and we should never as a specialty parsed this phraseology so much as to confuse non-rad oncs. Heck, even some of the rad oncs are confused.
EDIT: If you create an ecomp plan, and take the 95% isovol as a PTV for an inverse-opt plan and fix the field size to the same as the ecomp, the inverse-opt will be equal or better (but it's minorly better in the extreme usu)
FURTHER EDIT: ecomp is inverse optimization whereby the planning system inverse opt designs a fluence that gives 100% Rx dose homogeneity at a plane (a transmission penetration depth) perpendicular to the beam (it's described better than this in the Eclipse manual) for that beam only

 

[seper]

Inverse planning IMRT for 2 mini-tangents is a blatant waste of medical resources. Usually you'll need complex Physics Q&A, dose verification which requires machine time, etc.

 

[scarbrtj]

Inverse planning IMRT for 2 mini-tangents is a blatant waste of medical resources. Usually you'll need complex Physics Q&A, dose verification which requires machine time, etc.

Ecomp (inversely optimized as it is--the only optimization parameter being TPD vs PTV min/max/priority etc in other Eclipse GUI optimization methods) is fine, of course, for breast XRT. And it takes like 5 minutes (or less) to QA it. (And many academic centers don't even take that QA step you're alluding to anymore.) So it's not that blatant, all things considered. What sort of resources are you worried of wasting. We have a well-done randomized trial showing lower side effects than whole breast, using IMRT. At 15 fractions, this IMRT regimen will be much less "waste of resources" than 33 fractions of non-IMRT using multiple conedowns etc. At least, that seems obvious to me.

 

[seper]

I'm saying if you're trying to prevent acute dermatitis while treating 2 small tangents, any competent dosimetrist will produce a 3-control-point-per-field forward plan for you in 5 minutes. No need to fight insurance companies for IMRT approval, to make sure Physics is there to do whatever IMRT Q&A is needed on time, or try to explain you rationale to colleagues.

 

[scarbrtj]

I'm saying if you're trying to prevent acute dermatitis while treating 2 small tangents, any competent dosimetrist will produce a 3-control-point-per-field forward plan for you in 5 minutes. No need to fight insurance companies for IMRT approval, to make sure Physics is there to do whatever IMRT Q&A is needed on time, or try to explain you rationale to colleagues.

I'm not nuts-and-bolts familiar with the technique you mentioned (Eclipse or a different TPS?). With Eclipse ecomp, which generates fluences that convert to sliding window MLCs, many folks still QA the plan as it is a form of IMRT will all its attendant makes-a-physicist-nervy particulars. Again, the QA is quite simple. I would think you would still want to QA a field-in-field forward planned tx to make sure you're getting the in vivo distribution you want. I'm sure there's a QA of some sort even in that scenario beside just shooting a wide open BEV port film. Still bet an Eclipse ecomp could be done/planned quicker than the 3 control-point-per field method (and again I don't know its ins and outs). In my state, we can do IMRT and bill 3D due to some peculiar insurance co. agreements; as I'm always happier with IMRT for breast, that winds up being the course I pursue many times... no fights, or inordinate resource wasting, necessary.

 

[seper]

I've seen breast field-in-field plans done in Pinnacle, Monaco, and XiO. First, a tentative plan using 2 tangents is made without wedges. Then, a dosimetrist projects an area of inhomogeneity onto beam-eye view and manually creates 2-3 reduced beams with fixed MLC positions blocking those areas. "Open tangents" should deliver 70% of MU's or so to address the reproducibility of the target position. Resulting plans are usually much better than simple wedges. Just off-table ports to verify field shapes are needed for QA.

I'm not nuts-and-bolts familiar with the technique you mentioned (Eclipse or a different TPS?). With Eclipse ecomp, which generates fluences that convert to sliding window MLCs, many folks still QA the plan as it is a form of IMRT will all its attendant makes-a-physicist-nervy particulars. Again, the QA is quite simple. I would think you would still want to QA a field-in-field forward planned tx to make sure you're getting the in vivo distribution you want. I'm sure there's a QA of some sort even in that scenario beside just shooting a wide open BEV port film. Still bet an Eclipse ecomp could be done/planned quicker than the 3 control-point-per field method (and again I don't know its ins and outs). In my state, we can do IMRT and bill 3D due to some peculiar insurance co. agreements; as I'm always happier with IMRT for breast, that winds up being the course I pursue many times... no fights, or inordinate resource wasting, necessary.

 

[scarbrtj]

manually creates 2-3 reduced beams with fixed MLC positions blocking those areas.

An automated production of a fluence would be a significantly quicker process, as it will be done in like 1 second for ecomp e.g. (I have done manual FIF in the good 'ol days. It was pretty frustrating due to its iterativeness and resultant potential for coming up with suboptimal FIF segments.)

Just off-table ports to verify field shapes are needed for QA.

So you will have to film each FIF port; more time consuming than just portal imaging 2 sliding-window MLC fields and spot checking the beam fluences in the automated QA software. Also, I imagine the MD will want to check each FIF image (whereas the "traditional IMRT" process has evolved to free the MD from the entire checking-the-field process). For FIF intensity modulation, do you check each port or just let the physicist check the port films.

Again, to me, it just seems way simpler/quicker to do an optimized IMRT approach vs forward planning IMRT. YMMV, but not a blatant resource (if resource=manpower) waste for the former vs latter. That was the point I didn't wanna have dangling out there, that somehow the FIF approach is superior or more facile (acceptable, but not superior).

 

[SneakyBooger]

Noninferiority trials...

With so very few events (incl. N+ pts), should we be concerned that the result would (perhaps way too easily in a noninf trial) support the experimental treatment? Does it not present a concern that the LC is greatest for the smallest fields and lowest doses?

Has the UK NHS done any trials in which the experimental arm was found to be inferior - I am not aware of any. Is this an issue of confirmational bias?

What happens when B not inf to A
Then C not inf to B
Then D not inf to C
Then E not inf to D
Is E truly not inf to A?
Future generations may have to figure that out.

Perhaps once noninf established a superiority trial should ensue to detect smaller differences that noninf trials cannot? That doesn't seem to be the case.

Is this the fox guarding the henhouse? Canadian study looked at hypoFx breast and found that despite the lack of financial incentive to do so the majority of rad oncs continued to utilize conventional fractionation - why not.

Any concern that the payors are the source of trial funding?

Finally, are noninf trials ethical? There has been a significant debate regarding this in the med lit.

So much to think about on this slippery slope.

I may not know the answers but can think of some questions.

I sure do miss superiority trials, though I doubt others involved in healthcare do.

 

[evilbooyaa]

Superiority trials are great when you're trying to escalate level of treatment, because you want to be able to justify the increased cost, toxicity, whatever.

When you're de-escalating level of treatment, you need non-inferiority trials. You can discuss the optimal cut-off for a curable disease (see any discussion of PRIME II or CALGB studies omitting RT in the elderly - is a 7-8% absolute increase in local breast recurrence a big deal? Depends on your outlook. Goes from 2 to 10% LRR at 10 years without RT)

As far as I know, we're somewhat at step C in your graph. They are slowly trying to de-escalate radiation, first in # of fractions, then in size of field, in this early stage group.

Again, I would not use this data for node positive patients. 97-98% of patients were node negative. That's why the total number of events are so low.

 

[BobbyHeenan]

Superiority trials are great when you're trying to escalate level of treatment, because you want to be able to justify the increased cost, toxicity, whatever.

When you're de-escalating level of treatment, you need non-inferiority trials. You can discuss the optimal cut-off for a curable disease (see any discussion of PRIME II or CALGB studies omitting RT in the elderly - is a 7-8% absolute increase in local breast recurrence a big deal? Depends on your outlook. Goes from 2 to 10% LRR at 10 years without RT)

As far as I know, we're somewhat at step C in your graph. They are slowly trying to de-escalate radiation, first in # of fractions, then in size of field, in this early stage group.

Again, I would not use this data for node positive patients. 97-98% of patients were node negative. That's why the total number of events are so low.

Agree fully with not using this in node + patients....

One take away I've seen I think on this board is that to me this tells me that even if I'm doing whole breast, I can really block heart and lung aggressively as long as my seroma cavity is covered well and it *probably* won't impact failure. I already to DIBH on L sided cases which IMO does a great job of heart sparing, but I've always kind of felt this way (ie just block the heart at all costs if seroma can still be covered) and practiced this way (especially in low risk cases, I'm not going to give someone an MI for their DCIS), but now you have a reasonable extrapolation of data that says you're OK to do this.

 

[seper]

40/15 fx PBI is good for ASTRO "suitable" category.