IMRT for lung cancer?

[Neoplastic]

---

Members don't see this ad.

Just curious if any of you had thoughts on using IMRT in the lungs. There does not seem to be much evidence either way. Do any academic centers use IMRT consistently for lung cancer?

 

[Palex80]

I use it for large volume irradiations, that means Stage IIIA (bulky / multinodal) and Stage IIIB NSCLC in the definitive setting. I am not doing it for postoperative treatments.

 

[medgator]

I use it a lot for central located tumors, especially larger T2-T3 lesions that are abutting the heart, esophagus, spine etc. IMRT seems to be the best for those kind of cases.

 

[deleted4401]

Yeah V5<50, V10<40, if possible. I think there is some Chinese data on V10<50 predicting for lower rates of pneumonitis - http://www.ncbi.nlm.nih.gov/pubmed/20462424?dopt=Abstract.
Been trying to keep mean lung lower than 15 if possible, as well.

 

[user_name]

My training program transitioned fully to IMRT while I was there for all NSCLC cases except SBRT was typically noncoplanar static beams. We did almost no elective nodal. The transition was probably 5 years ago. SCLC is a different topic, I know.

When I went into private practice, the group did NO lung IMRT. I did a couple of cases that were needed, for example, a new contralateral primary near the cord. IMRT had been done for probably 7 years for H/N and prostate before I arrived.

I left that practice. My new practice is almost all IMRT even if chasing some elective nodes. I think a lot of the difference is payor mix, not just technology adoption. Many sites that were denied in the last job (several states away and differenct Medicare carrier) are paid without question here.

My biggest concern about IMRT is the confusing data on V5, V10, etc. There is no good direction in my mind. I think a big deal is actually reviewing the beam arrangements on plans. Some (most?) dosimetrists just want the prettiest plan. I have been given at least two plans where beams were coming needlessly from the anterior for a lesion nearly affixed to the posterior rib or a lateral beam from contralateral side. These are things we would never consider acceptable for a static plan, but they are given to us in an IMRT plan. Granted, the total dose may be smaller if a seven field plan versus a 3 field, but the V5 and 10 are definitely affected.

 

[medgator]

Are you you looking at lung V5 or any other parameters beyond MLD and V20 with IMRT?

Yeah V5<50, V10<40, if possible. I think there is some Chinese data on V10<50 predicting for lower rates of pneumonitis - http://www.ncbi.nlm.nih.gov/pubmed/20462424?dopt=Abstract.
Been trying to keep mean lung lower than 15 if possible, as well.

What SimulD said, but honestly, the data for V5/10 isn't quite at the level of MLD/V20 data. I think basic things are important when planning lung whether it's 3D or IMRT. Avoiding contralateral entrance/exit beams when possible, for one thing. If anything, I get more nervous about carbo/taxol than anything else when it comes to pneumonitis risk

http://www.ncbi.nlm.nih.gov/pubmed/22682812?dopt=Abstract

 

[Neoplastic]

Thank you all for your responses -- it is really interesting to me that such a fundamental question in our field has not been tested more rigorously -- especially given all the differences (dose rate, interplay, incidental nodal radiation, conformality at the edges) that could have a bearing on local control.

 

[radiaterMike]

Thank you all for your responses -- it is really interesting to me that such a fundamental question in our field has not been tested more rigorously -- especially given all the differences (dose rate, interplay, incidental nodal radiation, conformality at the edges) that could have a bearing on local control.

IMRT is less likely to augment efficacy (tumor control) for Stage III NSCLC if dose escalation does not improve outcomes ... IMRT can improve lung dose-volume measures and can allow delivery of 60+ Gy for tumors near the spine, for which there are decades worth of data demonstrating the importance (i.e. lower V20, maintain safe cord dose). For low dose lung exposure, IMRT leads to an increase, but this tends to be more so for V<5.

Late 2nd malignancies is much less of a concern for a disease with a median survival on 18 months, and 5-Y OS of <20%.

So, a study of IMRT vs. conventional 3D RT is probably not that appealing in that there is not much of interest to study.

 

[Neoplastic]

I respectfully disagree about the interestingness of the question. I think the worry is not whether imrt augments disease control but actually whether it is inferior to 3dcrt. I think it is a mistake to assume that equivalent outcomes are assured so long as the simple recipe of dose and fractionation are the same. The other factors I mentioned all could have a significant impact on local control of lung cancer. I think similar observations can be made about the proton-photon debate. At the very least, I would have expected more retrospective analyses in the literature.

 

[Gfunk6]

I think the worry is not whether imrt augments disease control but actually whether it is inferior to 3dcrt. .

And this is precisely why there will never be a randomized trial comparing the two. It is a tremendous financial disincentive, just like for protons.

Personally, I think IMRT for lung cancer should be reserved in cases where lung, spinal cord, or brachial plexus constraints cannot be met with 3DCRT.

 

[RadOncDoc21]

Fortunately we have places like Japan and Europe that do try to answer these types of questions.

 

[radiaterMike]

I agree that this is the reason why it wont be studied- but not solely due to financial reasons, but also because a study of inferiority would a prohibitive number of patients, as well as the fact that those patients who would benefit from reducing lung dose, cord dose, brahcial plexus dose, etc would not be enrolled for the simple reason that no radiation oncologist is going to offer what he/she thinks is an unsafe treatment.

I agree that patients need to be selected for IMRT- it is not a solution for all.

And this is precisely why there will never be a randomized trial comparing the two. It is a tremendous financial disincentive, just like for protons.

Personally, I think IMRT for lung cancer should be reserved in cases where lung, spinal cord, or brachial plexus constraints cannot be met with 3DCRT.

 

[TarHeelRadOnc]

I think part of the challenge of designing such a trial is the question of what the most optimal plan is; often times I can get a 3D-CRT plan to meet all the constraints I am looking for but when I compare to an IMRT plan it not only meets all the constraints but improves target coverage while reducing esophageal dose, V20, and MLD (usually these are N2 cases). Its hard to assess for plan quality in a trial when you are simply looking ot meet constraints and not maximize out the efficiency of IMRT.

In terms of the value of IMRT, I don't think its likely that IMRT would impact clinical outcomes compared with 3D-CRT but I do imagine a benefit in terms of pneumonitis, esophagitis, etc. Problem is that in my experience, different institutions code these things very differently and it would be almost impossible to get uniform reporting of toxicities over a 1-2 years which would be needed for the pneumonitis info.

Agree that MLD and V20 are generally improved for stage IIIA with multiple involved N2 nodal stations and for IIIB with IMRT. Esophageal mean more dependent on which stations involved than technique used, but agree that there may be minimal improvement. Disagree about target coverage. In my experience, target coverage is considerably worse with IMRT in lung compared to 3D-CRT. IMRT plans have more problems with dose buildup at the periphery of the PTV in lung (because of problems with electron equilibrium due to low hounsfield density of lung vs soft tissue). Plans tend to either have worse coverage or more in homogeneity (b/c normalization adjusted to achieve desired PTV coverage).

I think that the previously mentioned point about dose spill to adjacent "elective" nodal regions is a good one. Multiple studies suggest that dose to adjacent uninvolved mediastinal stations is ~40Gy with 3D-CRT. Tends to be considerably less with IMRT. Would take a large study, however, to show a difference in LRC based on this issue alone.

The point midway through this thread about Carbo-Taxol is also very pertinent. Recent Red J study suggested clinical pneumonitis (grade 2+) in range of 50% for patients >65yo treated with carbo-taxol based CRT. Chinese randomized study of CDDP-VP16 vs Carbo-Taxol showed similar rate of RP in Carbo-Taxol arm vs ~25% in SWOG arm. I have been trying to steer med oncs toward platinum-alimpta for patients with adenoCA who aren't SWOG chemo eligible. One of the cooperative groups (I think CALGB) has adopted this as their standard regimen.

Last point, V5 data is VERY weak. I don't think that is ready for prime time. Certainly don't think that it should influence dose or treatment technique decisions based on currently available date.

 

[CUBuffsgrad98]

A great question. I was trained to use 3D (been out for 1.5 years now), but starting to use IMRT a little more. RTOG 0617 (60 Gy vs 74 Gy), allowed for IMRT and stratified for this. Will be interesting when published if any difference in positive or negative outcomes. Both this study and CALGB 30610 (SCLC different dosing) allow for IMRT (again stratified), provided less than 1 cm of tumor movement with breathing. Neither has limits on anything but V20 (in both) or MLD (CALGB only), suggesting that those that wrote these, as well as IRBs, are less than swayed by any other criteria, even when using IMRT.

In talking to my physicists, they are very skeptical about this in terms of movement and heterogeneity as the tumor moves, the leaves are moving, which can lead to dose delivery inaccuracies. Regarding the above meta analysis, this is always interesting data, but the heterogeneity of these things makes it hard to really calculate. For example, I dont see anywhere in it where they report a patients PFTs, and in my experience, those with lower PFTs (possibly higher risk for pneumonitis) are often recommended carbo-taxol as med oncs feel it is more tolerable, and thus it may have nothing to do with the chemo.

 

[Palex80]

The type of IMRT used plays a majoe role in dose uncertainties with moving targets (such as lung cancer). Interplay effect is the key term here.

Rapid Arc / VMAT is considerably more prone to dose uncertainties than step-and-shoot IMRT. Of course, fractionation also plays a role.

 

[seper]

Well... that's IMO debatable. When claiming that IMRT is less prone to interplay effect than VMAT, have to specify what type of IMRT you're talking about - sliding window or step-and-shoot?

The type of IMRT used plays a majoe role in dose uncertainties with moving targets (such as lung cancer). Interplay effect is the key term here.

Rapid Arc / VMAT is considerably more prone to dose uncertainties than step-and-shoot IMRT. Of course, fractionation also plays a role.

 

[Palex80]

Well... that's IMO debatable. When claiming that IMRT is less prone to interplay effect than VMAT, have to specify what type of IMRT you're talking about - sliding window or step-and-shoot?

I did say "step-and-shoot" IMRT

 

[CUBuffsgrad98]

The type of IMRT used plays a majoe role in dose uncertainties with moving targets (such as lung cancer). Interplay effect is the key term here.

Rapid Arc / VMAT is considerably more prone to dose uncertainties than step-and-shoot IMRT. Of course, fractionation also plays a role.

There may be more dose uncertainties with these, but that shouldn't dismiss the uncertainty of step and shoot IMRT with a tumor that may show a lot of movement with breathing.

 

[Palex80]

There may be more dose uncertainties with these, but that shouldn't dismiss the uncertainty of step and shoot IMRT with a tumor that may show a lot of movement with breathing.

Ok, here's some literature on the subject:

Court et al., Med. Phys. 35 (2008) 1926-1931

In this graph from the paper you can see phantom measurements demonstrating, that RapidArc results in the largest dose uncertainties compared with step-and-shoot IMRT.

 

[CUBuffsgrad98]

Ok, here's some literature on the subject:

Court et al., Med. Phys. 35 (2008) 1926-1931

In this graph from the paper you can see phantom measurements demonstrating, that RapidArc results in the largest dose uncertainties compared with step-and-shoot IMRT.

This is interesting data, however, I somewhat disagree with your interpretation. I think this suggests that once you reach about 20 fractions (which would be a fairly low amount in a definitive lung cancer case), RapidArc (single or double), dynamic IMRT (simple, complex, hybrid, 200-600), step and shoot IMRT, conformal, and pretty much everything else are all the same (not inferior), with the sole exception of a complex single rapid arc.

 

[Palex80]

This is interesting data, however, I somewhat disagree with your interpretation. I think this suggests that once you reach about 20 fractions (which would be a fairly low amount in a definitive lung cancer case), RapidArc (single or double), dynamic IMRT (simple, complex, hybrid, 200-600), step and shoot IMRT, conformal, and pretty much everything else are all the same (not inferior), with the sole exception of a complex single rapid arc.

Well in my original post I did say "Of course, fractionation also plays a role."

Bearing in mind, that more and more people tend to use hypofractionated RT for lung cancer (lots of Phase I/II trials with integrated boost giving something like 10 x 5 Gy to the GTV), one can expect that there are lots of patients being treated in centers with large dose uncertainties because of the fast (perhaps too fast?) introduction of dynamic RT into clinical routine for lung cancer.

Believe me, I have seen clinics doing all strage stuff, once they get their hands on some new fancy technique. I have even seen people doing SBRT with 3 fractions for peripheral early stage NSCLC using tomotherapy. Now, imagine the dose uncertainties there... 😱