What margins do you use for GBM volume delineation?

There is an ASTRO consensus guidelines committee currently working on guidelines for glioblastoma. One of the key questions facing the committee is the optimal target volumes for potentially curative radiation therapy. Issues include treatment of FLAIR/T2/edema followed by a boost to the T1 enhancing volume versus enhancing abnormality only with wider margins as well as what those margins should be. The traditional margins utilized by RTOG are larger than those utilized by the American Brain Tumor Consortium (merger of NABTT and NABTC). At least three institutions from the ABTC have reported retrospective studies of smaller margins without an increase in marginal recurrences including Wake Forest, Emory, and University of Alabama at Birming...

This conversation is much more useful than the newly published ASTRO guideline in Practical Radiation Oncology:

http://www.practicalradonc.org/article/S1879-8500(16)30003-0/abstract

The entire discussion of target volume margins was reduced to a hedged commentary with one self-citation. Both this guideline and NCCN tilt non-specialists toward larger margins unnecessarily and contradict what it seems many academic CNS specialists do in practice. As a result, it means that 70% of community practice radiation oncologists following NCCN and ASTRO may not improve cancer control but only increase leukoencephalopathy risk.

Emory McDonald et al., PMID 20399036 n=62
Michigan Chan et al., PMID 11896114 n=34
Wake Forest Paulsson et al., PMID 23211224 n=161...

Great comments above. I have found it quite interesting that there are quite a lot of differences in the way that GBM volumes are done. Interestingly, if you look at the Canadian recommendations (Mason et al., PMID 17593983, they recommend treating only the T1post area of disease, with no consideration of the FLAIR. Similarly, the Stupp publications ([1],[2]) focus more on T1post (or "contrast-postive region on CT" as per the original paper) without a cone-down. However, the margins were quite large, in the range of 2-3 cm, which suggests that in many cases FLAIR was included. In both cases though, the dose has been 60 Gy. This contrasts with the RTOG protocol.

As mentioned above, recently it has been shown that margins can be reduced consid...

Interesting discussion and something I struggle with on a regular basis.

It is peculiar that what was a margin to block edge got transformed into a PTV margin. A 2 cm margin to block edge is more akin to a 1.5 cm margin expansion.

We know there are tumor cells in the edema as @Dr. First Last pointed out, but if these cells are under or untreated will they ever manifest as a recurrence? Perhaps not. Perhaps so, since data from several institutions (including our own) show that in the select group of 'long-term' survivors, distant recurrences are more apt to occur. What we don't know is if we somehow knew where GBM would recur (outside of a central recurrence), would treating these regions at initial diagnosis prevent that recurrence in the fut...

@Dr. First Last - These responses are from 2014 - 2017. Have practices changed by 2024?

Great discussion from 2014. Six years later, is there a consensus?

Great discussion. I will also add that POF analysis from our own institution suggests that long term survivors in the TMZ era are at higher risk for distant failure. Again, this is food for thought for the residents but to me that suggests that RT-TMZ is increasingly effective in controlling disease and it is the microscopic disease that did not receive therapeutic dose of radiation that is resulting in the distant failure. I agree that T2 is not the best reflection of microscopic disease and I know there are many groups studying advancing imaging to see if there are better imaging studies that reflect tumor biology - hopefully, the next generation can help us figure that out.

This is a terrific discussion. At my institution, we have begun an arduous retrospective study along with Dr. Li at MD Anderson to attempt to compare failure patterns and clinical outcomes in patients treated with larger RTOG volumes vs tighter margins in a large pooled analysis. We hope that this will help address this important question.

As someone trained at the institution where Dr. Kelly did the initial biopsy studies leading to our inclusion of the edema in the target volume, I will present a slightly different view. The original biopsy study confirmed high-grade glioma cells are present (not just low-grade tumor, as secondary GBM is only a small proportion of GBM), throughout (all the way to the periphery) the T2 signal hyperintensity on MRI. Similar to our practice in many other sites (breast, head and neck, etc), radiation oncologists began treating edema harboring microscopic disease to a lower dose and boosting the highest risk/gross disease. From the initial '2cm to block edge' on T2 which was what many adopted initially, as MRI fusion became more common place and...

For standard fractionation schemas:

Our current philosophy is Flair + 5 mm to CTV/5 mm PTV

If using anisotropy maps then area of anisotropy + 5 mm/5 mm PTV

For accelerated courses, I agree with @Dr. First Last - we focus on enhancement.

As for the reference to Pat Kelly's work, as a neurosurgeon myself (and radiation oncologist), I am well aware of the biopsy data - in fact, showing GBM cells in the contralateral hemisphere - I strongly feel that we swung overboard when we went from CT to MRI and our GBM fields were too large for too long.

The more restricted margins have not yielded more failures - but we are in the process of formally analyzing this.

Thanks for the really useful comments.

This is another recent and hopefully useful reference that also nicely summarises recent developments in target volume delineation and imaging techniques in gliomas.

Whitfield et al., PMID 24824451

Great question, and great timing! I was thinking of writing up a blog post on this topic because NCCN recommends much wider margins - with no data to back it up. In addition to the data listed above, here is an Italian study

Minniti et al., PMID 20855119

Here's another re: the need to include edema in your CTV:

Chang et al., PMID 17306935

Fascinating discussion thread, thanks for everyone's comments. Indeed the Kelly paper was key, although I personally struggle to reconcile the findings that many places, across the pond and above the border, only cover T1post with no differences in outcomes (examples below). One note for the Kelly paper, or at least the 1987 version in Mayo Proceedings is how easily one can call a small group of cells in the periphery on a biopsy specimen as high grade, since the features of necrosis and vascular proliferation can be problematic to call; which Kelly implies in the discussion.

Similarly, the choice to treat the FLAIR regions with 46 Gy has been mysterious if we presume all microscopic GBM cells after a GTR are radioresistant (a whole other de...

Excellent discussion, thanks @Dr. First Last and @Dr. First Last for your insights and some reading to look forward to.

As a community doc, my issue has been that the BTSG/RTOG definitions haven't correlated well to the evolution being explained above. My particular concern is that NCCN, which many now reference in clinical decisions, gives variable detail on "Principles of Radiation Therapy"

Lung is great in NCCN. CNS, not so much. Per NCCN: "The gross tumor volume (GTV) is best defined using pre-and postoperative MRI imaging using enhanced T1 and FLAIR/T2. The GTV is expanded by 2-3 cm (clinical target volume, CTV) to account for sub-diagnostic tumor infiltration." [BRAIN-C, page 1]

I agree that we should reassess what the patterns of failur...

Indeed - if anyone can find more references on things like the 46 Gy dose origin, send them along. I will ask Ed Shaw...

Thank you everyone for the interesting discussion. Apparently, there was a small randomized trial that compared the RTOG volumes vs more modest volumes that reported an OS advantage with a smaller PTV. The abstract was presented at SNO but I haven't seen a publication with the details. Slide 17 for reference: https://www.astro.org/uploadedFiles/Main_Site/Meetings_and_Events/2013_Spring_Refresher_Course/Meeting_Program/AdultBrainTumors.pdf

Thanks again for all the helpful comments!