Hi Martin,
 
it's great to get questions like that on the list!
Please don't feel repulsed by the medical issues. FMRI and tractography have the potential to enter decision-making for the patient's good. But it's a real interdiciplinary process (truely exciting, I think) where specialities need to learn from each other. Currently, folks on the image analysis side are often "scared" by medical implications extending to a real given patient. On the other hand, medical doctors often have too ambitious expectations to these "new", quite complicated technologies.
Commonly, this interaction has a "social" or "assuring" component to it: Of all the mapping or tractography requests I get hardly one-out-of-ten make sense from the imaging AND medical point of view. But from my point of view, as soon you are willing to give your opinion you do take some responsibility for the patient (regardless of whether you are a medic, psychologist, image analysist or whatever). Sometimes, the neurosurgeon is simply not sure how to deal with the case and may has the hope that resourcing to FMRI, DTI etc. will help him to solve the issue. As soon as you talk to him and as soon as he understands the limitations we are confronted with, he may make up his mind even without the mapping.
So I would certainly go into that in a case where a resection of a M1 lesion is scheduled. If there is perifocal edema and the lesion is right in M1, you can't pick the access to the lesion or limit the resection according to the tractography with the thought in the back of your mind that this will prevent any deficits or do the patient any good. On the other hand, I would strongly question whether a M1 high grade glioma should be resectioned. In most cases, the risks to further cripple the patient will prevail over the benefits you are hoping for. Quite different for a sub-Rolandic lesion: the tractography may tell you (possibly quite to your surprise!) that (at least some of) the pyramidal hand fibres descend down in front of the lesion. Although you won't know if you are missing fibres running down in the back (very unlikely if there is no motor deficit), you would still approach the lesion from the back because you know for sure that you can do harm if you are coming from the front.
I guess I am just arguing that if the neurosurgeon asks you to do something you should understand why he is asking you the question, what are the possible implications of your answer and if the request does in fact make sense from both the medical as well as the imaging side. I know - kind of tricky because medics are sometimes hard to deal with (I am allowed to say that since I belong to them) but give them a chance to learn from your expertise!
Cheers-
Andreas
PS: Needless to say that FSL offers you great tools to do that kind of work - probtrack being one of them;)

________________________________

Von: FSL - FMRIB's Software Library im Auftrag von Martin Kavec
Gesendet: Mi 05.07.2006 13:02
An: [log in to unmask]
Betreff: Re: [FSL] AW: [FSL] ProbTrack on a tumor patient



Hi Andreas,

thanks a lot for great insight into this topic. Perhaps I put my email
more like a doc, but I am not. So I will have to get your message second,
third, and more thought. Anyway it is great to know this side of the
"business"

Cheers - Martin

On Wed, 5 Jul 2006, Andreas Bartsch wrote:

> Hi,
> ok - so the goal is to pick up any fibres in the perifocal edema.
> First, if the lesion is in M1 and the neurosurgeon still wants to resect
> it, there will remain a postoperative motor deficit. The best you can
> attain is that it won't worsen substantially (the very best would be it
> will get a bit better when the edema eventually resolves). The worst
> case scenario is a (substantial!) worsening of the motor deficit.
> Because cure is unlikely the higher the grading of a malignant glioma
> and since you already know that you would be resecting within M1, most
> would be afraid to inflict an additional functional deficit on the
> patient (i.e. the approach needs to be weighted against the nil nocere
> principle). However, assuming that the neurosurgeon really wants to go
> for it (?) and given that the topographic relation between lesion and M1
> is definite I can't see any good argument to limit or guide your
> resection based on the tractography. The only guidance you can get is
> from the intraoperative monitoring, and even that will be very
> problematic. Even though probtrack may be better than streamlining,
> there is no reliable way to tell the amount of fibres disguised by the
> edema. Is it the surgical approach itself (not the extent of resection)
> you want to guide by the tractography? I would not really be in favor of
> that either (assuming the lesion is really in M1 and already scheduled
> for resection).
> However, you may be aware of all this and may have your specific reasons
> to go for it. If so, you can run probtrack in seed mask mode using a
> mask of the edema within M1. This should give you all the fibres
> originating from the seed. Additionally, you can use this mask as a seed
> and the cerebral peduncle ipsilaterally (=crus cerebri) as a single
> waypoint. You may then also look at the difference between the two
> results (using avwmaths_32R or above) to assure that the waypoint mode
> did not miss anything of the pyramidal tract.
> Hope this helps, keep us posted-
> Andreas
> PS: It is a different story if you are unsure about the topographic
> relation between lesion and eloquent (incl. motor) cortex. However, that
> is rarely the case for M1. It is also a different story if you are
> talking about a sub-Rolandic (subcentral) lesion. Here it is hard to
> tell whether the pyramidal tract passes in front or in the back of the
> lesion or both (see
> http://www3.interscience.wiley.com/cgi-bin/fulltext/112598789/PDFSTART
> <http://www3.interscience.wiley.com/cgi-bin/abstract/112598789/ABSTRACT?
> CRETRY=1&SRETRY=0>  for an example). In fact, this is the case where
> pyramidal tractography makes most sense to me. Nevertheless, there is
> really no way to make a proper informed decision based just on DTI if
> the edema is bad - unfortunately.
>
>
> ________________________________
>
> Von: FSL - FMRIB's Software Library im Auftrag von Martin Kavec
> Gesendet: Mi 05.07.2006 09:57
> An: [log in to unmask]
> Betreff: [FSL] ProbTrack on a tumor patient
>
>
>
> Hello,
>
> I have a 32 direction DTI images from a high grade glioma patient
> (scheduled for resection) with a lesion in a motor cortex area.
> The lesion itself consists of a necrotic core (very low anisotropy)
> surrounded by hyperintense tissue (with moderate anisotropy) on T2
> weighted
> images. I suspect that there are still functional fibers in this region,
> which I would like to pick up by ProbTrack. I ran bedpost on the data,
> but now I wonder what would be the most meaningful Mode in ProbTrack for
> finding tracks within this T2 hyperintense lesion.
>
> Thanks for suggestions.
>
> Martin
>
>
>