From Kent:

If I was just proposing a change to ITK this would be a bit easier -- I'd
just go through ITK's gerrit review system.

As it is this has a few more moving parts.

The problem was that MGHImageIO was writing out Image origins incorrectly.
Eun Young Kim discovered this while using SimpleITK within Slicer to read
and write MGZ files.

I re-wrote the 'WriteHeader' code where it converts the origin from ITK
conventions to FreeSurfer conventions, so the conversion code on the Write
side matches that on the Read side.

But getting this incorporated back into Slicer and ITK has more than one
step.

ITK:

1. Merge the pull request into the Slicer/itkMGHImageIO repository on
   github. The pull request is here:
   https://github.com/Slicer/itkMGHImageIO/pull/3
2. Get a topic through ITK Gerrit to update the SHA to match the revision
   of my change.

SLICER:

Slicer has an older version of MGHImageIO than the one used by ITK; I
don't know if Slicer uses this and turns on the MGHImageIO remote module
when it builds ITK.
A) If Slicer uses the old code in Slicer/Libs/MGHImageIO, then this should
be replaced by the code in the Github repository above.
B) If Slicer uses the ITK version via the ITK remote module for
itkMGHImageIO, then Š well there's a bit of a problem, as when the fix
gets propogated through into ITK, it will be at the Development head for
ITK, and be some time before it's part of a ITK Minor version release.
Slicer no doubt picks a particular ITK version and sticks with it until a
newer version comes along with sufficient new functionality and bug fixes
to justify re-validating ITK all over again with Slicer.

So I have the fix, I just need to know what I need to do for it to be
pushed where it needs to be.