questions > GEIIS violates the DICOM standard
Showing 1-8 of 8 posts
Display:
Results per page:
Jul 11, 2017  01:07 AM | cg17
GEIIS violates the DICOM standard
Hello,

I am trying to convert some DICOM files with dcm2niix v1.0.20170624 Clang8.1.0 (64-bit MacOS), but I get this:

Warning: GEIIS violates the DICOM standard. Inspect results and admonish your vendor.

I saw that someone ran into this problem in a past post, but I thought that versions after Nov 2016 would be able to fix this. Any ideas how I can get around this problem? 

Thanks so much!

C
Jul 11, 2017  02:07 PM | Chris Rorden
RE: GEIIS violates the DICOM standard
To be clear, this is a problem with your images (and PACS server) and not my software. Hopefully my software accurately detects this defect and converts the images correctly. My software converts DICOM to NIfTI but it does not change the DICOM data. Therefore broken DICOM images will remain broken. My software is correctly warning you that you need to upgrade the software on your PACS system, as it is corrupting your DICOM images. While my software may be able to handle this problem, if you do not upgrade your PACS you may have problems with other DICOM tools that expect valid DICOM files. Your images are a clear violation of the DICOM standard. This may cause serious archival issues. GE is a professional company that sold the PACS software with a DICOM conformance statement. If you contact them, they have a responsibility to patch your PACS system. I describe this more here:
  https://github.com/rordenlab/dcm2niix/issues/113
Jul 11, 2017  05:07 PM | cg17
RE: GEIIS violates the DICOM standard
Thanks for your quick response! This only happens to our older < 2007 images, so I suspect the system was upgraded around 2007. Is there any way to "fix" this older files so they can be used?
Jul 12, 2017  07:07 PM | Chris Rorden
RE: GEIIS violates the DICOM standard
The GEIIS added a thumbnail image, but unfortunately uses this used a different transfer syntax from the main image (which is not allowed). Therefore, you must remove all the tags associated with the sequence (SQ) Group,Element 0009,1010. This requires a DICOM-to-DICOM converter. Perhaps GE can provide you with a cleaner, or you could modify dcm2niix to identify this and remove it. You might also want to see if any of these solutions works: 
 http://dicom.offis.de/dcmtk.php.en
 http://www.dclunie.com/pixelmed/software/webstart/DicomCleanerUsage.html
 http://gdcm.sourceforge.net/html/gdcmconv.html
Jul 15, 2017  06:07 PM | cg17
RE: GEIIS violates the DICOM standard
Thanks so much! I try your suggestions out.
Sep 21, 2017  07:09 PM | Paul Wright - King's College London
RE: GEIIS violates the DICOM standard
Hi Chris

I have this error message with some older scan data. I could not find element 0009,1010 but I tried removing all of 0009 and then all private tags, but still got the GEIIS error. dcm2niix does not appear to work around the error as it just says 'no valid DICOMs found' and produces no files.

I'm using dcm2niix v1.0.20170624 GCC4.4.7 (64-bit Linux).

I can convert the files using old dcm2nii (12 12 2012) but if you can recommend a fix so I can use dcm2niix I would prefer that.

Many thanks
Paul
Sep 27, 2017  12:09 PM | Chris Rorden
RE: GEIIS violates the DICOM standard
Thanks for providing me with another variation of the GEIIS bug. My latest software release v1.0.20170923 handles these files. You can get the latest version on Github https://github.com/rordenlab/dcm2niix/releases
Sep 27, 2017  04:09 PM | Paul Wright - King's College London
RE: GEIIS violates the DICOM standard
Many thanks, Chris. The new version works for my dodgy files.