XNAT vs Gyrus DICOMs

Imaging data may be pulled from XNAT, WPIC-XXX@cortex (mounted gyrus /disk/mace2/scan_data), or flywheel.

The imaging data is identical but the dicoms are not.

XNAT puts it's own information in the header and that changes all the dicom offsets. But the tags and values are all the same. So dicom_info in matlab, pydicom in python, dcmdumps, and AFNI's dicom_hdr all will report the same values.

gyrus=2020.10.10-09.41.12/11688_20201010/T1_MPRAGE_Sag_Res1mm_256x256.10/MR.1.3.12.2.1107.5.2.43.167092.2020101010005271808660254
xnat=xnat/11688_20201010/10/DICOM/1.3.12.2.1107.5.2.43.167092.30000020100814342951600000022-10-1-5c9s0t.dcm
vimdiff <(dicom_hdr $gyrus) <(dicom_hdr $xnat)

The imaging data is also identical.

dcm2niix -o . -f gyruns /Volumes/Hera/Raw/MRprojects/SlipsPilot/2020.10.10-09.41.12/11688_20201010/T1_MPRAGE_Sag_Res1mm_256x256.10/
dcm2niix -o . -f xnat xnat/11688_20201010/10/DICOM/
3dBrickStat -slow '3dcalc(-a gyruns.nii -b xnat.nii -expr abs(a-b) )' # 0 -- all voxels have same value
diff gyruns.json xnat.json # No difference

0

More interesting to note: the dicoms file names from mace2/scan_date are not necessarily sorted by acquisition number (0020,0013 = REL Instance Number). I had to be careful when picking the dicom to compare. the first by timestamp (gyrus) and first by number (xnat) are not the same.

find /Volumes/Hera/Raw/MRprojects/SlipsPilot/2020.10.10-09.41.12/11688_20201010/T1_MPRAGE_Sag_Res1mm_256x256.10/ -iname "MR*" |
  head -n7 |
  xargs dicom_hinfo -tag 0020,0013  |
  sed 's:/.*T1:...:'

find xnat/11688_20201010/10/DICOM/ -iname "*dcm" |
  sort -V |
  head -n7 |
  xargs dicom_hinfo -tag 0020,0013  |
  sed 's:/.*DICOM:...:'