dicom_hdr should report tags with FD VR

[neurolabusc]

I recently helped an AFNI user who was puzzled that dcm2niix correctly detected that a sequence was acquired with TE=30ms for a dataset where the public tag 0080,0081 was not included.

When I told them that dcm2niix can also detect 0018,9082 they responded:

Dear Chris,

Thank you for quick response.

Tag 0018,9082 is “Unimplemented” for my dicoms:

dicom_hdr NA-PMP-012+00005+00001+00001.dcm | grep "9082"
0018 9082        8 [103148  ] //                               //Unimplemented

The confusion stems from the fact that AFNI's dicom_hdr does not report values for unknown tags with the FD (Float Double) VR. While I can understand that the historically rare Effective Echo Time tag may not exist in a dictionary, dicom_hdr should still resolve the values for tags with explicit VRs. For example dcmdump

dcmdump --search 0018,9082 NA-PMP-012+00005+00118+00118.dcm
(0018,9082) FD 30                                       #   8, 1 EffectiveEchoTime

or gdcmdump where

gdcmdump  NA-PMP-012+00005+00118+00118.dcm > file.txt

reveals:

# Dicom-Meta-Information-Header
# Used TransferSyntax: 
(0002,0000) UL 214                                                # 4,1 File Meta Information Group Length
(0002,0001) OB 00\01                                              # 2,1 File Meta Information Version
(0002,0002) UI [1.2.840.10008.5.1.4.1.1.4.1]                      # 28,1 Media Storage SOP Class UID
(0002,0003) UI [1.3.12.2.1107.5.2.50.176250.2022010510234581165456079]         # 54,1 Media Storage SOP Instance UID
(0002,0010) UI [1.2.840.10008.1.2.1]                              # 20,1 Transfer Syntax UID
...
(0008,1090) LO [MAGNETOM Vida ]                                   # 14,1 Manufacturer's Model Name
(0018,1020) LO [syngo MR XA20 ]                                   # 14,1-n Software Version(s)
...
        (0018,9082) FD 30                                         # 8,1 Effective Echo Time

I can completely understand the commitment required with supporting the complex, evolving and often mis-interpreted DICOM standard. However, now that Siemens is actively upgrading VE11 (classic DICOM, files are 2D slices or mosaics) systems to XA30 (enhanced multi-frame DICOM), I think we will see a very rapid change in the files users are dealing with. One option is to upgrade dicom_hdr. An alternative is to retire classic DICOM tools and inform users of actively developed open source tools (like gdcm).