XA30 myomaps BEATs triggers conversion warnings, unusable nifti

[bpinsard]

Describe the bug

I got in my hands XA30, Myomaps, BEATS_map (bunch of various naming), tfi2d1_109 data, apparently used for T1 mapping with the STAGE protocol on Siemens. Supposed to be cardiac imaging but used on brain (waiting on more info from the site).

dicoms are 5 slices x 10 InversionTimes

It stores 1 slice per dicom despite the ICE dims specifying 5: it triggers #742
It also triggers:

• Philips fMRI discrepancy between TR and other temporal measures #560 (Warning: Discrepancy between reported (0.45361s) and estimated (1.13361s) repetition time (issue 560).)
• Warning: Unable to determine slice direction: please check whether slices are flipped (derived image)
• Converting Siemens Vida data #236 Warning: Siemens XA exported as classic not enhanced DICOM (issue 236) : however the dicoms contains EnhancedMRImageStorage stuffs and shared/per-frame structures.

The resulting nifti is non-usable: while resulting in a [218, 256, 5, 10] nifti,

• the slices are not ordered correctly
• the affine has a zero in the s-form at the position [3,3], so fails on a bunch of nifti loader (related to slices issue above)
• the inversion times are not ordered (nor split in different files such as for echoes or complex, is that expected?).
• it takes only a single (first) value for InversionTime in the sidecar.

If I take only 5 dicoms with the same InversionTime, it generates a usable volume.

These are likely non-conform enhanced dicoms (at least for the ICE dims info).

v1.0.20240202

[neurolabusc]

You might want to see if the latest commits to the development branch help out a bit, as I recently included modifications to aid XA30 data saved as classic DICOMs. Regardless, I would suggest you have the data re-exported as enhanced DICOM, as the XA30 enhanced DICOMs include useful meta data, in particular for complex sequences like this. If these tips do not reoslve your issue, the only other solution would be for you to send a sample dataset to my institutional email.

[bpinsard]

Thanks, I will build the latest and check if that helps.
Do classic dicoms still have some EnhancedMRImageStorage structures (with shared and per-frame sections) ? If so, how do one differentiate between classic and enhanced? The rest of the session series are enhanced dicoms, so I would be surprised that it was separately exported as classic, but I don't have access to the history of that pilot dataset.

[neurolabusc]

@bpinsard I do not have XA30, so I can not provide much feedback without seeing a dataset. As I recall, the XA30 does provide an option to use classic for some but not all images (as I recall, there is a n option for using enhanced just for the BOLD fMRI, but it does not generalize to other EPI sequences like DWI).

[bpinsard]

build of development branch reduces the number of errors/warnings.

Warning: Discrepancy between reported (0.45361s) and estimated (1.13333s) repetition time (issue 560).
Warning: Images sorted by instance number [0020,0013](21001..21050), but AcquisitionTime [0008,0032] suggests a different order (150744..150710)
::autoBids:Siemens CSAseqFname:'%SiemensSeq%\BEAT_map' pulseSeq:'*tfi2d1_109' seqName:''
Warning: Siemens MoCo? Bogus slice timing (range -1..33155, TR=1133.33 seconds)
Warning: Siemens XA exported as classic not enhanced DICOM (issue 236)

The resulting nifti is now usable and seems structured appropriately by InversionTime.
Thanks! ( closing the issue)

[bpinsard]

One thing though: it only lists the first InversionTime in the sidecar.
Related question: is there a way to print debug which dicoms files goes at which position in the resulting 4d stack, just to ensure that their position makes sense?

[neurolabusc]

@bpinsard I can not really comment more without seeing sample images. You can observe a lot by running dcm2niix with logorrheic vebosity (-v 2) or use dcmdump or gdcmdump to view the headers (the latter has the -csa flag which is useful for Siemens). I do encourage you to reach out to the Siemens Research Collaboration Manager affiliated with your center, as it is their role to help with these issues.