Nifti files modified/created by DSI Sutdio not recognized by other programs

[Gautier Kempinaire]

Dear Dr. Yeh,

Thank you very much for creating this amazing program and for your support over the years.

We are setting up a study investigating the use of tractography to reduce the adverse effects of stereotactic radiation therapy in brain tumors. For this purpose, we need a 3D T1 or T2 FLAIR Nifti/DICOM file with tracts marked on it.

I have tried various exportation methods using the "Mark Tracts on Slices" function, but I am unable to find a satisfying solution for our project:

1. Slices > Save Slices to DICOM: I recall this only works on 2D series. Trying this on a 3D sequence simply causes DSI Studio to crash.
2. Export > then selecting T1 or FLAIR: This only works if I don't resample the DWI at the T2 Reconstruction step. However, my 3D T1 gets resampled to the DWI resolution (from 0.9mm³ to either 2mm³ or 2.4mm³), which is not ideal. If resampling occurs at any point, the resulting Nifti files are not recognized by either 3D Slicer or ITK-Snap (Error: Wrong Format. The IO library for the format '' cannot read the image file.)
3. Slices > Save as > .nii: The native resolution of the T1/FLAIR is kept, but the Nifti file is once again not recognized by any programs.
4. DICOM Conversion Errors: I get the same "File not recognized" error when trying to convert these files to DICOM using multiple Python scripts (nii2dcm, nifti2dicom, nifti_to_dicom).
5. Custom Scripts: I tried rewriting scripts myself using nibabel and ITK, to no avail.
6. Software Versions: I tried this with various "Hou" versions, resulting in the same problem.

I suspect that when DSI Studio alters Nifti files, some vital metadata is lost along the way, making them unreadable by standard libraries.

Out of desperation, I tried an online converter (https://www.onlineconverter.com/nifti-to-dicom) and it worked perfectly: I obtained a 3D DICOM series that is accepted by our PACS, Syngo.via, and our StealthStation. Sadly, they do not provide an offline solution, and we need our project to be future-proof and avoid GDPR issues.

If I missed an already existing solution within the software, can you please point me toward it?

If not, I don't know if there is a possible fix for how DSI Studio writes Nifti files.

Best regards,

Kempinaire Gautier
CHU Liège, Belgium

[Frank Yeh]

>
> Slices > Save Slices to DICOM: I recall this only works on 2D series. Trying this on a 3D sequence simply causes DSI Studio to
crash.

I can help fix this error, but may need a sample DICOM data and its
associated .fz file.
Most DICOM files we used here are 2D series.


Best regards,
Frank

[Gautier Kempinaire]

Dear Dr. Yeh,

Sorry for the late reply, but we investigated a bit more while extracting the anonymized data. We think there is a compatibility issue between DSI Studio and the Enhanced DICOM format from our Siemens Vida XA60 scanner.

Maybe you already know what I've written below, but it might still be useful for other users:

Not all software, PACS, and servers can handle this data format. dcm2niix was updated quickly to deal with issues, allowing the conversion of Enhanced DICOM to NIfTI without any loss of information . Until now, we were only using NIfTI files for our projects, so there were no issues; however, asking DSI Studio to interact with Enhanced DICOM files seems to be the source of our crashes.

When exporting DICOM from XA MRI, we can also export in the Interoperability DICOM format at the cost of losing some DICOM headers, but this makes the files compatible with software that accepts VE's version of DICOM (Skyra, Prisma, Trio, etc.). It doesn't really change anything for structural sequences, but for DWI, the DICOM fields for gradient directions feeding the bval and bvec files are lost during the conversion, making them unusable (https://github.com/rordenlab/dcm2niix/tree/master/Siemens#siemens-x-series).

For our test patient, I exported the DWI and fMRI in Enhanced and the T1/T2 FLAIR in Interoperability, and everything works perfectly in DSI Studio. The marked DICOM files generated this way are accepted in 3D Slicer RT modules and in the Syngo.via RT Planning workflow.

I guess the best fix would be to make DSI Studio compatible with Siemens XA's Enhanced DICOM files, but I don't know how hard or time-consuming that would be, since exporting in Interoperability is an existing workaround. Alternatively, I think a pop-up informing the user of the incompatibility with the Enhanced DICOM format could solve a headache for future users of XA MRIs staying in the "DICOM realm."

I'll still upload the files to the Dropbox ifthe scanners you work with are either in VE or non-Siemens, in case you want to look at the files.

Best regards,

Kempinaire Gautier.

[Frank Yeh]

Thanks. I am listing this issue on my to do list. Hope to get back to you soon.

Frank

--
You received this message because you are subscribed to the Google Groups "DSI Studio" group.
To unsubscribe from this group and stop receiving emails from it, send an email to dsi-studio+...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/dsi-studio/a2d1f0d2-8cdb-4e45-be94-e46659de8758n%40googlegroups.com.