DCM2NIIx error from Philips DICOM ?QSM data -- can this be skipped?

Hello neuroimaging gurus! And thank you in advance for any help you could provide.

I’m getting an error from DCM2NIIX (error code 4) while trying to convert Philips DICOM data, which includes a QSM protocol that should be kept as DICOMs rather than converting to NIFTI. I think because of this, the latest version of DCM2NIIx now gives me an error (the earlier version did not) as follows: “Conversion aborted due to corrupt file”, which seems to stem from the following python code:

//20190524: Philips MR 55.1 creates non-image files that report kDim1/kDim2 - we can detect them since 0008,0016 reports “RawDataStorage”
printError(“Conversion aborted due to corrupt file: %s %d %d\n”, fname, d.xyzDim[1], d.xyzDim[2]);

Presumably, this is related to the QSM files. Is there any way to tell DCM2NIIX to skip these files? Or is this something more nefarious? Or is it okay to just use an older version of DCM2NIIx? Other programs like Horos seem to open the DICOMS fine, so they don’t seem truly “corrupted” from a data integrity standpoint.

Thank you again!

I suggest you generate an issue at the dcm2niix GitHub page. The primary issue is that Philips has started to generate PS_#### and XX_#### files along with the traditional IM_#### DICOM images. The challenge is that the PS_##### and XX_##### files mimic properties of DICOM files but do not contain raw data. If you do a web search, you will see these files have broken a lot of import tools. I had used the RawDataStorage tag to discriminate true DICOM images from these derived files. I would be grateful if you could (privately) send me a sample image. I do not have access to Philips hardware, so have to reverse engineer their logic.

Thanks very much, Chris. I opened an issue on GitHub. You are correct-- the source of (at least one) error is an XX_#### file. I’ll follow-up with you separately.

Thanks. For others that find this chain in the future, see dcm2niix github issue 328 for comments and resolution.