questions > ArterialSpinLabelingType in output json file
Showing 1-2 of 2 posts
Display:
Results per page:
Feb 10, 2025  03:02 PM | Jason Smith
ArterialSpinLabelingType in output json file

Hello


We are using dcm2niix ("ConversionSoftwareVersion": "v1.0.20220720") to convert multidelay PCASL data and create the associated json file.


The output json file has 2 fields labeled "ArterialSpinLabelingType", the first with the correct "PCASL" value and a second with "PASL" as the value which ends up overwriting the first when the file is later edited. This is an easy fix on our end so it's not a huge problem.


Also the field "PostLabelDelay" (which I think should be PostLabelingDelay maybe?) has a single incorrect value that doesn't match any of the actual delays. There are some other fields that ASLPrep thinks are supposed to be in there, but I don't know how "standard" they are, and again this is easily fixed after the fact.


Thank you


j


 


 

Feb 10, 2025  05:02 PM | Chris Rorden
RE: ArterialSpinLabelingType in output json file

Note that the ASL information depends a lot of vendor (GE, Siemens, Philips) and the software model (e.g. Siemens VE11 is very different from XA30). The first thing I would strongly recommend is to use the latest stable release (v1.0.20241211) - with the adoption of enhanced DICOM and the ratification of the ASL BIDS specification, there were a lot of changes in the intervening 29 months In particular, it is unfortunate that the BIDS specification renamed some of the tags populated by dcm2niix, breaking backward compatibility. 
  https://github.com/rordenlab/dcm2niix/re...

I would also suggest you try the ExploreASL wrapper for dcm2niix, from the team that helped create the BIDS specification. While they use dcm2niix (with its permissive license), their code uses an institutional license that prevents me from understanding some of the sequence details it extracts. Therefore, it is likely to provide a more thorough solution
  https://github.com/ExploreASL/ExploreASL


For most vendors and models, the DICOM data is underspecified with regards to the desired fields, and will require some manual intervention. The research collaboration manager (if you use Siemens) or clinical scientist (for Philips) affiliated with your center can help.

If you can share the dataset where dcm2niix assumes "PASL" when it should be "PCASL", send a link to my institutional email:
  https://sc.edu/study/colleges_schools/ar...

and I can try to do better.