Submission Sahlgrenska

[oscarjalnefjord]

Name: Oscar Jalnefjord

Number of datasets (sets of inversion recovery measurements: 1

Submission Checklist:

• [x] Have you collected all the raw data exported directly from scanner into a single folder?
• [x] For each dataset, did you create one (or two, if complex) NIFTI files with each inversion time measurement in the "time" dimension?
• [x] For each NIFTI file, did you create a YAML file with the same filename containing the necessary details about your dataset for the challenge? (See sample submission on OSF.io for a template)

When your data is ready for review, please comment in your issue and email a link to a Google Drive folder (or other online data sharing service) containing your data to: rrsg.challenge@gmail.com.

[oscarjalnefjord]

The data is almost ready for review, I just have a few questions:

1. What defines "version" under "sample" in the yaml file? Is it serial number <0042 vs. >0041 or something else?
2. What should be put into the field "name" under "sequence" in the yaml file? (I run a Philips scanner)
3. (just a minor note) I went for complex data (real/imag) and used dcm2niix followed by fslsplit and fslmerge to get the corresponding nifti files. When I inspected the final nifti files in fsleyes I noticed that the real and imaginary data had different slice position. I guess it wont be an issue? The origin of this is that dcm2niix put the real/imag data in the 3rd dimension as only a single slice was scanned.

[mathieuboudreau]

What defines "version" under "sample" in the yaml file? Is it serial number <0042 vs. >0041 or something else?

@agahkarakuzu You're the one who told me that the NIST phantom had different "versions". Do you know if this tracked somewhere, or how users could find out which one they're using?

What should be put into the field "name" under "sequence" in the yaml file? (I run a Philips scanner)

I don't think this is super important, it's just to gather more information about what folks did. When I converted my DICOMS to nifti (acquired on Siemens), the exported json file accompanying the header information of the DICOM has a parameter called "ScanningSequence" (see below), and that's what I placed there. I'll either clarify this point in the template, or omit it, after I see a few more submissions. (This is also partly why I encourage folks to upload their raw data as well)

(just a minor note) I went for complex data (real/imag) and used dcm2niix followed by fslsplit and fslmerge to get the corresponding nifti files. When I inspected the final nifti files in fsleyes I noticed that the real and imaginary data had different slice position. I guess it wont be an issue? The origin of this is that dcm2niix put the real/imag data in the 3rd dimension as only a single slice was scanned.

Interesting - I didn't know that! I think it should still be fine, what do you think @agahkarakuzu ? Once you share your data, I'll be able to test it a bit more to know.

Thanks!

[agahkarakuzu]

Reference values are going to be different for different versions.

There are two prototype versions I know, Phannie and Phreddie, you can see the reference here. The data archive has examples for both. Commercialized versions of these phantoms were sold by HPD, now QualibreMD.

If any center has one of these prototypes, version should refer to Phannie or Phreddie. If not, serial number seems relevant, see the description below:

You can download the latest specs from here.

Bottomline, if people have a prototype phantom, the version is Phannie or Phreddie. If have the commercial one, then the serial number. The spec that comes with the commercial phantom contain the details.

[mathieuboudreau]

Great, thanks Agah. Which version do we have? We should add this to an FAQ page/section on this repo.

[agahkarakuzu]

@mathieuboudreau the serial number is written on the Plate 5 (which is the plate of the T1 array). Our phantom is from HPD. For the SN, I will check one of the highres images I acquired before and see if it is visible on the image.

[agahkarakuzu]

I could not find the serial number on my laptop, it is probably on my workstation at MHI. I have the page for reference values here:

[oscarjalnefjord]

Thanks for the clarifications

The data is available for review now (emailed link)

[mathieuboudreau]

Hi @oscarjalnefjord ,

Your submission was properly formatted, thank you for your attention to detail! I did a few minor adjustments, because the template changed slightly today.

Your data was uploaded to the RRSG/qMRSG OSF.io repo: https://osf.io/pg74t/

I tested it in the processing pipeline, and it ran successfully.

Pipeline output:

Run it here: