fs-aseg2dicom-seg.md

1. Description of the platform/product:

• name and version of the software: FreeSurfer 6.0
• free?: yes
• commercial?: no
• open source?: yes
• what DICOM library do you use?: DCMQI to write DICOM-SEG
• Description of the relevant features of the platform: Freesurfer's subcortical segmentation results can be converted to DICOM-SEG using fs-aseg.json that has been kindly provided by Emily Lindemer.
• are both single and multiple segments supported?: Yes, however Freesurfer's subcortical segmentation process assumes a mutually exclusive label set.
• do you support both BINARY and FRACTIONAL segmentation types? BINARY only.
• do you support compressed objects? I beleive itkimage2segimage can read compressed itk objects but does not write compressed DICOM objects(?)
• do you render the segment using the color specified in the DICOM object? N/A (write-only), however Freesurfer's recoomended color scheme is preserved.
• how do you communicate segment semantics to the user? The fs-aseg.json maps freesurfer's subcortical labels to SNOMED codes.
• how do you support the user in defining the semantics of the object at the time segmentation is created? Freesurfer's subcortical label set is predefined, so this is pre-computed.

2. Read task: Not supported

3. Write task:

The file dcmqi-fs-aseg.example.tar.gz contains the following:

• ./dicom-anon: Direcotry containting input dicoms for a FreeSurfer-compatible T1 weighted MPRAGE sequence.
• aseg-t1space.nii.gz: FreeSurfer 6.0 subcortical segmentations (this is aseg.mgz output from FreeSurfer transformed back to the original input DICOM coordinate system and converted to nifti)
• fs-aseg.json: Mappings from FreeSurfer aseg labels to SNOMED codes provided by Emily Lindemer.

Running the example:

tar -zxvf ./dcmqi-fs-aseg.example.tar.gz
cd ./dcmqi-fs-aseg-example
docker pull qiicr/dcmqi
docker run \
  -v $PWD:/tmp/dcmqi/ \
  qiicr/dcmqi \
  itkimage2segimage \
    --inputDICOMDirectory /tmp/dcmqi/dicom-anon \
    --inputMetadata /tmp/dcmqi/fs-aseg.json \
    --inputImageList /tmp/dcmqi/aseg.nii.gz \
    --outputDICOM /tmp/dcmqi/aseg.dcm

Verifying the output:

dciodvfy ./aseg.dcm &> dciodvfy-output.txt

Currently dciodvfy genertes the following errors:

• Thousands of erros that look like:

Error - Illegal root for UID - "5.555.5.2017.3.22.??.?.??????" in (0x0008,0x1155) Referenced SOP Instance UID

• Other errors:

Error - Value invalid for this VR - (0x0008,0x0050) SH Accession Number  SH [0] = <Accession Number 1> - Length invalid for this VR = 18, expected <= 16
Warning - Value dubious for this VR - (0x0008,0x0090) PN Referring Physician's Name  PN [0] = <Referring Physician's Name 1> - Retired Person Name form
Warning - Value dubious for this VR - (0x0010,0x0010) PN Patient's Name  PN [0] = <Patient's Name 1> - Retired Person Name form
Warning - Value dubious for this VR - (0x0070,0x0084) PN Content Creator's Name  PN [0] = <Reader1> - Retired Person Name form
Error - Dicom dataset contains invalid data values for Value Representations
Segmentation
Error - Unrecognized enumerated value <CLEANED> for value 1 of attribute <Patient's Sex>