mica_pet

MICA PET - Positron Emission Tomography Images

Processing workflow for the PET MK-6240 (tau tracer) images.

How to retrieve the data

Use the find_hrrt command at the command line, you'll get better mileage:

cat `which find_hrrt`
#!/bin/sh

if [ "$1" == "-h" ] ; then
   echo
   echo "You can provide a string representing the hrrt study to look for!"
   echo
   exit
fi

/usr/bin/find /hrrt/scans{0,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33} -maxdepth 1 -name \*$1\*

exit

PET data organization

The raw files obtained from the PET scanner are stored at:

/data_/mica3/MICA-PET/sourcedata

The processed data is stored:

/data_/mica3/MICA-PET/derivatives/micapet

The BIDS is stored:

/data_/mica3/MICA-PET/rawdata

PET features

The PET images are transformed to NIFTIS from the ECAT (.v) files with dcm2niix. This transformation keeps the data in nano Curies. On the other hand, minc files units are becquerels.

Parameters

Dimensions: 256 x 256 x 207
Voxel size: 1.21875 x 1.21875 x 1.21875

Unit reference

cc=cm3=mL
Bq: Becquerels
nCi: nano Curies
1 nCi = 37 Bq

3D non motion corrected data

1. TX256.v Linear atenuation map, 4D_MC is corregistered to this image. great for corregistering to MRI!
2. EM_3D.v 4 frames, 20 minutes of acquisition each one (Bq/cc).
3. EM_3D_AVG.v average of the four frames (Bq/cc).

4D motion corrected data

1. EM_4Dfbp_MC01.v Filter Back Projection image. It is the inversion of the radon transformation.
2. EM_4D_MC01.v 4 frames, 20 minutes of acquisition each one (nCi/cc).
3. EM_4D_MC01_AVG.v Average of 4 four frames (nCi/cc units).

PET json default info

{
	"Manufacturer": "Siemens",
	"ManufacturersModelName": "328",
	"ProcedureStepDescription": "unknown",
	"ProtocolName": "HC10-MK-HRRT",
	"SeriesNumber": 1,
	"ImageComments": "F-18",
	"IsotopeHalfLife": 6588,
	"Dosage": 370,
	"RepetitionTime": 300,
	"ReconMatrixPE": 256
}

PET processing

The functions to process the PET data can be found under the mica-pet GitHub repository. The processing workflow consists in two steps, a transformation from ECAT to NIFTI, and a set of registrations to the T1nativepro space and the to the surface.

1. dcm2niix
2. fslmaths -Tmean compute average of EM_4D_MC01.v
3. PET to T1 registration

• PET to native MRI space (ANTs)
• PET to Freesurface (surface) space (bbreg)

From PET raw data to T1nativepro space

micapet_01.sh

Normalize PET and projects it to the surface

micapet_02.sh

PET processing workflow

BIDS PET

Specification of BIDS conform PET data.