Brain Atrophy - Processing Options for Longitudinal Assessment

Brain Atrophy - Longitudinal

The picture below shows the setup for assessing changes in brain atrophy over time, using T₁-weighted images.

Setup for longitudinal brain atrophy assessment

The input images should, ideally, have been acquired using a 3-D (not multi-slice) pulse sequence, and must be acquired using the same MRI scanner and imaging protocol at every time-point. Follow these steps for longitudinal atrophy assessment:

Set the number of time-points using the spinner. In this example, there are two time-points, but you can set this to more than two if your have more follow-up images.
Load the T₁-weighted images in the input image selection panels. In the example above, these are called "BaselineT1.nii" and "FollowUpT1.nii" for the two time-points.
. If the T₁-weighted images were acquired multi-slice (with slice selection), rather than 3-D (with the slice dimension being phase-encoded), then select this option.
Leave the slider at its default value of 5. This should not need to be changed.
Set the base name and folder for the output images.
Optional. If you want to use FLAIR images as well, increase the number of contrasts to 2, and select the FLAIR images.

Now press the Button to start the atrophy analysis

button to start the analysis, which will take some time. When the analysis is complete, you will see a pop-up message showing the results:

Results for longitudinal brain atrophy assessment

These results show, for each time-point:

The grey-matter (GM) volume.
The white-matter (WM) volume.
The cerebro-spinal fluid (CSF) volume.
The intra-cranial volume (ICV), which is the sum of the volumes of the 3 compartments above.
The brain parenchymal fraction (BPF), which is: (GM Volume + WM Volume) / ICV.
The grey-matter fraction (GMF), which is: GM Volume / ICV.
The white-matter fraction (WMF), which is: WM Volume / ICV.

Also shown for each pair of time-points:

The percentage change in grey-matter volume assessed using the Jacobian integration method.

To make a permanent record of these results, you can:

Write to a text file report. A File chooser will appear, for you to choose a log file name. The default file extension for log files is ".log". If the chosen file already exists, an entry will be appended to the log file.
Write to a PDF file report. A File chooser will appear, for you to choose a PDF file name. If the chosen file already exists, an entry will be appended to the PDF file. A PDF report will also include an illustration the GM/WM/CSF segmentation.

For input images called "BaselineT1.nii" and "FollowUpT1.nii", and a base name of "Longitudinal", these output images will be created in the selected folder:

LongitudinalGMPrior.nii - the grey-matter prior probability image, registered to the average co-registered T₁-weighted images.
LongitudinalWMPrior.nii - the white-matter prior probability image, registered to the average co-registered T₁-weighted images.
LongitudinalCSFPrior.nii - the CSF prior probability image, registered to the average co-registered T₁-weighted images.
LongitudinalLVPrior.nii - the lateral ventricles prior probability image, registered to the average co-registered the T₁-weighted images.
LongitudinalPosition.nii - an image of the pixel positions in template image space.
Longitudinal_Det1_2.nii - showing the determinants of the Jacobian for the deformation that maps the first time-point to the second.

rBaselineT1_pGM.nii and code>rFollowUpT1_pGM.nii- the grey-matter posterior probability images for the two time-points.
rBaselineT1_pWM.nii and code>rFollowUpT1_pWM.nii- the white-matter posterior probability images for the two time-points.
rBaselineT1_pCSF.nii and code>rFollowUpT1_pCSF.nii- the CSF posterior probability images for the two time-points.
rBaselineT1_pOTHER.nii and code>rFollowUpT1_pOTHER.nii- the other (non-tissue class) posterior probability images for the two time-points.
rBaselineT1_Classes.nii and rFollowT1_Classes.nii - colour images showing the final segmented tissue classes at each time-point.

rBaselineT1.nii and rFollowUpT1.nii- co-registered versions of the input images produced during Part 1 of the processing (see below).
bcrBaselineT1.nii and bcrFollowUpT1.nii- bias corrected versions of the input images. Bias field correction occurs as a side-effect of the Expectation-Maximization algorithm.

The procedure for longitudinal atrophy assessment is summarised in the two flow-charts below.

The flow chart for part 1 of the longitudinal brain atrophy
assessment
Part 1 of Longitudinal Atrophy Processing

Part 1 involves co-registration of the T₁-weighted images images and averaging to provide a registration target for the FLAIR images.

The flow chart for part 2 of the longitudinal brain atrophy
assessment
Part 2 of Longitudinal Atrophy Processing

In Part 2, the template tissue class prior probabilities are registered to the average T1 and these, together with the co-registered T₁-weighted and optional FLAIR images are input to the Expectation-Maximization classification algorithm.

The red arrows indicate a registration step, with the arrow pointing towards the registration target. The use of the FLAIR image is optional. E-M: expectation-maximization classification algorithm.

Jim Home