Processing Overview

<< Click to Display Table of Contents >>

Navigation:  PNROD Data Processing > Workflow for dual-Modality Studies >

Processing Overview

The two images of the same subject brain which are used during processing are called Input and Anatomical:

Input Image

The Input image series is the target of the analysis. It will be used for VOI statistics as well as parametric mapping. Note that in the user interface the label occasionally appears capitalized INPUT for clarity.

Anatomical Image

The Anatomical image is used for deriving the brain VOIs by adjusting it to an atlas. Use of an Anatomical reference image is recommended if it supports better normalization than using the Input directly.    

The processing steps, reflected by the user interface, are the following (with the Anatomical image being MR):

1.Brain Atlas Adjustment page, INPUT layout:

a.Loading of the Input image series which may be static or dynamic.

b.Dynamic Input image case: Averaging of the series in a specified acquisition range. The averaged image will be used in all following steps except for the final statistics calculation and parametric mapping.

c.Cropping of the image to retain mainly the brain part, potentially with some neighboring tissue which helps the normalization (e.g. skull for CT, Harderian glands for PET).

2.Brain Atlas Adjustment page, ANATOMICAL MR layout:

a.Loading of the Anatomical image series.

b.Cropping of the brain area to a similar extent as for the Input image.

c.Start of the automatic rigid matching of the Input image to the Anatomical image.

3.Brain Atlas Adjustment page, MATCHED INPUT layout:

a.Visual, interactive assessment of the alignment.

b.Manual improvement of the alignment if necessary.

c.Configuration of the atlas, the normalization method and the normalization template.

d.Start of the normalization procedure which matches the Anatomical image to the selected atlas template.

4.Brain Atlas Adjustment page, NORMALIZED layout:

a.Visual assessment of the normalization.

b.In case of a bad match, return to the MATCHED INPUT layout, change the normalization approach, and normalize again. Repeat until a satisfactory match is obtained. If no automatic procedure works, resort to manual rigid matching.

c.Start transformation of the atlas brain regions so that they can be overlaid on the Anatomical image as segments.

5.Brain Atlas Adjustment page, BRAIN SEGMENTS layout:

a.Visual assessment of the segments overlaid on the Anatomical image.

b.Selection of the image space (Atlas, Input, Anatomical) where the VOIs are generated.

c.Start generation of the outline VOIs in the result space.

6.Brain Atlas Adjustment page, BRAIN VOIS layout:

a.Visual assessment of the brain VOI outlines overlaid on a fusion of the Anatomical and the Input image by the user. If the match is not satisfactory, return to steps 3 and 4 and improve matching and/or normalization.

b.If needed, manually adjust the whole VOI set or individual structures. For instance, all VOIs can be scaled at once.

c.Start calculation of the VOI statistics on the Input image.

7.Results Statistics page:

a.Inspect the results.

b.Save the statistics table (or time curves in dynamic case).

c.In the dynamic case transfer the curves (usually PET tissue time-activity curves) to the PKIN tool (option) and continue there with modeling.

8.Save the complete configuration of the processing with Save protocol in the bottom line.

Parametric Mapping of Dynamic Data

In the case of a dynamic Input series, return to

9.Brain Atlas Adjustment page, BRAIN VOIS layout:

a.Select an appropriate parametric mapping method.

b.Start parametric mapping.

10.Parametric Mapping page:

a.Perform the Parametric Mapping workflow as in PXMOD.

b.Save the parametric maps.

11.Save the complete configuration of the processing including parametric mapping with Save protocol in the bottom line.