Brain Follow-Up - Output Image Review

Having performed the brain follow-up analysis as detailed in the Introduction, and Options sections, you can now review the output image for change.

Several output images are created, with the number depending on the number of time points and image contrasts analysed. In this simple example, we analysed only two time points and a single image contrast (proton-density-weighted MR images). The input images are as shown below:
Baseline image, pre-registration 2 year follow-up image, pre-registration
Baseline Two-year follow-up
These two scans were performed 2 years apart. Although the slices were carefully re-positioned for the second scan, the patient's head is in a different position.

The Brain Follow-Up tool creates co-registered versions of each of the input images, with names equal to the input images, but with the prefix 'r'. These co-registered images are shown below:
Baseline image, post-registration 2 year follow-up image, post-registration
Baseline, after co-registration
 Two-year follow-up, after co-registration
At first glance, there appears to be little difference between the images at the two time points. However, the Brain Follow-Up tool also creates difference images (after correcting any bias field differences). In this case, we gave the output image base name 'Test', so the difference between time points 2 and 1 is called 'Test1_2'. This image is shown below for the same slice:
Difference between Year 2 scan and baseline
Difference between the 2-year
follow-up scan and baseline,
after bias field correction.
In the difference image, the new periventricular lesion that has formed is readily apparent. Review of all the image slices would reveal all newly-formed, enlarged, reduced or resolved lesions. If you use the option Option to calculate the magnitude difference images, it is only possible to distinguish between new and resolved lesions, and between enlarged and reduced lesions by simultaneously looking at the difference image, together with the two registered time point images. Review is greatly assisted by loading these three images into linked, spawned displays.

Assisted Review

You can get Jim to load these three images (the difference image, together with the two registered time point images) into three linked display windows by clicking on the Button to display images for review button. This will automatically load the difference image into Jim's main display, and the registered time point images into two spawned slave displays. The ROI Toolkits will also be launched ready for you to make measurements of lesion volume changes on the registered images, confirming the changes by referring to the difference image.

Note: you can only review the last set of results produced by the Brain Follow-Up tool in this way, and if you make any changes to the setup between performing the analysis and requesting the review, then the review may fail to display the correct images.

Jim Home