Research Overview

Multiple Sclerosis (MS) is a chronic disease of the central nervous system. The progression of the disease is often tracked by observing changes in lesions (patches of inflammation where the neurons have been stripped of their insulating layer of protein).

Dr. Sturm has developed a lesion browsing tool to identify and visualize lesions from MRI images. Users may browse through an MRI stack, or they may search for lesions of interest. The lesions can be viewed separately, measured, and compared to lesions from other MRIs of the same patient. The browsing tool runs as a plug-in to ImageJ, an open-source, Java-based image processing program.

The goal of the summer was to improve the tool in several ways:

• The algorithm used to locate lesions segments the MRI images into gray matter, white matter, cerebral fluid, and lesions. Because lesions are only located in specified areas of the brain (within white matter), including more information about the boundaries of the brain in the program was provided to produce more useful results.
• Lesions of interest were stored in an array. I modified the data structure used to result in more efficient storage and sorting.
• The visualization of the lesions previously was of limited help because it displayed the lesions “floating in space”, i.e. as isolated spots on the screen. I included the border of the brain as reference points to make it easier to classify and compare the lesions.
• A graphical user interface was designed and created to allow users to easily access and use all of the program features.

Final Paper