|Ph.D Student||Goldberg-Zimring Daniel|
|Subject||Quantitative Characterization of Multiple Sclerosis|
Lesions Geometry in the 3-D Space
|Department||Department of Biomedical Engineering||Supervisors||PROF. Haim Azhari|
|PROF. Ron Achiron|
Multiple Sclerosis (MS) is the most common demyelinating disease of the central nervous system white matter, and MRI is the most suitable technique for monitoring it.
In this research, a new method to quantitatively characterize the 3-D shape of MS lesions detected by MRI was developed. With this method sets of lesion contours taken from segmented MR images were used to generate an analytical approximation of the lesion’s geometry by spherical harmonics (SH). Using this approximation, the lesion volume and indices, which are invariant to space rotation and quantitatively describe its shape, were obtained. The method was implemented to sets of in vivo and in vitro MRI brain images, and to computer generated synthetic lesions. The results indicate that good geometrical approximations and good consistency in volume estimation can be obtained. A validation of the method was obtained by studying computer-generated lesions and by comparing the volume and 3-D shape of MS lesions detected by MRI to pathological measurements. It was also demonstrated that the utilized indices are relatively insensitive to changes in lesion orientation, but significantly sensitive to morphological changes.
Size and shape analysis were also implemented to analyze the changes in individual MS lesions’ 3-D geometry over time in a patient studied by MRI 24 times during one year. The results demonstrate that the changes corresponding to lesions’ shape can be 1.4 to 8.0 times higher than those corresponding changes in size/volume.
In conclusion, the suggested method offers new tools for monitoring MS burden and activity by MRI, and for the systematic exploration of characteristic patterns of lesion evolution.