Abstract #3982
Pattern Recognition Classification of Weighted MR Images of OARSI Scored Human Articular Cartilage at 3T
Vanessa A. Lukas 1 , Beth G. Ashinsky 1 , Christopher E. Coletta 2 , Julianne M. Boyle 1 , David A. Reiter 1 , Corey P. Neu 3 , Richard G. Spencer 1 , and Ilya G. Goldberg 2
1
Magnetic Resonance Imaging and Spectroscopy
Section, National Institute on Aging, National
Institutes of Health, Baltimore, Maryland, United
States,
2
Image Informatics and Computational
Biology Unit, National Institute on Aging, National
Institutes of Health, Baltimore, Maryland, United
States,
3
Weldon School of Biomedical
Engineering, Purdue University, West Lafayette, Indiana,
United States
An important limitation in the application of MRI to the
early detection and monitoring of osteoarthritis (OA) is
the substantial overlap in parameter values between
different degrees of cartilage degradation. In several
studies, multiparametric analysis as been shown to
markedly improve discrimination ability. We extend this
through application of an established pattern
recognition algorithm, wndchrm, to
T
1
,
T
2
,
T
2
*
,
ADC and MT weighted images obtained on OARSI-graded
human cartilage explants. We found that wndchrm, which
detects differences in textures and intensity patterns
between images through examination of multiple image
transforms, results in substantially higher
classification performance than conventional univariate
analysis.
This abstract and the presentation materials are available to members only;
a login is required.
Join Here
