The aim of this work was to develop a fast and accurate method for tissue segmentation in magnetic resonance imaging (MRI) based on a four‐dimensional (4D) feature map and compare it with that derived from a 3D feature map. High‐resolution MRI was performed in 5 normal individuals, in 12 patients with brain multiple sclerosis (MS), and 9 patients with malignant brain tumors. Three inputs (proton‐density, T2‐weighted fast spin‐echo, and T1‐weighted spin‐echo MR images) were routinely utilized. As a fourth input, either magnetization transfer MRT was used or T1‐weighted post‐contrast MRI (in patients only). A modified k‐nearest neighbor segmentation algorithm was optimized for maximum computation speed and high‐quality segmentation. In that regard, we a) discarded the redundant seed points; b) discarded the points within 0.5 standard deviation from the cluster center that were non‐overlapping with other tissue; and c) removed outlying seed points outside 5 times the standard deviation from the cluster center of each tissue class. After segmentation, a stack of color‐coded segmented images was created. Our new technique utilizing all four MRI inputs provided better segmentation than that based on three inputs (P < 0.001 for MS and P < 0.001 for tumors). The tissues were smoother due to the reduction of statistical noise, and the delineation of the tissues became sharper. Details that were previously blurred or invisible now became apparent. In normal persons a detailed depiction of deep gray matter nuclei was obtained. In malignant tumors, up to five abnormal tissue types were identified: 1) solid tumor core, 2) cyst, 3) edema in white matter 4) edema in gray matter, and 5) necrosis. Delineation of MS plaque in different stages of demyelination became much sharper. In conclusion, the proposed methodology warrants further development and clinical evaluation. J. Magn. Reson. Imaging 1999;9:768–776. © 1999 Wiley‐Liss, Inc.
[1]
Peter E. Hart,et al.
Nearest neighbor pattern classification
,
1967,
IEEE Trans. Inf. Theory.
[2]
H E Cline,et al.
3D reconstruction of the brain from magnetic resonance images using a connectivity algorithm.
,
1987,
Magnetic Resonance Imaging.
[3]
W. Lorensen,et al.
Two algorithms for the three-dimensional reconstruction of tomograms.
,
1988,
Medical physics.
[4]
Lactate observation in vivo by spectral editing in real time.
,
1988,
Magnetic resonance imaging.
[5]
R. Kikinis,et al.
Three-dimensional segmentation of MR images of the head using probability and connectivity.
,
1990,
Journal of computer assisted tomography.
[6]
Bob S. Hu,et al.
Pulsed saturation transfer contrast
,
1992,
Magnetic resonance in medicine.
[7]
B. Dawson-Saunders,et al.
Basic and Clinical Biostatistics
,
1993
.
[8]
MRI contrast enhancement in the central nervous system : a case study approach
,
1993
.
[9]
Feroze B. Mohamed,et al.
Tissue Segmentation in MRI as an Informative Indicator of Disease Activity in the Brain
,
1995,
ICIAP.
[10]
R. Knobler,et al.
Improved intracranial lesion characterization by tissue segmentation based on a 3D feature map
,
1997,
Magnetic resonance in medicine.
[11]
J Gordon,et al.
In vivo validation of tissue segmentation based on a 3D feature map using both a hamster brain tumor model and stereotactically guided biopsy of brain tumors in man
,
1998,
Journal of magnetic resonance imaging : JMRI.