Reliability and validity of an algorithm for fuzzy tissue segmentation of MRI.

PURPOSE A new multistep, volumetric-based tissue segmentation algorithm that results in fuzzy (or probabilistic) voxel description is described. This algorithm is designed to accurately segment gray matter, white matter, and CSF and can be applied to both single channel high resolution and multispectral (multiecho) MR images. METHOD The reliability and validity of this method are evaluated by assessing (a) the stability of the algorithm across time, rater, and pulse sequence; (b) the accuracy of the method when applied to both real and synthetic image datasets; and (c) differences in specific tissue volumes between individuals with a specific genetic condition (fragile X syndrome) and normal control subjects. RESULTS The algorithm was found to have high reliability, accuracy, and validity. The finding of increased caudate gray matter volume associated with the fragile X syndrome is replicated in this sample. CONCLUSION Since this segmentation approach incorporates "fuzzy" or probabilistic methods, it has the potential to more accurately address partial volume effects, anatomical variation within "pure" tissue compartments, and more subtle changes in tissue volumes as a result of disease and treatment. The method is a component of software that is available in the public domain and has been implemented on an inexpensive personal computer thus offering an attractive and promising method for determining the status and progression of both normal development and pathology of the CNS.

[1]  J Toriwaki,et al.  Automatic segmentation of head MRI images by knowledge guided thresholding. , 1991, Computerized medical imaging and graphics : the official journal of the Computerized Medical Imaging Society.

[2]  D. Marquardt.  An Algorithm for Least-Squares Estimation of Nonlinear Parameters , 1963 .

[3]  Godfrey D. Pearlson,et al.  User-friendly method for rapid brain and CSF volume calculation using transaxial MRI images , 1991, Psychiatry Research: Neuroimaging.

[4]  G J Barker,et al.  Simulation of MRI cluster plots and application to neurological segmentation. , 1996, Magnetic resonance imaging.

[5]  Allan L. Reiss,et al.  Neurodevelopmental effects of the FMR-1 full mutation in humans , 1995, Nature Medicine.

[6]  L O Hall,et al.  Review of MR image segmentation techniques using pattern recognition. , 1993, Medical physics.

[7]  J. Rapoport,et al.  A technique for single-channel MR brain tissue segmentation: application to a pediatric sample. , 1996, Magnetic resonance imaging.

[8]  N C Andreasen,et al.  Automatic atlas-based volume estimation of human brain regions from MR images. , 1996, Journal of computer assisted tomography.

[9]  James R. MacFall,et al.  Accuracy and reproducibility of brain and tissue volumes using a magnetic resonance segmentation method , 1996, Psychiatry Research: Neuroimaging.

[10]  K. Lim,et al.  Segmentation of MR Brain Images into Cerebrospinal Fluid Spaces, White and Gray Matter , 1989, Journal of computer assisted tomography.

[11]  T Sandor,et al.  Segmentation of brain CT images using the concept of region growing. , 1991, International journal of bio-medical computing.

[12]  G J Barker,et al.  Improvements to the quality of MRI cluster analysis. , 1994, Magnetic resonance imaging.

[13]  L M Fletcher,et al.  A multispectral analysis of brain tissues , 1993, Magnetic resonance in medicine.

[14]  D. Kennedy,et al.  Anatomic segmentation and volumetric calculations in nuclear magnetic resonance imaging. , 1989, IEEE transactions on medical imaging.

[15]  P. Ring,et al.  Texture Analysis in Quantitative MR Imaging , 1995, Acta radiologica.

[16]  Nancy C. Andreasen,et al.  Segmentation techniques for the classification of brain tissue using magnetic resonance imaging , 1992, Psychiatry Research: Neuroimaging.

[17]  N. Otsu.  A threshold selection method from gray level histograms , 1979 .

[18]  Max A. Viergever,et al.  Probabilistic segmentation of partial volume voxels , 1994, Pattern Recognit. Lett..