Diffusion tensor imaging-based tissue segmentation: validation and application to the developing child and adolescent brain

We present and validate a novel diffusion tensor imaging (DTI) approach for segmenting the human whole-brain into partitions representing grey matter (GM), white matter (WM) and cerebrospinal fluid (CSF). The approach utilizes the contrast among tissue types in the DTI anisotropy vs. diffusivity rotational invariant space. The DTI-based whole-brain GM and WM fractions (GMf and WMf) are contrasted with the fractions obtained from conventional magnetic resonance imaging (cMRI) tissue segmentation (or clustering) methods that utilized dual echo (proton density-weighted (PDw)), and spin-spin relaxation-weighted (T2w) contrast, in addition to spin-lattice relaxation weighted (T1w) contrasts acquired in the same imaging session and covering the same volume. In addition to good correspondence with cMRI estimates of brain volume, the DTI-based segmentation approach accurately depicts expected age vs. WM and GM volume-to-total intracranial brain volume percentage trends on the rapidly developing brains of a cohort of 29 children (6-18 years). This approach promises to extend DTI utility to both micro and macrostructural aspects of tissue organization.     

Brain tissue segmentation based on DTI data

We present a method for automated brain tissue segmentation based on the multi-channel fusion of diffusion tensor imaging (DTI) data. The method is motivated by the evidence that independent tissue segmentation based on DTI parametric images provides complementary information of tissue contrast to the tissue segmentation based on structural MRI data. This has important applications in defining accurate tissue maps when fusing structural data with diffusion data. In the absence of structural data, tissue segmentation based on DTI data provides an alternative means to obtain brain tissue segmentation. Our approach to the tissue segmentation based on DTI data is to classify the brain into two compartments by utilizing the tissue contrast existing in a single channel. Specifically, because the apparent diffusion coefficient (ADC) values in the cerebrospinal fluid (CSF) are more than twice that of gray matter (GM) and white matter (WM), we use ADC images to distinguish CSF and non-CSF tissues. Additionally, fractional anisotropy (FA) images are used to separate WM from non-WM tissues, as highly directional white matter structures have much larger fractional anisotropy values. Moreover, other channels to separate tissue are explored, such as eigenvalues of the tensor, relative anisotropy (RA), and volume ratio (VR). We developed an approach based on the Simultaneous Truth and Performance Level Estimation (STAPLE) algorithm that combines these two-class maps to obtain a complete tissue segmentation map of CSF, GM, and WM. Evaluations are provided to demonstrate the performance of our approach. Experimental results of applying this approach to brain tissue segmentation and deformable registration of DTI data and spoiled gradient-echo (SPGR) data are also provided.     

Automatic segmentation and reconstruction of the cortex from neonatal MRI

Segmentation and reconstruction of cortical surfaces from magnetic resonance (MR) images are more challenging for developing neonates than adults. This is mainly due to the dynamic changes in the contrast between gray matter (GM) and white matter (WM) in both T1- and T2-weighted images (T1w and T2w) during brain maturation. In particular in neonatal T2w images WM typically has higher signal intensity than GM. This causes mislabeled voxels during cortical segmentation, especially in the cortical regions of the brain and in particular at the interface between GM and cerebrospinal fluid (CSF). We propose an automatic segmentation algorithm detecting these mislabeled voxels and correcting errors caused by partial volume effects. Our results show that the proposed algorithm corrects errors in the segmentation of both GM and WM compared to the classic expectation maximization (EM) scheme. Quantitative validation against manual segmentation demonstrates good performance (the mean Dice value: 0.758+/-0.037 for GM and 0.794+/-0.078 for WM). The inner, central and outer cortical surfaces are then reconstructed using implicit surface evolution. A landmark study is performed to verify the accuracy of the reconstructed cortex (the mean surface reconstruction error: 0.73 mm for inner surface and 0.63 mm for the outer). Both segmentation and reconstruction have been tested on 25 neonates with the gestational ages ranging from approximately 27 to 45 weeks. This preliminary analysis confirms previous findings that cortical surface area and curvature increase with age, and that surface area scales to cerebral volume according to a power law, while cortical thickness is not related to age or brain growth.     

Magnetic resonance image tissue classification using a partial volume model

We describe a sequence of low-level operations to isolate and classify brain tissue within T1-weighted magnetic resonance images (MRI). Our method first removes nonbrain tissue using a combination of anisotropic diffusion filtering, edge detection, and mathematical morphology. We compensate for image nonuniformities due to magnetic field inhomogeneities by fitting a tricubic B-spline gain field to local estimates of the image nonuniformity spaced throughout the MRI volume. The local estimates are computed by fitting a partial volume tissue measurement model to histograms of neighborhoods about each estimate point. The measurement model uses mean tissue intensity and noise variance values computed from the global image and a multiplicative bias parameter that is estimated for each region during the histogram fit. Voxels in the intensity-normalized image are then classified into six tissue types using a maximum a posteriori classifier. This classifier combines the partial volume tissue measurement model with a Gibbs prior that models the spatial properties of the brain. We validate each stage of our algorithm on real and phantom data. Using data from the 20 normal MRI brain data sets of the Internet Brain Segmentation Repository, our method achieved average kappa indices of kappa = 0.746 +/- 0.114 for gray matter (GM) and kappa = 0.798 +/- 0.089 for white matter (WM) compared to expert labeled data. Our method achieved average kappa indices kappa = 0.893 +/- 0.041 for GM and kappa = 0.928 +/- 0.039 for WM compared to the ground truth labeling on 12 volumes from the Montreal Neurological Institute's BrainWeb phantom.     

White matter lesion extension to automatic brain tissue segmentation on MRI

A fully automated brain tissue segmentation method is optimized and extended with white matter lesion segmentation. Cerebrospinal fluid (CSF), gray matter (GM) and white matter (WM) are segmented by an atlas-based k-nearest neighbor classifier on multi-modal magnetic resonance imaging data. This classifier is trained by registering brain atlases to the subject. The resulting GM segmentation is used to automatically find a white matter lesion (WML) threshold in a fluid-attenuated inversion recovery scan. False positive lesions are removed by ensuring that the lesions are within the white matter. The method was visually validated on a set of 209 subjects. No segmentation errors were found in 98% of the brain tissue segmentations and 97% of the WML segmentations. A quantitative evaluation using manual segmentations was performed on a subset of 6 subjects for CSF, GM and WM segmentation and an additional 14 for the WML segmentations. The results indicated that the automatic segmentation accuracy is close to the interobserver variability of manual segmentations.     

Comparative assessment of statistical brain MR image segmentation algorithms and their impact on partial volume correction in PET

Magnetic resonance imaging (MRI)-guided partial volume effect correction (PVC) in brain positron emission tomography (PET) is now a well-established approach to compensate the large bias in the estimate of regional radioactivity concentration, especially for small structures. The accuracy of the algorithms developed so far is, however, largely dependent on the performance of segmentation methods partitioning MRI brain data into its main classes, namely gray matter (GM), white matter (WM), and cerebrospinal fluid (CSF). A comparative evaluation of three brain MRI segmentation algorithms using simulated and clinical brain MR data was performed, and subsequently their impact on PVC in 18F-FDG and 18F-DOPA brain PET imaging was assessed. Two algorithms, the first is bundled in the Statistical Parametric Mapping (SPM2) package while the other is the Expectation Maximization Segmentation (EMS) algorithm, incorporate a priori probability images derived from MR images of a large number of subjects. The third, here referred to as the HBSA algorithm, is a histogram-based segmentation algorithm incorporating an Expectation Maximization approach to model a four-Gaussian mixture for both global and local histograms. Simulated under different combinations of noise and intensity non-uniformity, MR brain phantoms with known true volumes for the different brain classes were generated. The algorithms' performance was checked by calculating the kappa index assessing similarities with the "ground truth" as well as multiclass type I and type II errors including misclassification rates. The impact of image segmentation algorithms on PVC was then quantified using clinical data. The segmented tissues of patients' brain MRI were given as input to the region of interest (RoI)-based geometric transfer matrix (GTM) PVC algorithm, and quantitative comparisons were made. The results of digital MRI phantom studies suggest that the use of HBSA produces the best performance for WM classification. For GM classification, it is suggested to use the EMS. Segmentation performed on clinical MRI data show quite substantial differences, especially when lesions are present. For the particular case of PVC, SPM2 and EMS algorithms show very similar results and may be used interchangeably. The use of HBSA is not recommended for PVC. The partial volume corrected activities in some regions of the brain show quite large relative differences when performing paired analysis on 2 algorithms, implying a careful choice of the segmentation algorithm for GTM-based PVC.     

Infant brain probability templates for MRI segmentation and normalization

Spatial normalization and segmentation of infant brain MRI data based on adult or pediatric reference data may not be appropriate due to the developmental differences between the infant input data and the reference data. In this study we have constructed infant templates and a priori brain tissue probability maps based on the MR brain image data from 76 infants ranging in age from 9 to 15 months. We employed two processing strategies to construct the infant template and a priori data: one processed with and one without using a priori data in the segmentation step. Using the templates we constructed, comparisons between the adult templates and the new infant templates are presented. Tissue distribution differences are apparent between the infant and adult template, particularly in the gray matter (GM) maps. The infant a priori information classifies brain tissue as GM with higher probability than adult data, at the cost of white matter (WM), which presents with lower probability when compared to adult data. The differences are more pronounced in the frontal regions and in the cingulate gyrus. Similar differences are also observed when the infant data is compared to a pediatric (age 5 to 18) template. The two-pass segmentation approach taken here for infant T1W brain images has provided high quality tissue probability maps for GM, WM, and CSF, in infant brain images. These templates may be used as prior probability distributions for segmentation and normalization; a key to improving the accuracy of these procedures in special populations.     

Age- and sex-related effects on the neuroanatomy of healthy elderly

Effects of age and sex, and their interaction on the structural brain anatomy of healthy elderly were assessed thanks to a cross-sectional study of a cohort of 662 subjects aged from 63 to 75 years. T1- and T2-weighted MRI scans were acquired in each subject and further processed using a voxel-based approach that was optimized for the identification of the cerebrospinal fluid (CSF) compartment. Analysis of covariance revealed a classical neuroanatomy sexual dimorphism, men exhibiting larger gray matter (GM), white matter (WM), and CSF compartment volumes, together with larger WM and CSF fractions, whereas women showed larger GM fraction. GM and WM were found to significantly decrease with age, while CSF volume significantly increased. Tissue probability map analysis showed that the highest rates of GM atrophy in this age range were localized in primary cortices, the angular and superior parietal gyri, the orbital part of the prefrontal cortex, and in the hippocampal region. There was no significant interaction between "Sex" and "Age" for any of the tissue volumes, as well as for any of the tissue probability maps. These findings indicate that brain atrophy during the seventh and eighth decades of life is ubiquitous and proceeds at a rate that is not modulated by "Sex".     

Semi-automatic brain region extraction (SABRE) reveals superior cortical and deep gray matter atrophy in MS

In multiple sclerosis (MS), atrophy occurs in various cortical and subcortical regions. However, it is unclear whether this is mostly due to gray (GM) or white matter (WM) loss. Recently, a new semi-automatic brain region extraction (SABRE) technique was developed to quantify parenchyma volume in 13 hemispheric regions. This study utilized SABRE and tissue segmentation to examine whether regional brain atrophy in MS is mostly due to GM or WM loss, correlated with disease duration, and moderated by disease course. We studied 68 MS patients and 39 normal controls with 1.5 T brain MRI. As expected, MS diagnosis was associated with significantly lower (P < 0.001) regional brain parenchymal fractions (RBPFs). While significant findings emerged in 11 GM comparisons, only four WM comparisons were significant. The largest mean RBPF percent differences between groups (MS < NC) were in the posterior basal ganglia/thalamus region (-19.3%), superior frontal (-15.7%), and superior parietal (-14.3%) regions. Logistic regression analyses showed GM regions were more predictive of MS diagnosis than WM regions. Eight GM RBPFs were significantly correlated (P < 0.001) with disease duration compared to only one WM region. Significant trends emerged for differences in GM, but not WM between secondary progressive (SP) and relapsing-remitting MS patients. Percent differences in GM between the two groups were largest in superior frontal (-9.9%), medial superior frontal (-6.5%), and superior parietal (-6.1%) regions, with SP patients having lower volumes. Overall, atrophy in MS is diffuse and mostly related to GM loss particularly in deep GM and superior frontal-parietal regions.     

Characterization of a sequential pipeline approach to automatic tissue segmentation from brain MR Images

Objective Quantitative analysis of gray matter and white matter in brain magnetic resonance imaging (MRI) is valuable for neuroradiology and clinical practice. Submission of large collections of MRI scans to pipeline processing is increasingly important. We characterized this process and suggest several improvements. Materials and methods To investigate tissue segmentation from brain MR images through a sequential approach, a pipeline that consecutively executes denoising, skull/scalp removal, intensity inhomogeneity correction and intensity-based classification was developed. The denoising phase employs a 3D-extension of the Bayes–Shrink method. The inhomogeneity is corrected by an improvement of the Dawant et al.’s method with automatic generation of reference points. The N3 method has also been evaluated. Subsequently the brain tissue is segmented into cerebrospinal fluid, gray matter and white matter by a generalized Otsu thresholding technique. Intensive comparisons with other sequential or iterative methods have been carried out using simulated and real images. Results The sequential approach with judicious selection on the algorithm selection in each stage is not only advantageous in speed, but also can attain at least as accurate segmentation as iterative methods under a variety of noise or inhomogeneity levels. Conclusion A sequential approach to tissue segmentation, which consecutively executes the wavelet shrinkage denoising, scalp/skull removal, inhomogeneity correction and intensity-based classification was developed to automatically segment the brain tissue into CSF, GM and WM from brain MR images. This approach is advantageous in several common applications, compared with other pipeline methods.     

Impact of acquisition protocols and processing streams on tissue segmentation of T1 weighted MR images

The segmentation of T1-weighted images into gray matter (GM), white matter (WM), and cerebrospinal fluid (CSF) is a fundamental processing step in neuroimaging, the results of which affect many other structural imaging analyses. Variability in the segmentation process can decrease the power of a study to detect anatomical differences, and minimizing such variability can lead to more robust results. This paper outlines a straightforward strategy that can be used (1) to select more optimal data acquisition and processing protocols and (2) to quantify the impact of such optimization. Using this approach with multiple scans of a single subject, we found that the choice of a segmentation algorithm had the largest impact on variability, while the choice of a pulse sequence had the second largest impact. The data indicate that the classification of GM is the most variable, and that the optimal protocol may differ across tissue types. Therefore, the intended use of segmentation data should play a role in optimization. Examples are provided to demonstrate that the minimization of variability is not sufficient for optimization; the overall accuracy of the approach must also be considered. Simple volumetric computations are included to illustrate the potential gain of optimization; these results show that volume estimates from optimal pathways were on average three times less variable than estimates from suboptimal pathways. Therefore, the simple strategy illustrated here can be applied to many studies to optimize tissue segmentation, which should lead to a net increase in the power of structural neuroimaging studies.     

Investigating association of brain volumes with intracranial capacity in schizophrenia

Intracranial volume (ICV) is usually treated as a global or nuisance covariate in almost all volumetric studies of schizophrenia. However, validation for this analytic method has seldom been accomplished. In this study, we aimed to determine the effects of ICV on gray matter (GM) and white matter (WM) volumes. Sixty-three patients with schizophrenia and sixty normal controls were recruited; and high resolution T1 weighted images were obtained by 3T-MRI. After segmentation and normalization of the images into GM, WM, and cerebrospinal fluid (CSF), multiple regression analyses of global GM and WM volumes were performed using explanatory variables such as diagnosis, ICV, and diagnosis-ICV interaction. In addition, associations between regional GM and WM volumes with ICV were also investigated using voxel-based morphometry (VBM). No significant interaction between diagnosis and ICV was found for global GM volume, whereas interactions were detected in restricted GM areas using VBM. On the other hand, an interaction between ICV and diagnosis was found in WM not only for regional volumes, but also for global WM volume. The regression slope of global WM volumes against ICV was steeper in patients with schizophrenia than in healthy controls. These results imply that ICV should be carefully evaluated in the analyses of volumetric studies of schizophrenia, especially when analyzing WM volumes.     

Mapping brain structure and personality in late adulthood

Cerebral gray matter (GM) volume decreases in normal aging with a parallel increase in intracranial cerebrospinal fluid (CSF) volume. There is considerable interindividual variation in these changes, and the consequences of age-related GM shrinkage and CSF expansion are unclear. The present study examined whether late adulthood brain structural differences are related to differences in temperament and character. Personality structures of 42 healthy aged adults (mean age 60 years) were examined together with global and regional GM, CSF, and white matter (WM) volumes calculated from structural magnetic resonance images using voxel-based morphometry (VBM). A positive relationship was seen between GM volume at the border of the temporal, parietal, and frontal cortices, and self-transcendence, a character personality trait that reflects mature creativity and spiritualism. The relationship remained significant after a conservative correction for multiple comparisons and it was seen both using uncorrected raw values and after a correction for the effects of age and sex. The results suggest that high self-transcendence, which has adaptive advantages in the later part of life, is associated with relatively greater temporal cortical GM volumes.     

Longitudinal changes in patients with traumatic brain injury assessed with diffusion-tensor and volumetric imaging

Traumatic brain injury (TBI) is associated with brain volume loss, but there is little information on the regional gray matter (GM) and white matter (WM) changes that contribute to overall loss. Since axonal injury is a common occurrence in TBI, imaging methods that are sensitive to WM damage such as diffusion-tensor imaging (DTI) may be useful for characterizing microstructural brain injury contributing to regional WM loss in TBI. High-resolution T1-weighted imaging and DTI were used to evaluate regional changes in TBI patients compared to matched controls. Patients received neuropsychological testing and were imaged approximately 2 months and 12.7 months post-injury. Paradoxically, neuropsychological function improved from Visit 1 to Visit 2, while voxel-based analyses of fractional anisotropy (FA), and mean diffusivity (MD) from the DTI images, and voxel-based analyses of the GM and WM probability maps from the T1-weighted images, mainly revealed significantly greater deleterious GM and WM change over time in patients compared to controls. Cross-sectional comparisons of the DTI measures indicated that patients have decreased FA and increased MD compared to controls over large regions of the brain. TBI affected virtually all of the major fiber bundles in the brain including the corpus callosum, cingulum, the superior and inferior longitudinal fascicules, the uncinate fasciculus, and brain stem fiber tracts. The results indicate that both GM and WM degeneration are significant contributors to brain volume loss in the months following brain injury, and also suggest that DTI measures may be more useful than high-resolution anatomical images in assessment of group differences.     

Multi-spectral brain tissue segmentation using automatically trained k-Nearest-Neighbor classification

Conventional k-Nearest-Neighbor (kNN) classification, which has been successfully applied to classify brain tissue in MR data, requires training on manually labeled subjects. This manual labeling is a laborious and time-consuming procedure. In this work, a new fully automated brain tissue classification procedure is presented, in which kNN training is automated. This is achieved by non-rigidly registering the MR data with a tissue probability atlas to automatically select training samples, followed by a post-processing step to keep the most reliable samples. The accuracy of the new method was compared to rigid registration-based training and to conventional kNN-based segmentation using training on manually labeled subjects for segmenting gray matter (GM), white matter (WM) and cerebrospinal fluid (CSF) in 12 data sets. Furthermore, for all classification methods, the performance was assessed when varying the free parameters. Finally, the robustness of the fully automated procedure was evaluated on 59 subjects. The automated training method using non-rigid registration with a tissue probability atlas was significantly more accurate than rigid registration. For both automated training using non-rigid registration and for the manually trained kNN classifier, the difference with the manual labeling by observers was not significantly larger than inter-observer variability for all tissue types. From the robustness study, it was clear that, given an appropriate brain atlas and optimal parameters, our new fully automated, non-rigid registration-based method gives accurate and robust segmentation results. A similarity index was used for comparison with manually trained kNN. The similarity indices were 0.93, 0.92 and 0.92, for CSF, GM and WM, respectively. It can be concluded that our fully automated method using non-rigid registration may replace manual segmentation, and thus that automated brain tissue segmentation without laborious manual training is feasible.     

An automated tool for detection of FLAIR-hyperintense white-matter lesions in Multiple Sclerosis

In Multiple Sclerosis (MS), detection of T2-hyperintense white matter (WM) lesions on magnetic resonance imaging (MRI) has become a crucial criterion for diagnosis and predicting prognosis in early disease. Automated lesion detection is not only desirable with regard to time and cost effectiveness but also constitutes a prerequisite to minimize user bias. Here, we developed and evaluated an algorithm for automated lesion detection requiring a three-dimensional (3D) gradient echo (GRE) T1-weighted and a FLAIR image at 3 Tesla (T). Our tool determines the three tissue classes of gray matter (GM) and WM as well as cerebrospinal fluid (CSF) from the T1-weighted image, and, then, the FLAIR intensity distribution of each tissue class in order to detect outliers, which are interpreted as lesion beliefs. Next, a conservative lesion belief is expanded toward a liberal lesion belief. To this end, neighboring voxels are analyzed and assigned to lesions under certain conditions. This is done iteratively until no further voxels are assigned to lesions. Herein, the likelihood of belonging to WM or GM is weighed against the likelihood of belonging to lesions. We evaluated our algorithm in 53 MS patients with different lesion volumes, in 10 patients with posterior fossa lesions, and 18 control subjects that were all scanned at the same 3T scanner (Achieva, Philips, Netherlands). We found good agreement with lesions determined by manual tracing (R2 values of over 0.93 independent of FLAIR slice thickness up to 6 mm). These results require validation with data from other protocols based on a conventional FLAIR sequence and a 3D GRE T1-weighted sequence. Yet, we believe that our tool allows fast and reliable segmentation of FLAIR-hyperintense lesions, which might simplify the quantification of lesions in basic research and even clinical trials.     

Semiautomatic brain region extraction: a method of parcellating brain regions from structural magnetic resonance images

Structural MR imaging has become essential to the evaluation of regional brain changes in both healthy aging and disease-related processes. Several methods have been developed to measure structure size and regional brain volumes, but many of these methods involve substantial manual tracing and/or landmark identification. We present a new technique, semiautomatic brain region extraction (SABRE), for the rapid and reliable parcellation of cortical and subcortical brain regions. We combine the SABRE parcellation with tissue compartment segmentation [NeuroImage 17 (2002) 1087] to produce measures of gray matter (GM), white matter (WM), ventricular CSF, and sulcal CSF for 26 brain regions. Because SABRE restricts user input to a few easily identified landmarks, inter-rater reliability is high for all volumes, with all coefficients between 0.91 and 0.99. To assess construct validity, we contrasted SABRE-derived volumetric data from healthy young and older adults. Results from the SABRE parcellation and tissue segmentation showed significant differences in multiple brain regions in keeping with regional atrophy described in the literature by researchers using lengthy manual tracing methods. Our findings show that SABRE is a reliable semiautomatic method for assessing regional tissue volumes that provides significant timesavings over purely manual methods, yet maintains information about individual cortical landmarks.     

Validating cortical surface analysis of medial prefrontal cortex.

This paper describes cortical analysis of 19 high resolution MRI subvolumes of medial prefrontal cortex (MPFC), a region that has been implicated in major depressive disorder. An automated Bayesian segmentation is used to delineate the MRI subvolumes into cerebrospinal fluid (CSF), gray matter (GM), white matter (WM), and partial volumes of either CSF/GM or GM/WM. The intensity value at which there is equal probability of GM and GM/WM partial volume is used to reconstruct MPFC cortical surfaces based on a 3-D isocontouring algorithm. The segmented data and the generated surfaces are validated by comparison with hand segmented data and semiautomated contours, respectively. The L(1) distances between Bayesian and hand segmented data are 0.05-0.10 (n = 5). Fifty percent of the voxels of the reconstructed surface lie within 0.12-0.28 mm (n = 14) from the semiautomated contours. Cortical thickness metrics are generated in the form of frequency of occurrence histograms for GM and WM labelled voxels as a function of their position from the cortical surface. An algorithm to compute the surface area of the GM/WM interface of the MPFC subvolume is described. These methods represent a novel approach to morphometric chacterization of regional cortex features which may be important in the study of psychiatric disorders such as major depression.     

Grey and white matter atrophy in early clinical stages of primary progressive multiple sclerosis

BACKGROUND: There is little information available on grey and white matter (GM and WM) atrophy in primary progressive multiple sclerosis (PPMS) and on their relationships with clinical and other magnetic resonance imaging (MRI) measures. AIM: To evaluate disease progression in the early phase of PPMS, focusing on axonal loss as assessed by volumetric MRI measures of WM and GM, and to determine their relationships with clinical outcomes and lesion load measures. METHODS: Forty-three patients with PPMS within 5 years of symptom onset and 45 control subjects were studied. Three-dimensional brain scans were acquired and segmented into WM, GM, and cerebrospinal fluid (CSF) using SPM99. Brain parenchymal (BPF), WM (WMF), and GM fractions (GMF) normalized against total intracranial volumes were estimated. T2-weighted (T2) and enhancing lesion loads were also determined. Expanded Disability Status Scale (EDSS) and Multiple Sclerosis Functional Composite (MSFC) scores were recorded in all patients. RESULTS: There were significant differences between patients and controls in BPF, WMF, and GMF values (P < 0.001). BPF (r = -0.469; P = 0.002) and WMF (r = -0.532; P < 0.001) but not GMF (r = -0.195; P = 0.2) correlated with EDSS scores. BPF (r = 0.518; P = 0.001), WMF (r = 0.483; P = 0.001), and GMF (r = 0.337; P = 0.031) correlated with MSFC scores. Correlations with enhancing lesion and T2 loads were only significant for BPF and WMF. CONCLUSIONS: Brain atrophy is seen in the early stages of PPMS and affects both GM and WM. WM atrophy appears more closely related to clinical outcome and WM focal damage than GM atrophy in this patient group.     

Validation of partial tissue segmentation of single-channel magnetic resonance images of the brain

We describe and evaluate a practical, automated algorithm based on local statistical mixture modeling for segmenting single-channel, T1-weighted volumetric magnetic resonance images of the brain into gray matter, white matter, and cerebrospinal fluid. We employed a stereological sampling method to assess, prospectively, the performance of the method with respect to human experts on 10 normal T1-weighted brain scans acquired with a three-dimensional gradient echo pulse sequence. The overall kappa statistic for the concordance of the algorithm with the human experts was 0.806, while that among raters, excluding the algorithm, was 0.802. The algorithm had better agreement with the modal expert decision (kappa = 0.878). The algorithm could not be distinguished from the experts by this measure. We also validated the algorithm on a simulated MR scan of a digital brain phantom with known tissue composition. Global gray matter and white matter errors were 1% and <1%, respectively, and correlation coefficients with the underlying tissue model were 0.95 for gray matter, 0.98 for white matter, and 0.95 for cerebrospinal fluid. In both approaches to validation, we evaluated both local and global performance of the algorithm. Human experts generated slightly higher global gray matter proportion estimates on the test brain scans relative to the algorithm (3.7%) and on the simulated MR scan relative to the true tissue model (4.4%). The algorithm underestimated gray in some subcortical nuclei which contain admixed gray and white matter. We demonstrate the reliability of the method on individual 1 NEX data sets of the test subjects, and its insensitivity to the precise values of initial model parameters. The output of this algorithm is suitable for quantifying cerebral cortical tissue, using a commonly performed commercial pulse sequence.