Elias R. Melhem

Imaging cortical association tracts in the human brain using diffusion‐tensor‐based axonal tracking

Diffusion‐tensor fiber tracking was used to identify the cores of several long‐association fibers, including the anterior (ATR) and posterior (PTR) thalamic radiations, and the uncinate (UNC), superior longitudinal (SLF), inferior longitudinal (ILF), and inferior fronto‐occipital (IFO) fasciculi. Tracking results were compared to existing anatomical knowledge, and showed good qualitative agreement. Guidelines were developed to reproducibly track these fibers in vivo. The interindividual variability of these reconstructions was assessed in a common spatial reference frame (Talairach space) using probabilistic mapping. As a first illustration of this technical capability, a reduction in brain connectivity in a patient with a childhood neurodegenerative disease (X‐linked adrenoleukodystrophy) was demonstrated. Magn Reson Med 47:215–223, 2002.

Classification of brain tumor type and grade using MRI texture and shape in a machine learning scheme

The objective of this study is to investigate the use of pattern classification methods for distinguishing different types of brain tumors, such as primary gliomas from metastases, and also for grading of gliomas. The availability of an automated computer analysis tool that is more objective than human readers can potentially lead to more reliable and reproducible brain tumor diagnostic procedures. A computer‐assisted classification method combining conventional MRI and perfusion MRI is developed and used for differential diagnosis. The proposed scheme consists of several steps including region‐of‐interest definition, feature extraction, feature selection, and classification. The extracted features include tumor shape and intensity characteristics, as well as rotation invariant texture features. Feature subset selection is performed using support vector machines with recursive feature elimination. The method was applied on a population of 102 brain tumors histologically diagnosed as metastasis ( 24 ), meningiomas ( 4 ), gliomas World Health Organization grade II ( 22 ), gliomas World Health Organization grade III ( 18 ), and glioblastomas ( 34 ). The binary support vector machine classification accuracy, sensitivity, and specificity, assessed by leave‐one‐out cross‐validation, were, respectively, 85%, 87%, and 79% for discrimination of metastases from gliomas and 88%, 85%, and 96% for discrimination of high‐grade (grades III and IV) from low‐grade (grade II) neoplasms. Multiclass classification was also performed via a one‐vs‐all voting scheme. Magn Reson Med, 2009.

Leftward asymmetry in relative fiber density of the arcuate fasciculus.

Left hemispheric language dominance is well established, but the structural substrate for this functional asymmetry is uncertain. We report a strong asymmetry in the relative fiber density of the arcuate fasciculus, a white matter pathway associated with language that connects the frontal, temporal, and parietal lobes. Measured with diffusion tensor tractography, nearly all study participants demonstrated greater relative fiber density in the left arcuate fasciculus than in the right arcuate fasciculus. In comparison, we found no asymmetry in the corticospinal tract, an important white matter pathway with no known role in language. Combined with data on volumetric and activation asymmetry, greater connectivity may provide the elements of a neural system model for language lateralization.

Diffusion tensor imaging of periventricular leukomalacia shows affected sensory cortex white matter pathways

Abstract—The authors used diffusion-tensor imaging to examine central white matter pathways in two children with spastic quadriplegic cerebral palsy. Corticospinal tracts projecting from cortex to brainstem resembled controls. In contrast, posterior regions of the corpus callosum, internal capsule, and corona radiata were markedly reduced, primarily in white matter fibers connected to sensory cortex. These findings suggest that the motor impairment in periventricular leukomalacia may, in part, reflect disruption of sensory connections outside classic pyramidal motor pathways.

Computer-Assisted Segmentation of White Matter Lesions in 3D MR images, Using Support Vector Machine

RATIONALE AND OBJECTIVES Brain lesions, especially white matter lesions (WMLs), are associated with cardiac and vascular disease, but also with normal aging. Quantitative analysis of WML in large clinical trials is becoming more and more important. MATERIALS AND METHODS In this article, we present a computer-assisted WML segmentation method, based on local features extracted from multiparametric magnetic resonance imaging (MRI) sequences (ie, T1-weighted, T2-weighted, proton density-weighted, and fluid attenuation inversion recovery MRI scans). A support vector machine classifier is first trained on expert-defined WMLs, and is then used to classify new scans. RESULTS Postprocessing analysis further reduces false positives by using anatomic knowledge and measures of distance from the training set. CONCLUSIONS Cross-validation on a population of 35 patients from three different imaging sites with WMLs of varying sizes, shapes, and locations tests the robustness and accuracy of the proposed segmentation method, compared with the manual segmentation results from two experienced neuroradiologists.

High-Dimensional Spatial Normalization of Diffusion Tensor Images Improves the Detection of White Matter Differences: An Example Study Using Amyotrophic Lateral Sclerosis

Spatial normalization of diffusion tensor images plays a key role in voxel-based analysis of white matter (WM) group differences. Currently, it has been achieved using low-dimensional registration methods in the large majority of clinical studies. This paper aims to motivate the use of high-dimensional normalization approaches by generating evidence of their impact on the findings of such studies. Using an ongoing amyotrophic lateral sclerosis (ALS) study, we evaluated three normalization methods representing the current range of available approaches: low-dimensional normalization using the fractional anisotropy (FA), high-dimensional normalization using the FA, and high-dimensional normalization using full tensor information. Each method was assessed in terms of its ability to detect significant differences between ALS patients and controls. Our findings suggest that inadequate normalization with low-dimensional approaches can result in insufficient removal of shape differences which in turn can confound FA differences in a complex manner, and that utilizing high-dimensional normalization can both significantly minimize the confounding effect of shape differences to FA differences and provide a more complete description of WM differences in terms of both size and tissue architecture differences. We also found that high-dimensional approaches, by leveraging full tensor features instead of tensor-derived indices, can further improve the alignment of WM tracts.

Analysis of MRI patterns aids prediction of progression in X-linked adrenoleukodystrophy

Background: X-linked adrenoleukodystrophy (X-ALD) has variants with widely different outcomes, hampering clinical counseling and evaluation of therapies. Objective: To evaluate the degree to which MRI patterns can predict lesion progression. Methods: Two hundred six boys and men with cerebral X-ALD (median age 12.2 years, mean age 18.5 years, age range 1.7 to 73.8 years) were studied. In 140 individuals, follow-up MRI were available. Data after bone marrow transplantation (BMT) were excluded. The patterns of MRI abnormalities were subdivided into five groups based on the anatomic location of the initial T2 signal hyperintensity (pattern 1: parieto-occipital white matter, pattern 2: frontal white matter, pattern 3: corticospinal tract, pattern 4: cerebellar white matter, pattern 5: concomitant parieto-occipital and frontal white matter). The X-ALD MRI Severity Scale, a 34-point scale previously described, was used in the analysis. Results: Pattern 1 patients had rapid progression if contrast enhancement was present and if the MRI abnormality manifested at an early age. The latter was also true for pattern 2 patients. Based on these variables, predictive formulas were constructed for these two patterns using multiple regressions. MRI progression was much slower in pattern 3 and 4 patients, whereas in the few pattern 5 patients, it was more rapid than in any other of the patterns. Patterns 1 and 5 occurred mainly in childhood, patterns 2 and 4 in adolescence, and pattern 3 in adults. Conclusions: MRI progression in X-ALD depends on patient age, initial MRI Severity Scale score, and anatomic location of the lesion. When used in combination, these data aid the prediction of disease course and the selection of patients for BMT.

Grading of CNS neoplasms using continuous arterial spin labeled perfusion MR imaging at 3 Tesla.

To differentiate glioma grade based on blood flow measured using continuous arterial spin labeled (CASL) perfusion MRI, implemented at 3 Tesla for improved signal‐to‐noise ratio (SNR) and spin labeling effect.

Differentiation between glioblastomas and solitary brain metastases using diffusion tensor imaging

The purpose of this study is to determine whether diffusion tensor imaging (DTI) metrics including tensor shape measures such as linear and planar anisotropy coefficients (CL and CP) can help differentiate glioblastomas from solitary brain metastases. Sixty-three patients with histopathologic diagnosis of glioblastomas (22 men, 16 women, mean age 58.4 years) and brain metastases (13 men, 12 women, mean age 56.3 years) were included in this study. Contrast-enhanced T1-weighted, fluid-attenuated inversion recovery (FLAIR) images, fractional anisotropy (FA), apparent diffusion coefficient (ADC), CL and CP maps were co-registered and each lesion was semi-automatically subdivided into four regions: central, enhancing, immediate peritumoral and distant peritumoral. DTI metrics as well as the normalized signal intensity from the contrast-enhanced T1-weighted images were measured from each region. Univariate and multivariate logistic regression analyses were employed to determine the best model for classification. The results demonstrated that FA, CL and CP from glioblastomas were significantly higher than those of brain metastases from all segmented regions (p<0.05), and the differences from the enhancing regions were most significant (p<0.001). FA and CL from the enhancing region had the highest prediction accuracy when used alone with an area under the curve of 0.90. The best logistic regression model included three parameters (ADC, FA and CP) from the enhancing part, resulting in 92% sensitivity, 100% specificity and area under the curve of 0.98. We conclude that DTI metrics, used individually or combined, have a potential as a non-invasive measure to differentiate glioblastomas from metastases.

Multiparametric Tissue Characterization of Brain Neoplasms and Their Recurrence Using Pattern Classification of MR Images

RATIONALE AND OBJECTIVES Treatment of brain neoplasms can greatly benefit from better delineation of bulk neoplasm boundary and the extent and degree of more subtle neoplastic infiltration. Magnetic resonance imaging (MRI) is the primary imaging modality for evaluation before and after therapy, typically combining conventional sequences with more advanced techniques such as perfusion-weighted imaging and diffusion tensor imaging (DTI). The purpose of this study is to quantify the multiparametric imaging profile of neoplasms by integrating structural MRI and DTI via statistical image analysis methods to potentially capture complex and subtle tissue characteristics that are not obvious from any individual image or parameter. MATERIALS AND METHODS Five structural MRI sequences, namely, B0, diffusion-weighted images, fluid-attenuated inversion recovery, T1-weighted, and gadolinium-enhanced T1-weighted, and two scalar maps computed from DTI (ie, fractional anisotropy and apparent diffusion coefficient) are used to create an intensity-based tissue profile. This is incorporated into a nonlinear pattern classification technique to create a multiparametric probabilistic tissue characterization, which is applied to data from 14 patients with newly diagnosed primary high-grade neoplasms who have not received any therapy before imaging. RESULTS Preliminary results demonstrate that this multiparametric tissue characterization helps to better differentiate among neoplasm, edema, and healthy tissue, and to identify tissue that is likely to progress to neoplasm in the future. This has been validated on expert assessed tissue. CONCLUSION This approach has potential applications in treatment, aiding computer-assisted surgery by determining the spatial distributions of healthy and neoplastic tissue, as well as in identifying tissue that is relatively more prone to tumor recurrence.