Search

A number of noninvasive imaging techniques have been used for the evaluation of bone marrow, including magnetic resonance imaging (MRI) and bone marrow scintigraphy. The appearance of bone marrow on MRI varies considerably depending on the proportion of red and yellow marrow, and the composition of the red marrow and its distribution with relation to age and sex. The composition of bone marrow also can vary under physiological and pathological conditions. MRI is a highly sensitive technique for evaluating the bone marrow, but it is limited in its practical use for whole-body bone marrow screening. Bone marrow scintigraphy with radiolabeled compounds such as technetium-99m-labeled nanocolloid and monoclonal antibodies has the advantage of evaluating the entire bone marrow, and has been used for the diagnosis of various bone marrow disorders. In addition, (18)F-fluorodeoxyglucose positron emission tomography (FDG-PET) imaging can be used to evaluate bone marrow metabolism and disease and to provide information about the state of the primary tumor, lymph nodes, and distant metastases. Understanding of the appearance of normal bone marrow, including age- and sex-specific differences with each of these imaging modalities, is essential to permit accurate diagnosis of benign and malignant bone marrow disorders. We present a review of MRI and scintigraphy of normal bone marrow with some emphasis on FDG-PET imaging in assessing marrow activity in normal and abnormal states and also present preliminary data regarding normal age-related changes in bone marrow through use of FDG-PET, as well as the role of segmentation of bone marrow on MRI for quantitative calculation of the metabolic volumetric product for red marrow metabolism using FDG-PET.

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.

The Medical Image Processing Group at the University of Pennsylvania has been developing (and distributing with source code) medical image analysis and visualization software systems for a long period of time. Our most recent system, 3DVIEWNIX, was first released in 1993. Since that time, a number of significant advancements have taken place with regard to computer platforms and operating systems, networking capability, the rise of parallel processing standards, and the development of open-source toolkits. The development of CAVASS by our group is the next generation of 3DVIEWNIX. CAVASS will be freely available and open source, and it is integrated with toolkits such as Insight Toolkit and Visualization Toolkit. CAVASS runs on Windows, Unix, Linux, and Mac but shares a single code base. Rather than requiring expensive multiprocessor systems, it seamlessly provides for parallel processing via inexpensive clusters of work stations for more time-consuming algorithms. Most importantly, CAVASS is directed at the visualization, processing, and analysis of 3-dimensional and higher-dimensional medical imagery, so support for digital imaging and communication in medicine data and the efficient implementation of algorithms is given paramount importance.

The Medical Image Processing Group (MIPG) at the University of Pennsylvania has been developing (and distributing with source code) medical image analysis and visualization software systems for a long period of time. Our most recent system, 3DVIEWNIX, was first released in 1993. Since that time, a number of significant advancements have taken place with regard to computer platforms and operating systems, networking capability, the rise of parallel processing standards, and the development of open source toolkits. The development of CAVASS by our group is the next generation of 3DVIEWNIX. CAVASS will be freely available, open source, and is integrated with toolkits such as ITK and VTK. CAVASS runs on Windows, Unix, and Linux but shares a single code base. Rather than requiring expensive multiprocessor systems, it seamlessly provides for parallel processing via inexpensive COWs (Cluster of Workstations) for more time consuming algorithms. Most importantly, CAVASS is directed at the visualization, processing, and analysis of 3D and higher dimensional medical imagery, so support for DICOM data and the efficient implementation of algorithms is given paramount importance.

This paper presents (1) an improved hierarchical method for segmenting the component tissue regions in fast spin echo T2 and PD images of the brain of Multiple Sclerosis (MS) patients, and (2) a methodology to characterize the disease utilizing the distributions of standardized T2 and PD intensities in the segmented tissue regions. First, the background intensity inhomogeneities are corrected and the intensity scales are standardized for all acquired images. The segmentation method imposes a feedback-like procedure on our previously developed hierarchical brain tissue segmentation method. With gradually simplified patterns in images and stronger evidences, pathological objects are recognized and segmented in an interplay fashion. After the brain parenchymal (BP) mask is generated, an under-estimated gray matter mask (uGM) and an over-estimated white matter mask (oWM) are created. Pure WM (PWM) and lesion (LS) masks are extracted from the all-inclusive oWM mask. By feedback, accurate GM and WM masks are subsequently formed. Finally, partial volume regions of GM and WM as well as Dirty WM (DWM) masks are generated. Intensity histograms and their parameters (peak height, peak location, and 25th, 50th and 75th percentile values) are computed for both T2 and PD images within each tissue region. Tissue volumes are also estimated. Spearman correlation coefficient rank test is then utilized to assess if there exists a trend between clinical states and the image-based parameters. This image analysis method has been applied to a data set consisting of 60 patients with MS and 20 normal controls. LS related parameters and clinical Extended Disability Status Scale (EDSS) scores demonstrate modest correlations. Almost every intensity-based parameter shows statistical difference between normal control and patient groups with a level better than 5%. These results can be utilized to monitor disease progression in MS.

The purpose of this work was to optimize and increase the accuracy of tissue segmentation of the brain magnetic resonance (MR) images based on multispectral 3D feature maps. We used three sets of MR images as input to the in-house developed semi-automated 3D tissue segmentation algorithm: proton density (PD) and T2-weighted fast spin echo and, T1-weighted spin echo. First, to eliminate the random noise, non-linear anisotropic diffusion type filtering was applied to all the images. Second, to reduce the nonuniformity of the images, we devised and applied a correction algorithm based on uniform phantoms. Following these steps, the qualified observer "seeded" (identified training points) the tissue of interest. To reduce the operator dependent errors, cluster optimization was also used; this clustering algorithm identifies the densest clusters pertaining to the tissues. Finally, the images were segmented using k-NN (k-Nearest Neighborhood) algorithm and a stack of color-coded segmented images were created along with the connectivity algorithm to generate the entire surface of the brain. The application of pre-processing optimization steps substantially improved the 3D tissue segmentation methodology.

Our aim was to develop an accurate multispectral tissue segmentation method based on 3D feature maps. We utilized proton density (PD), T2-weighted fast spin-echo (FSE), and T1-weighted spin-echo images as inputs for segmentation. Phantom constructs, cadaver brains, an animal brain tumor model and both normal human brains and those from patients with either multiple sclerosis (MS) or primary brain tumors were analyzed with this technique. Initially, misregistration, RF inhomogeneity and image noise problems were addressed. Next, a qualified observer identified samples representing the tissues of interest. Finally, k-nearest neighbor algorithm (k-NN) was utilized to create a stack of color-coded segmented images. The inclusion of T1 based images, as a third input, produced significant improvement in the delineation of tissues. In MS, our 3D technique was found to be far superior to that based on any combination of 2D feature maps (P < 0.001). We identified at least two distinctly different classes of lesions within the same MS plaque, representing different stages of the disease process. Further, we obtained the regional distribution of MS lesion burden and followed its changes over time. Neuropsychological aberrations were the clinical counterpart of the structural changes detected in segmentation. We could also delineate the margins of benign brain tumors. In malignant tumors, up to four abnormal tissues were identified: 1) a solid tumor core, 2) a cystic component, 3) edema in the white matter, and 4) areas of necrosis and hemorrhage. Subsequent neurosurgical exploration confirmed the distribution of tissues as predicted by this analysis.