Head holder for PET, CT, and MR studies

A new head holder for the fixation and repositioning of the patient head in PET, CT, and MR scanners has been designed and tested. With this device, a bidimensional correlation between functional and anatomical brain images can also be obtained. Head fixation and repositioning are achieved using the patient's dental morphology as an anatomical reference. With a dentistic material, a mold of the patient's teeth is obtained in a few minutes. The molding substance rests on a plastic support, fixed to the head holder. Each time the patient undergoes a new study, his/her personal mold is used, ensuring accurate head repositioning. External markers fixed on the head holder (made visible in lateral PET and CT projection images, midsaggital MR images, and also on the axial images) make it possible to record and recognize the angular orientation and the position of the brain in the three-dimensional space, to correlate images of the same patient obtained with different neuroimaging modalities, and to accurately reposition patients for follow-up studies. The head holder was tested on several subjects. Fixation and repositioning accuracy of within 2.5 mm were achieved in the three-dimensional space. Orientation accuracy was 1 degree.     

A combined region growing and deformable model method for extraction of closed surfaces in 3D CT and MRI scans

Image segmentation of 3D medical images is a challenging problem with several still not totally solved practical issues, such as noise interference, variable object structures and image artifacts. This paper describes a hybrid 3D image segmentation method which combines region growing and deformable models to obtain accurate and topologically preserving surface structures of anatomical objects of interest. The proposed strategy starts by determining a rough but robust approximation of the objects using a region-growing algorithm. Then, the closed surface mesh that encloses the region is constructed and used as the initial geometry of a deformable model for the final refinement. This integrated strategy provides an alternative solution to one of the flaws of traditional deformable models, achieving good refinements of internal surfaces in few steps. Experimental segmentation results of complex anatomical structures on both simulated and real data from MRI scans are presented, and the method is assessed by comparing with standard reference segmentations of head MRI. The evaluation was mainly based on the average overlap measure, which was tested on the segmentation of white matter, corresponding to a simulated brain data set, showing excellent performance exceeding 90% accuracy. In addition, the algorithm was applied to the detection of anatomical head structures on two real MRI and one CT data set. The final reconstructions resulting from the deformable models produce high quality meshes suitable for 3D visualization and further numerical analysis. The obtained results show that the approach achieves high quality segmentations with low computational complexity.     

An automated method for rotational correction and centering of three-dimensional functional brain images.

The display and analysis of functional brain images often benefit from head rotational correction and centering. An automated method was developed to align brain PET images into a standard three-dimensional orientation. The algorithm performs transverse and coronal rotational correction as well as centering of a brain image set. Optimal rotational correction and centering are determined by maximizing a bilateral hemispheric similarity index, the stochastic sign change criterion. Testing of this algorithm on simulated symmetrical brain image sets showed errors less than 1.0 degree and 0.5 pixels for rotational correction and centering, respectively. With actual PET data, the algorithm results correlated well with those obtained by visual inspection. Testing on asymmetrical brain image sets with simulated lesions indicated that performance of the algorithm is not sensitive to focal asymmetries. This automated method provides objective, reproducible image alignment into a standard orientation and facilitates subsequent data analysis techniques for functional brain images.     

Dual-modality brain PET-CT image segmentation based on adaptive use of functional and anatomical information

Dual medical imaging modalities, such as PET-CT, are now a routine component of clinical practice. Medical image segmentation methods, however, have generally only been applied to single modality images. In this paper, we propose the dual-modality image segmentation model to segment brain PET-CT images into gray matter, white matter and cerebrospinal fluid. This model converts PET-CT image segmentation into an optimization process controlled simultaneously by PET and CT voxel values and spatial constraints. It is innovative in the creation and application of the modality discriminatory power (MDP) coefficient as a weighting scheme to adaptively combine the functional (PET) and anatomical (CT) information on a voxel-by-voxel basis. Our approach relies upon allowing the modality with higher discriminatory power to play a more important role in the segmentation process. We compared the proposed approach to three other image segmentation strategies, including PET-only based segmentation, combination of the results of independent PET image segmentation and CT image segmentation, and simultaneous segmentation of joint PET and CT images without an adaptive weighting scheme. Our results in 21 clinical studies showed that our approach provides the most accurate and reliable segmentation for brain PET-CT images.     

Real-time 3D image registration for functional MRI.

Subject head movements are one of the main practical difficulties with brain functional MRI. A fast, accurate method for rotating and shifting a three-dimensional (3D) image using a shear factorization of the rotation matrix is described. Combined with gradient descent (repeated linearization) on a least squares objective function, 3D image realignment for small movements can be computed as rapidly as whole brain images can be acquired on current scanners. Magn Reson Med 42:1014-1018, 1999.     

A simple method to improve image nonuniformity of brain MR images at the edges of a head coil

One of the major sources of image nonuniformity in the high field MR scanners is the radiofrequency (RF) coil inhomogeneity. It degrades conspicuity of lesion(s) in the MR images of the brain and surrounding tissues and reduces accuracy of image postprocessing particularly at the edges of the coil. In this investigation, we have devised and tested a simple method to correct for nonuniformity of MR images of the brain at the edges of the RF head coil. Initially, a cylindrical oil phantom, which fit exactly in the head coil, was scanned on a 1.5 T imager. Then, a correction algorithm identified a reference pixel value in the phantom at the most homogeneous region of the RF coil. Next, every pixel inside the phantom was normalized relative to this reference value. The resulting set of coefficients or "correction matrices" was obtained for different types of MR contrast agent. Finally, brain MR images of normal subjects and multiple sclerosis patients were acquired and processed by the corresponding correction matrices obtained with different pulse sequences. Application of correction matrices to brain MR images showed a gain in pixel intensity particularly in the slices at the edge of the coil.     

Gamma camera-mounted anatomical X-ray tomography: technology, system characteristics and first images

Scintigraphic diagnosis, based on functional image interpretation, becomes more accurate and meaningful when supported by corresponding anatomical data. In order to produce anatomical images that are inherently registered with images of emission computerised tomography acquired with a gamma camera, an X-ray transmission system was mounted on the slip-ring gantry of a GEMS Millennium VG gamma camera. The X-ray imaging system is composed of an X-ray tube and a set of detectors located on opposite sides of the gantry rotor that moves around the patient along with the nuclear detectors. A cross-sectional anatomical transmission map is acquired as the system rotates around the patient in a manner similar to a third-generation computerised tomography (CT) system. Following transmission, single-photon emission tomography (SPET) or positron emission tomography (PET) coincidence detection images are acquired and the resultant emission images are thus inherently registered to the anatomical maps. Attenuation correction of the emission images is performed with the same anatomical maps to generate transmission maps. Phantom experiments of system performance and examples of first SPET and coincidence detection patient images are presented. Despite limitations of the system when compared with a state of the art CT scanner, the transmission anatomical maps allow for precise anatomical localisation and for attenuation correction of the emission images.     

Gamma camera-mounted anatomical X-ray tomography: technology, system characteristics and first images

Scintigraphic diagnosis, based on functional image interpretation, becomes more accurate and meaningful when supported by corresponding anatomical data. In order to produce anatomical images that are inherently registered with images of emission computerised tomography acquired with a gamma camera, an X-ray transmission system was mounted on the slip-ring gantry of a GEMS Millennium VG gamma camera. The X-ray imaging system is composed of an X-ray tube and a set of detectors located on opposite sides of the gantry rotor that moves around the patient along with the nuclear detectors. A cross-sectional anatomical transmission map is acquired as the system rotates around the patient in a manner similar to a third-generation computerised tomography (CT) system. Following transmission, single-photon emission tomography (SPET) or positron emission tomography (PET) coincidence detection images are acquired and the resultant emission images are thus inherently registered to the anatomical maps. Attenuation correction of the emission images is performed with the same anatomical maps to generate transmission maps. Phantom experiments of system performance and examples of first SPET and coincidence detection patient images are presented. Despite limitations of the system when compared with a state of the art CT scanner, the transmission anatomical maps allow for precise anatomical localisation and for attenuation correction of the emission images. Received 27 October 1999 and in revised form 10 February 2000     

Region-growing segmentation of brain vessels: an atlas-based automatic approach

PURPOSE: To propose an atlas-based method that uses both phase and magnitude images to integrate anatomical information in order to improve the segmentation of blood vessels in cerebral phase-contrast magnetic resonance angiography (PC-MRA). MATERIAL AND METHODS: An atlas of the whole head was developed to store the anatomical information. The atlas divides a magnitude image into several vascular areas, each of which has specific vessel properties. It can be applied to any magnitude image of an entire or nearly entire head by deformable matching, which helps to segment blood vessels from the associated phase image. The segmentation method used afterwards consists of a topology-preserving, region-growing algorithm that uses adaptive threshold values depending on the current region of the atlas. This algorithm builds the arterial and venous trees by iteratively adding voxels that are selected according to their grayscale value and the variation of values in their neighborhood. The topology preservation is guaranteed because only simple points are selected during the growing process. RESULTS: The method was performed on 40 PC-MRA images of the brain. The results were validated using maximum-intensity projection (MIP) and three-dimensional surface rendering visualization, and compared with results obtained with two non-atlas-based methods. CONCLUSION: The results show that the proposed method significantly improves the segmentation of cerebral vascular structures from PC-MRA. These experiments tend to prove that the use of vascular atlases is an effective way to optimize vessel segmentation of cerebral images.     

Automatic registration of brain magnetic resonance images based on Talairach reference system

PURPOSE: To demonstrate a robust registration method of brain magnetic resonance (MR) images based on the Talairach reference system with automatic determinations of the fiducial points. MATERIALS AND METHODS: Eight specified landmark points of the Talairach reference system are determined after successfully extracting the midsagittal plane from three-dimensional MR imaging (MRI) data. Projection information of the image intensity is used to determine the midline of the cerebrum in axial and coronal view images, which is a necessary step for extraction of the midsagittal plane. To find the landmarks of anterior commissure (AC) and posterior commissure (PC) in the midsagittal plane, we adopt two-step shape matching that properly finds locations of the corpus callosum (CC), and then AC and PC, respectively. The shape matching is performed on the edge-enhanced midsagittal plane image to minimize dependency on image intensity variation. Other landmark points of the Talairach reference system are determined by fitting the intensity curve of the cutview with the Gaussian model. RESULTS: The proposed method automatically determines seven landmark points, except the inferior point (IP), and the brain MR images can be successfully registered with the Talairach reference system. CONCLUSION: The suggested registration method can be applied to any MR images for functional studies. It can also be applied to patients unless their brains are highly deformed or have a highly deformed CC.     

Proportional localization system for anatomical interpretation of cerebral computed tomograms

The existence of individual variations in size and shape of the human brain constitutes a problem for the anatomical interpretation of brain reconstructed images obtained from scanning devices; it is, for example, responsible for most of the inaccuracies in reading CT scans. One way to account for these variations is to use a proportional localization system. In the 1960s a group of neurosurgeons developed such a system based on two pivotal intracerebral structures, the anterior and the posterior commissures; they published an atlas consisting of horizontal, coronal, and sagittal brain sections interpreted in the proportional system. The atlas also included standard proportional brain schemes based on anatomical and radiological studies on large numbers of individuals. In this article we report a target localization experiment that we carried out to determine if this atlas could be used as a reference for a more accurate interpretation of CT and, eventually, of positron emission tomography (PET) and nuclear magnetic resonance (NMR) scans. Ten radiopaque small targets were inserted through the skull in the cortex of three cadavers; head CT was performed, and the atlas was used for predicting the cortical location of the targets seen on the CT images: The predictions were confirmed. These results strongly support the use of the proportional atlas for the interpretation of CT as well as of PET and NMR scans.     

Neuroanatomical localization for clinical SPECT perfusion brain imaging: a practical proportional grid method.

For the purpose of facilitating anatomical localization in interpretation of 99Tcm-hexamethylpropyleneamine oxime (HMPAO) brain single photon emission tomographic (SPECT) scans, a stereotaxic proportional grid system was applied in the form of an interactive computer program. This method takes advantage of a rotating gamma camera system which permits planar scout imaging for the determination of anatomical reference lines, and standardization of tomographic slices for brain size. Using measurements made on a lateral planar HMPAO image, proportional grids were constructed onto standardized transaxial images. This method was implemented for 33 clinical HMPAO SPECT studies. It required less than 15 min of an operator's time. This simple and practical neuroanatomical localization technique can be instrumental as an aid to the interpretation of routine clinical HMPAO SPECT images.     

Cortical region of interest definition on SPECT brain images using X-ray CT registration

Summary We present a method for brain single photon emission computed tomography (SPECT) analysis based on individual registration of anatomical (CT) and functional (¹³³Xe regional cerebral blood flow) images and on the definition of three-dimensional functional regions of interest. Registration of CT and SPECT is performed through adjustment of CT-defined cortex limits to the SPECT image. Regions are defined by sectioning a cortical ribbon on the CT images, copied over the SPECT images and pooled through slices to give 3D cortical regions of interest. The proposed method shows good intra- and interobserver reproducibility (regional intraclass correlation coefficient 0.98), and good accuracy in terms of repositioning (3.5 mm) as compared to the SPECT image resolution (14 mm). The method should be particularly useful for analysing SPECT studies when variations in brain anatomy (normal or abnormal) must be accounted for.     

Three-dimensional alignment of functional and morphological tomograms

A method has been developed to create corresponding brain slices from morphological [CT, magnetic resonance (MR)] and functional [positron emission tomography (PET), single photon emission computed tomography] tomographic studies in individual patients. It does not require special headholders or definition of specific landmarks and is fully retrospective. Three-dimensional image registration in corresponding orientation is achieved by linear interpolation of original slices and a variety of interactively controlled video display options. These include simultaneous display of multiple slices and brain cuts in all three dimensions for comparison of positioning. Brain contours in one imaging modality may be enhanced by appropriate filtering and superimposed onto reference images of another modality. Matching accuracy depends on image resolution; misalignment of 4 mm was detected unambiguously in sample studies (fluoro-2-deoxy-D-glucose PET matched with MR). The technique is equally well applicable to normals and to patients with structural brain lesions. Additional options for shaded surface display enhance the power to identify neuroanatomical structures in functional image analysis. As demonstrated in the example of MR-guided PET, this modeling procedure can be successfully used for identification of brain structures on functional images, even in patients with pathologically altered brain morphology.     

Correlation methods for the centering, rotation, and alignment of functional brain images

Simple methods are described using correlation analysis to rotate functional brain images to a standard vertical orientation, identify the antero-posterior centerline, and align multiple images from the same brain level. Image rotation and centering are performed by determining the angle of rotation and centerline coordinate that result in maximal left-right correlation. Testing of this method on sets of multiple images acquired simultaneously through different brain levels suggests that the optimal rotation can be determined within 1 degree and the centerline within 0.3 mm. Spatial alignment of two or more images from the same brain level of a single subject is accomplished by finding the translation and rotation that yield the highest correlation between the images. Testing of the alignment method on sets of simultaneously acquired images at multiple brain levels suggests that the optimal translation can be determined within 0.45-0.69 mm and the optimal rotation within 0.8 degrees. These methods are completely objective and can easily be automated.     

Multimodality imaging of brain structures for stereotactic surgery

An image analysis system was developed for stereotactic neurosurgery that allows the simultaneous display of brain images from different imaging devices obtained in different orientations. The system is based on a stereotactic frame and a microcomputer and features an easy user interface together with point registration and region of interest analysis in three-dimensional space. A dynamic multi-image environment allows for simultaneous display of magnetic resonance, computed tomography, digital subtraction angiography, and positron emission tomography images in multiple windows, adjusted for common coordinates with reference to stereotactic frame fiducial markers. Linkages between images allow information interchange between different modalities and different views: Points and regions defined in one image can be transferred to others, and cursor coordinates in one image can be calculated and dynamically projected in other images. Phantom studies show that the system distortions are minor and that the system is suitable for clinical use. The system provides exceptional advantages over previous imaging procedures for stereotactic surgery.     

Improved image registration by using fourier interpolation

This paper presents a technique for performing two-dimensional rigid-body image registration for functional magnetic resonance images (fMRI). The method provides accurate motion correction without local distortion. The approach is to perform the translation and rotation in the Fourier domain. For images sampled on a grid, such as in echo-planar imaging (EPI), one potential stumbling block to this approach is the computational burden of reconstruction, since the rotated image will no longer be on the Cartesian grid. A method of approximating rotations via local translations (shearing) is presented, which keeps the data on the Cartesian grid. This can provide quite accurate approximations with only a moderate amount of computation. A mean squared error (MSE) criterion is used for determining the registration parameters. This method is tested on several sets of simulated images and shown to have an accuracy ranging from 0.02 to 0.3 pixels for images with SNRs ranging from 100 to 10, respectively. They techniques have been tested on several sets of images. They are shown to work well on real subjects, for both echo-planar and spiral data acquisition schemes. The techniques are used in an activation study in which the subject moved his head during image collection. After use of this registration technique, the activation is easily detected.     

Three-dimensionla images of the brain surface from standard CT examination

Three-dimensional (3D) image rendering was performed using data from standard CT examinations of the head. In this new method for 3D imaging of the brain surface interactive image segmentation and integral rendering were combined. Interactive segmentation operations meant that 3D images could be obtained within a short period of time. When combined with integral rendering this permits good-quality 3D views, even when generated from standard thick slices, together with visualisation of both the shape and densities of the brain surface. Surface infarctions can be evaluated in terms of their anatomical localisation to gyri and sulci, thus allowing their relationship to functional areas to be better defined. The technique can be an additional, easily obtainable tool, even in routine practice, for a better understanding of neurological signs. Correspondence to: R. Pozzi Mucelli     

A stereotaxic method of anatomical localization by means of H2 15O positron emission tomography applicable to the brain activation study in cats: Registration of images of cerebral blood flow to brain atlas

In the neuronal activation study of normal animals, precise anatomical correlation, preferentially to a detailed brain atlas, is required for the activation foci co-registration. To obtain precise regional correlation between H₂ ¹⁵O-PET images and the brain atlas, a method of stereotaxic image reorientation was applied to an activation study with vibrotactile stimulation. Cats anesthetized with halothane underwent repeated measurements of regional cerebral blood flow (rCBF) in the resting condition and during vibration of the right forepaw. The image set was adjusted three-dimensionally to the atlas. The postmortem brain was sectioned according to the atlas planes. The activated areas were determined by the stimulus-minus-resting subtraction images, and the areas were projected to the atlas. The PET images of the cat brain were compatible both to the postmortem brain slices and to the brain atlas. The activation foci obtained from the subtraction images corresponded to the area around the coronal sulcus, which is electrophysiologically known as the primary sensory area as described in the atlas. There were precise regional correlations between the PET image and anatomy in a PET activation study of the cat by means of stereotaxic image reorientation.     

Combining anatomy and function: the path to true image fusion

Modern imaging technologies visualize different aspects of disease in a non-invasive way. Considerable progress has been made in the fusion of images from different imaging modalities using software approaches. One goal of fusion software is to align anatomical and functional images and allow improved spatial localization of abnormalities. The resulting correlation of the anatomical and functional images may clarify the nature of the abnormality and help diagnose or stage the underlying disease. Whereas successful image fusion software has been developed for the brain, only limited success has been achieved for image alignment in other parts of the body. The development and current status of alternative approaches are presented. Dual-modality imaging is described with devices where two modalities are combined and mounted in a single gantry. The use of existing scanner technology ensures that no compromises are made in the clinical efficacy of either the anatomical or functional imaging modality. A combined positron emission tomography (PET) and computed tomography (CT) scanner has been developed and is undergoing clinical evaluation. Combining PET with MR is technologically more challenging because of the strong magnetic fields restricting the use of certain electronic components. An overview of the current status and future prospects of dual-modality imaging devices is presented.