Peter Hastreiter

Quantification of, visualization of, and compensation for brain shift using intraoperative magnetic resonance imaging.

OBJECTIVEModern neuronavigation systems lack spatial accuracy during ongoing surgical procedures because of increasing brain deformation, known as brain shift. Intraoperative magnetic resonance imaging was used for quantitative analysis and visualization of this phenomenon. METHODSFor a total of 64 patients, we used a 0.2-T, open-configuration, magnetic resonance imaging scanner, located in an operating theater, for pre- and intraoperative imaging. The three-dimensional imaging data were aligned using rigid registration methods. The maximal displacements of the brain surface, deep tumor margin, and midline structures were measured. Brain shift was observed in two-dimensional image planes using split-screen or overlay techniques, and three-dimensional, color-coded, deformable surface-based data were computed. In selected cases, intraoperative images were transferred to the neuronavigation system to compensate for the effects of brain shift. RESULTSThe results demonstrated that there was great variability in brain shift, ranging up to 24 mm for cortical displacement and exceeding 3 mm for the deep tumor margin in 66% of all cases. Brain shift was influenced by tissue characteristics, intraoperative patient positioning, opening of the ventricular system, craniotomy size, and resected volume. Intraoperative neuronavigation updating (n = 14) compensated for brain shift, resulting in reliable navigation with high accuracy. CONCLUSIONWithout brain shift compensation, neuronavigation systems cannot be trusted at critical steps of the surgical procedure, e.g., identification of the deep tumor margin. Intraoperative imaging allows not only evaluation of and compensation for brain shift but also assessment of the quality of mathematical models that attempt to describe and compensate for brain shift.

Intraoperative compensation for brain shift

BACKGROUND Tumor removal, brain swelling, the use of brain retractors, and cerebrospinal-fluid drainage all result in an intraoperative brain deformation that is known as brain shift. Thus, neuronavigation systems relying on preoperative image data have a decreasing accuracy during the surgical procedure. Intraoperative image data represent the correct anatomic situation, so their use may compensate for the effects of brain shift. METHODS In a series of 16 brain tumor patients, we used intraoperative magnetic resonance (MR) imaging to obtain 3-D data, which were then transferred to the microscope-based neuronavigation system. With the help of bone fiducial markers these images were registered intraoperatively, updating the neuronavigation system. RESULTS In all patients the updating of the neuronavigation system with the intraoperative MR data was successful. It led to reliable neuronavigation with high accuracy; the mean registration error of the update procedure in all patients was 1.1 mm. The updating procedure added about 15 minutes to the operation time. In all patients the area suggestive of remaining tumor was reached and the additional tumor could be resected, resulting in a complete tumor removal in 14 patients. In the remaining patients extension of the tumor into eloquent brain areas prevented a complete excision. CONCLUSIONS The update of a neuronavigation system with intraoperative MR images reliably compensates for the effects of brain shift. This method allows completion of tumor removal in some difficult brain tumors.

Correlation of Sensorimotor Activation with Functional Magnetic Resonance Imaging and Magnetoencephalography in Presurgical Functional Imaging: A Spatial Analysis

In this study we investigated the spatial heterotopy of MEG and fMRI localizations after sensory and motor stimulation tasks. Both methods are frequently used to study the topology of the primary and secondary motor cortex, as well as a tool for presurgical brain mapping. fMRI was performed with a 1.5T MR system, using echo-planar imaging with a motor and a sensory task. Somatosensory and motor evoked fields were recorded with a biomagnetometer. fMRI activation was determined with a cross-correlation analysis. MEG source localization was performed with a single equivalent current dipole model and a current density localization approach. Distances between MEG and fMRI activation sites were measured within the same anatomical 3-D-MR image set. The central region could be identified by MEG and fMRI in 33 of 34 cases. However, MEG and fMRI localization results showed significantly different activation sites for the motor and sensory task with a distance of 10 and 15 mm, respectively. This reflects the different neurophysiological mechanisms: direct neuronal current flow (MEG) and secondary changes in cerebral blood flow and oxygenation level of activated versus non activated brain structures (fMRI). The result of our study has clinical implications when MEG and fMRI localizations are used for pre- and intraoperative brain mapping. Although both modalities are useful for the estimation of the motor cortex, a single modality may err in the exact topographical labeling of the motor cortex. In some unclear cases a combination of both methods should be used in order to avoid neurological deficits.

Interactive exploration of volume line integral convolution based on 3D-texture mapping

Line integral convolution (LIC) is an effective technique for visualizing vector fields. The application of LIC to 3D flow fields has yet been limited by difficulties to efficiently display and animate the resulting 3D-images. Texture-based volume rendering allows interactive visualization and manipulation of 3D-LIC textures. In order to ensure the comprehensive and convenient exploration of flow fields, we suggest interactive functionality including transfer functions and different clipping mechanisms. Thereby, we efficiently substitute the calculation of LIC based on sparse noise textures and show the convenient visual access of interior structures. Further on, we introduce two approaches for animating static 3D-flow fields without the computational expense and the immense memory requirements for pre-computed 3D-textures and without loss of interactivity. This is achieved by using a single 3D-LIC texture and a set of time surfaces as clipping geometries. In our first approach we use the clipping geometry to pre-compute a special 3D-LIC texture that can be animated by time-dependent color tables. Our second approach uses time volumes to actually clip the 3D-LIC volume interactively during rasterization. Additionally, several examples demonstrate the value of our strategy in practice.

Combining local and remote visualization techniques for interactive volume rendering in medical applications

For a comprehensive understanding of tomographic image data in medicine, interactive and high-quality direct volume rendering is an essential prerequisite. This is provided by visualization using 3D texture mapping which is still limited to high-end graphics hardware. In order to make it available in a clinical environment, we present a system which uniquely combines local desktop computers and remote high-end graphics hardware. In this context, we exploit the standard visualization capabilities to a maximum which are available in the clinical environment. For 3D representations of high resolution and quality we access the remote specialized hardware. Various tools for 2D and 3D visualization are provided which meet the requirements of a medical diagnosis. This is demonstrated with examples from the field of neuroradiology which show the value of our strategy in practice.

Magnetic source imaging supports clinical decision making in glioma patients

OBJECTIVE This study addresses the potential utility of preoperative functional imaging with magnetoencephalography (MEG) for the selection of glioma patients who are likely to benefit from resective surgical treatment regarding postoperative morbidity. METHODS One hundred and nineteen patients with gliomas adjacent to sensorimotor, visual and speech related brain areas were investigated preoperatively with a MAGNES II biomagnetometer. In each patient the pre-surgical evaluation was focussed on the visual, sensorimotor cortex and/or of the speech related brain areas. A grading system was then used according to the distance of the MEG activation sources to the nearest tumour border to determine the further treatment. The therapeutic options consisted in conservative treatment, stereotactic biopsy and/or a radiation and chemotherapy, substantial cytoreduction and the gross total removal of the lesion. RESULTS From 119 investigated patients, 55 patients (46.2%) were not considered for surgery due to tumour invasion to functional cortex. Sixty four patients (53.8%) were chosen for resective surgery. In the surgical group only four patients (6.2%) suffered from neurological deterioration. CONCLUSIONS Magnetic source imaging (MSI) proved to be a valuable help in the clinical decision making process of lesions adjacent to functional important brain areas. The relative high number of patients in whom MSI warns of the postoperative crippling sequelae may lead to a better selection of patients who benefit from resective surgery. This method may help to find the patients for whom conservative treatment seems to be more favourable concerning quality of life in the surviving time.

Integrated registration and visualization of medical image data

Different imaging modalities give insight to vascular, anatomical and functional information which assist diagnosis and therapy planning in medicine. Registration and consecutive visualization allow to combine the image data and thereby convey more meaningful images to the clinician. Applying a voxel based approach based on mutual information, accurate and retrospective registration is provided. However, optimization and consecutive visualization procedures require a huge amount of trilinear interpolation operations to re-sample the data. Ensuring fast performance which is fundamental for medical routine, we suggest an integrated approach which takes advantage of the imaging and texture mapping subsystem of graphics computers. All trilinear interpolation is completely performed with hardware assisted 3D texture mapping. The 1D and 2D histograms of the datasets which are necessary for the calculation of mutual information are obtained with different hardware accelerated imaging operations. For the simultaneous and interactive visualization of the registered datasets a new approach was developed which allows for versatile fusion operations. Using similar procedures supported by hardware, contributes considerably to accelerate registration and visualization. Implementing our approach within a previously presented framework (Hastreiter et al., 1996) (Sommer et al., 1998) based on OpenInventor and OpenGL provides intuitive manipulation. Clinical examples show the value of our approach in practice.

Classification of neurovascular compression in typical hemifacial spasm: three-dimensional visualization of the facial and the vestibulocochlear nerves

OBJECT In this paper, the authors introduce a method of noninvasive anatomical analysis of the facial nerve-vestibulocochlear nerve complex and the depiction of the variable vascular relationships by using 3D volume visualization. With this technique, a detailed spatial representation of the facial and vestibulocochlear nerves was obtained. Patients with hemifacial spasm (HFS) resulting from neurovascular compression (NVC) were examined. METHODS A total of 25 patients (13 males and 12 females) with HFS underwent 3D visualization using magnetic resonance (MR) imaging with 3D constructive interference in a steady state (CISS). Each data set was segmented and visualized with respect to the individual neurovascular relationships by direct volume rendering. Segmentation and visualization of the facial and vestibulocochlear nerves were performed with reference to their root exit zone (REZ), as well as proximal and distal segments including corresponding blood vessels. The 3D visualizations were interactively compared with the intraoperative situation during microvascular decompression (MVD) to verify the results with the observed microneurosurgical anatomy. RESULTS Of the 25 patients, 20 underwent MVD (80%). Microvascular details were recorded on the affected and unaffected sides. On the affected sides, the anterior inferior cerebellar artery (AICA) was the most common causative vessel. The posterior inferior cerebellar artery, vertebral artery, internal auditory artery, and veins at the REZ of the facial nerve (the seventh cranial nerve) were also found to cause vascular contacts to the REZ of the facial nerve. In addition to this, the authors identified three distinct types of NVC within the REZ of the facial nerve at the affected sides. The authors analyzed the varying courses of the vessels on the unaffected sides. There were no bilateral clinical symptoms of HFS and no bilateral vascular compression of the REZ of the facial nerve. The authors discovered that the AICA is the most common vessel that interferes with the proximal and distal portions of the facial nerve without any contact between vessels and the REZ of the facial nerve on the unaffected sides. CONCLUSIONS Three-dimensional visualization by direct volume rendering of 3D CISS MR imaging data offers the opportunity of noninvasive exploration and anatomical categorization of the facial nerve-vestibulocochlear nerve complex. Furthermore, it proves to be advantageous in establishing the diagnosis and guiding neurosurgical procedures by representing original MR imaging patient data in a 3D fashion. This modality provides an excellent overview of the entire neurovascular relationship of the cerebellopontine angle in each case.

Hybrid Visualization for White Matter Tracts using Triangle Strips and Point Sprites

Diffusion tensor imaging is of high value in neurosurgery, providing information about the location of white matter tracts in the human brain. For their reconstruction, streamline techniques commonly referred to as fiber tracking model the underlying fiber structures and have therefore gained interest. To meet the requirements of surgical planning and to overcome the visual limitations of line representations, a new real-time visualization approach of high visual quality is introduced. For this purpose, textured triangle strips and point sprites are combined in a hybrid strategy employing GPU programming. The triangle strips follow the fiber streamlines and are textured to obtain a tube-like appearance. A vertex program is used to orient the triangle strips towards the camera. In order to avoid triangle flipping in case of fiber segments where the viewing and segment direction are parallel, a correct visual representation is achieved in these areas by chains of point sprites. As a result, high quality visualization similar to tubes is provided allowing for interactive multimodal inspection. Overall, the presented approach is faster than existing techniques of similar visualization quality and at the same time allows for real-time rendering of dense bundles encompassing a high number of fibers, which is of high importance for diagnosis and surgical planning

Visualization of white matter tracts with wrapped streamlines

Diffusion tensor imaging is a magnetic resonance imaging method which has gained increasing importance in neuroscience and especially in neurosurgery. It acquires diffusion properties represented by a symmetric 2nd order tensor for each voxel in the gathered dataset. From the medical point of view, the data is of special interest due lo different diffusion characteristics of varying brain tissue allowing conclusions about the underlying structures such as while matter tracts. An obvious way to visualize this data is to focus on the anisotropic areas using the major eigenvector for tractography and rendering lines for visualization of the simulation results. Our approach extends this technique to avoid line representation since lines lead 10 very complex illustrations and furthermore are mistakable. Instead, we generate surfaces wrapping bundles of lines. Thereby, a more intuitive representation of different tracts is achieved.