We present an object-oriented environment built on top of OpenGL and intended for virtual planning of image-guided surgery. We enumerate typical requirements from this field and show how they are coped with by our developments. A number of already derived surgery planning applications demonstrates the flexibility of our environment.
The possible complications of a pedicle screw fixation system include injury to neurologic and vascular structures resulting from inaccurate placement of the instrumentation. In a review of 617 surgical cases in which pedicle screw implants were used, Esses and co-authors reported a overall complication rate of 27.4%. The most common intraoperative problem was unrecognized screw misplacement (5.2%). Fracturing the pedicle during screw insertion and iatrogenic cerebrospinal fluid leak occurred in 4.2% of cases. Such a complication rate is not acceptable in clinical practice. In this paper, we discuss a computer-assisted spine surgery system designed for real-time intraoperative localization of surgical instruments on precaptured images used during surgery. Localization was achieved by combining image-guided stereotaxis with advanced opto-electronic position sensing techniques. The insertion of pedicle screws can be directly monitored by interactive navigation using specially equipped surgical tools. Our preliminary results showed no misplacement of pedicle screw, which have further confirmed the clinical potential of this system.
Magnetoencephalography is a new functional brain imaging technology which allows the non-invasive determination of cortical and subcortical neuronal assemblies involved in the processing of pain. In healthy subjects 4 pain relevant activity centers could be reliably identified which were activated by brief radiant infrared laser heat pulses: the ipsilateral SI area, the bilateral SII areas and a premotoric frontal activity. The localisation these cortical areas was based on the individual head and cortex anatomy.
A technique is presented for elastic alignment applicable to human brains. The transformation which minimizes the distance measure D(u) between template and reference is determined, thereby simultaneously satisfying smoothness constraints derived from an elastic potential known from the theory of kontinuum mechanics. The resulting partial differential equations, with up to 3·220 unknowns are directly solved for each voxel, that is, without interpolation, by an adapted full multigrid-method (FMG) providing a perfect alignment. For further increases of resolution, the full advantages of the FMG are maintained, that is, parallelization and linear effort with O(N), N being the number of grid-points.
This paper describes a robust fully automatic method for segmenting the brain from head MR images, which works even in the presence of RF inhomogeneities. It has been successful in segmenting the brain in every slice from head images acquired from three different MRI scanners, using different resolution images and different echo sequences. The three-stage integrated method employs image processing techniques based on anisotropic filters, „snakes“ contouring techniques, and a-priori knowledge. First the background noise is removed leaving a head mask, then a rough outline of the brain is found, and finally the rough brain outline is refined to a final mask.
Modern computers are unable to store in main memory the complete data of high resolution medical images. Even on secondary memory (disk), such large datasets are sometimes stored in a compressed form. At rendering time, parts of the volume are requested by the rendering algorithm and are loaded from disk. If one is not careful, the same regions may be (decompressed and) loaded to memory several times. Instead, a coherent algorithm should be designed that minimizes this thrashing and optimizes the time and effort spent to (uncompress and) load the volume. We present an algorithm that divides the volume into cubic cells, each (compressed and) stored on disk, in contrast to the more common slice-based storage. At rendering time, each cell is allocated a queue of rays. For a sequence of images, all rays are spawned and queued at the cells they intersect first. Cells are loaded, one at a time, in front-to-back (FTB) order. A loaded cell is rendered by all rays found in its queue. We analyze the algorithm in detail and demonstrate its advantages over existing ray casting volume rendering methods.
This paper describes our recent work on real-time Surgery Simulation using Fast Finite Element models of linear elasticity [1]. In addition we discuss various improvements in terms of speed and realism.
Interventional, fast MR images were registered to pre-operative CT and MR images of a primate cadaver. Localisation of an MR compatible, visible, flexible endoscope was achieved to within 2mm in the nasal cavity. Orthogonal views of the endoscope's position can be displayed alongside oriented video endoscope images to give 3D visualisation.
A computer integrated prostatectomy system named PROBOT has been produced to aid in the resection of prostatic tissue. The system is image guided, model based, with simulation and online video monitoring. The development and trial of the system have not only demonstrated the successful robotic imaging and resection of the prostate, but have also shown that soft tissue robotic surgery in general, can be successful.
In order to study the effects of acute electrical cochlear stimulation on the topography of the cat auditory cortex, we measured reflectance changes by means of optical imaging of intrinsic signals. Following single pulse electrical stimulation at selected sites of a multichannel implant device, we found topographically restricted response areas. Systematic variation of the stimulation pairs and thus of the cochlear frequency sites revealed a systematic and corresponding shift of the response areas. Increasingly higher stimulation currents evoked increasingly larger response areas resulting in decreasing spatial, i.e. cochleotopic selectivity. The results indicate that optical imaging intrinsic signals is useful to visualize effects of cochlear stimulation, which results in a profound cochleotopic selectivity. The implications of these findings are discussed in respect to underlying mechanisms of sound sensation mediated by cochlear implants.
A neural network approach has been developed to predict diagnostic image information to assist in the assessment of coronary artery disease. The predicted information represents the redistribution (or reversibility) of perfusion in the myocardium. A multilayer, backpropagation neural network is trained to predict the redistribution information from two other types of images: stress perfusion and myocardial thickening using SPECT imaging. The significance of this approach is two-fold: (i) the predicted reversibility information obviates the additional acquisition of delayed images (with the patient at rest), and (ii) the neural network approach represents a novel way with which to analyze and predict images from other images. This paper presents the methods that underlie the approach, and discusses the most recent experimental results that demonstrate its viability.
This paper concerns the visualization of group differences in the shapes of outlines without landmarks. Two helpful normalizations, Procrustes registration and the thin-plate spline, are available from the biometrics of landmarks. In this new context, the associated displays have both advantages and disadvantages of legibility or interpretation. I introduce a filter design intended to compromise among the many roles such visualizations are expected to serve. These concerns are illustrated using shapes of the corpus callosum in midsagittal images of the human brain, a data set of some relevance for the neurobiology of schizophrenia.
Craniofacial surgery requires careful preoperative planning in order to restore functionality and to improve the patient's aesthetics. In this article we present two different tissue models, integrated in an interactive surgical simulation system, which allow the preoperative visualization of the patient's postoperative appearance. We combine a reconstruction of the patient's skull from computer tomography with a laser scan of the skin, and bring them together into the system. The surgical procedure is simulated and its impact on the skin is calculated with either a MASS SPRING or a FINITE ELEMENT tissue model.
We describe techniques for visualising and measuring the asymmetry of the cerebral hemispheres based on MRI scans. These techniques improve on previous approaches in two ways: firstly, measurements are not limited to voxel discretisation scales; secondly, symmetry measurements are inherently 3D. This avoids the errors in slice-based measurements arising from shape distortions introduced by misalignment between the head and MRI machine. We focus on ‘sparse’ MRI data sets in which the slices are non-contiguous, since these constitute an important source of data in longitudinal neurological studies. Two visualisations of asymmetry are presented. The first is 3D and enables an immediate qualitative appreciation of the disposition of the asymmetry. The second is a 2D rendering of the symmetry map, a quantitative measure of asymmetry across the brain.
We consider elastic registration of medical image data based on thin-plate splines using a set of corresponding anatomical point landmarks. Previous work on this topic has concentrated on using interpolation schemes. Such schemes force the corresponding landmarks to exactly match each other and assume that the landmark positions are known exactly. However, in real applications the localization of landmarks is always prone to some error. Therefore, to take into account these localization errors, we have investigated the application of an approximation scheme which is based on regularization theory. This approach generally leads to a more accurate and robust registration result. In particular, outliers do not disturb the registration result as much as is the case with an interpolation scheme. Also, it is possible to individually weight the landmarks according to their localization uncertainty. In addition to this study, we report on investigations into semi-automatic extraction of anatomical point landmarks.
In this paper, we present the cross-validation of three deformable template superimposition techniques a,b and c, used to study 3D CT images of the bony skull. Method (a) relies on the manual identification by anatomists of anthropometric landmarks, method (b) on “crest lines”, which have a pure geometric definition, and method (c) is based on 3D non-rigid intensity based matching. We propose to define and compute a distance between methods a, b, c, and also to compute three representations I a , Ī b , Ī c of an “average” skull model based superimposition via these three methods. The overall aim is to determine if the three methods, all developed independently, give mutually coherent image superimposition results.
We present an approach for analyzing the morphology of anatomical structures of the brain, which uses an elastic transformation to normalize brain images into a reference space. The properties of this transformation are used as a quantitative description of the size and shape of brain structures; inter-subject comparisons are made by comparing the transformations themselves. The utility of this technique is demonstrated on a small group of eigth men and eight women, by comparing the shape and size of their corpus callosum, the main structure connecting the two hemispheres of the brain. Our analysis found the posterior region of the female corpus callosum to be larger than its corresponding region in the males. The average callosal shape of each group was also found, demonstrating visually the callosal shape differences between the two groups.
This surgical planning and training system integrates an electronic atlas of brain structure with a virtual stereotactic frame and patient data, for intuitive, reach-in manipulation. The objective is plans prepared faster; better, more accurate choice of target points; improved avoidance of sensitive structures; fewer sub-optimal frame attachments; and speedier, more effective training. If validated by clinical study now under way, this will improve medical efficacy and reduce costs.
We have used a twelve degrees of freedom global affine registration technique incorporating a multiresolution optimisation of mutual information to register MR and CT images with uncertain voxel dimensions and CT gantry tilt. Visual assessment indicates improved accuracy, without loss of robustness compared to rigid body registration.
L-systems are explored as a tool for the modelling of both branching and non-branching anatomical structures. Branching anatomical structures are modelled using parametric L-systems that are normally used to describe tree growth and structure. Non-branching structures are modelled using generalised cylinder L-systems. Both representations produce highly parametric models of the anatomy that can be used for efficient image generation and model matching. Examples are given to illustrate the power of this approach in modelling three-dimensional anatomy such as the ribcage, the lung airway, and the heart.
