Technical aspects of computer-assisted magnetic resonance (MR) image-guided surgery are discussed with highlighted applications. The applications presented with clinical feasibility studies include multi-modality image registration for prostate cancer biopsy, organ deformation tracking for craniotomy, thermal monitoring for interstitial laser ablation therapy of brain tumor. A preliminary study of MR-compatible surgical manipulator is also reported. The clinical feasibility study of computer-assisted MR-guided surgery suggested the its potential contribution to accuracy, less-invasiveness, efficacy in surgical procedures.
We developed intraoperative 3-dimensional (3D) echoic image using the information of 2-dimentional echo and image analyzing system (SAS-200). The principle of SAS-200 is that magnetic field sensor, which is setting on echoic probe, recognize the position of the probe as 3D coordinate. This method was adopted to the patients with liver tumors and the donors for living-related liver transplantation. Using this method, the relation between a main tumor and feeding vessels was visualized as 3D-US image.A 3D-US image was useful to make an accurate navigation for liver surgery.
Several types of supporting systems for surgical planning have been reported. This paper describes a new surgical supporting device which features image superimposition using a half-mirror. By using the device, graphic image of internal organs can be superimposed on the patient's live image. In order that the device allows users to move around, the position of the user toward the patient was measured consecutively by a small CCD camera placed on the top of a head mounted display (HMD). The CCD camera imaged a patient model with pre-arranged markers on its surface and the position of the model was measured with the accuracy of around 0.5cm by an image processing software. Based on this positional data, three dimensional model data were processed on a workstation to produce graphic image for superimposition. The image was presented to the user through the HMD which features the “see-through” function. The error in navigation by the device was found to be as large as 1cm, particularly at the peripheral view field, and this problem must be solved in the future study.
Until recently, we have only inferred the movement of each skeletal muscle from electromyogram measurements (EMG), anatomical information, and observations of the human body surface. However, under the present circumstances it is problematic to comprehend the connection and interaction between the human skeleton and skeletal muscles or between human skeletal muscles themselves. Thus we aimed to develop a method to quantitatively visualize these interactions in space and time sequential domains by utilizing computer graphic techniques. First, each human skeleton and muscle model was reconstructed from MRI data sets. Second, coordinates derived from a mesh framework of a muscle were converted into algorithm. This algorithm reduced all coordinates of the muscle framework while maintaining the original muscle shape. Afterwards, we were able to construct a realistic skeletal muscle model. This muscle model can contract or extend while avoiding the bone or adjacent skeletal muscle by taking physical interference into account. We also made a series of skeletal muscle models for each bone. As a result, we could observe the activity and contribution of each muscle in bone movement. In the future, by applying this system we will be able to monitor the movements of people with physical disorders.