Towards a uniform visual enabling concept for brain surgery

Sankt Augustin 10 July 1998 A comprehensive team of the German National Research Centre for Information Technology (GMD) started in May 1997 with the project Visual Enabling for Precision Surgery (VEP). The ultimate goal is to integrate various existing visualization and registration techniques into one single computer-based three-dimensional image guidance system for neurosurgical procedures. The research is specifically focused on the rather complex laser-induced interstitial thermotherapy (LITT) for brain tumour coagulation. The partners are trying to set up a configuration for pre-operative diagnosis and planning, as well as intra-operative therapy and control, with incorporation of an open magnetic resonance imaging (MRI) system. In this VEP project, research people, industry and hospitals are striving to achieve a perfect interaction between the human medical specialist and the computer.

Advertisement

A comprehensive team of the German National Research Centre for Information Technology (GMD) started in May 1997 with the project Visual Enabling for Precision Surgery (VEP). The ultimate goal is to integrate various existing visualization and registration techniques into one single computer-based three-dimensional image guidance system for neurosurgical procedures. The research is specifically focused on the rather complex laser-induced interstitial thermotherapy (LITT) for brain tumour coagulation. The partners are trying to set up a configuration for pre-operative diagnosis and planning, as well as intra-operative therapy and control, with incorporation of an open magnetic resonance imaging (MRI) system. In this VEP project, research people, industry and hospitals are striving to achieve a perfect interaction between the human medical specialist and the computer.

In essence, the VEP system constitutes a 3D reference scene of combined 3D and 2D data sets, originating from different imaging sources, to offer the surgeon accurate orientation in the planning and performance of neurosurgical procedures. Since medical and synthetic images are merged with the real vision of the patient's brain, the interdisciplinary project team is actually dealing with augmented reality. Successful brain surgery implies sophisticated imaging techniques, such as computed tomography (CT), MRI, positron-emission tomography (PET), and ultrasound. All these modalities have to be timely presented to the surgeon in a well structured way, giving him just the information he needs for the next step to be taken during the intervention.

A closer look at the LITT procedure learns us that the patient gets an optical fibre implanted into the brain tumour for coagulation with laser energy. In the pre-operative phase, the surgeon determines both the exposition times and the distribution parameters to avoid neighbouring brain tissue to be affected by the coagulation. For that purpose, MR or CT based image planes are combined into one 3D data set, in order to exactly identify the position of the tumour with respect to the surrounding brain areas. The latter are defined, using several imaging techniques as well as angiography for the vascular structures. The 3D data set is fused with this information to an integrated representation, registered with supplementary land marks and enhanced by synthetic structures, which display the selected access path and the surgical instruments.

At that point, everything is ready for experimental testing and simulation of the thermotherapy, in order to control its effect through visualization. This stage is followed by the real intervention. First, the surgeon has to precisely fit a catheter or optical fibre tip into the tumour. He controls this operation by means of 2D images from the open MRI system. For accurate spatial and structural orientation, he compares the action with both the pre-operative 3D reference scene and the synthetic imagery, showing the access path and surgical devices. Second, he coagulates the tumour tissue with laser energy, while constantly monitoring the process of temperature distribution, blood perfusion, dose adaptation, laser exposition time, and so on, according to the initial plan.

The pre-operative data set doesn't always correspond with the exact position of the tumour during the intervention, due to insertion of various surgical instruments changing the form of the soft surrounding tissues. In order to acquire up-to-date images, intervention microscopes are used, as well as the open MR system, which offers a larger perspective ahead and to the sides than the microscope but a poorer quality than the pre-operative MR. As a result, the open MR slices are fused with the pre-operative high resolution 3D reference scene, to optimally control the therapy proceedings. Next to this, it is also possible to introduce ultrasound as a method for real time intra-operative feedback. Integration with other visual material however is difficult because of the inferior resolution and the image distortions.

The average physician is still a bit reluctant to benefit from the current imaging modalities, since he experiences them as a hindering barrier instead of a useful tool in his contact with the patient. A cognitive and adequate system like VEP, however, is able to provide the surgeon with the right images at the right time, while keeping his head and hands free for the patient during the intervention. He only needs to visually compare the computer-enhanced information with the actual situation to decide on the following step in the procedure. Additional voice input will enable him to interact with the system. Supplementary training facilities by means of interactive CD-Roms will convince sceptic doctors of the system's visual enabling power. The GMD people are offering you a full description of the VEP project at their Web site.


Leslie Versweyveld

[Medical IT News][Calendar][Virtual Medical Worlds Community][News on Advanced IT]