[Image-guided surgery].
Initially, stereotactic surgery was developed to treat functional brain diseases only. The localisation of targets was based on stereotactic atlases and radiographs. The introduction of computer based imaging techniques, such as CT and MRI, have offered the possibility to "see" anomalies and to approach them stereotactically. The working principle of such a procedure consists of three steps. It is assumed that brain tissue does not move with respect to the skull. 1. Acquisition of the images and their registration with the patient, usually based on a series of reference points (fiducials) that belong to a stereotactic localizer attached to the base ring of the stereotactic frame, 2. planning and simulation of the surgical intervention, mostly based on two-dimensional (2D) images resliced along arbitrary directions, and 3. intra-operative guidance of the instruments mounted onto the stereotactic frame. This procedure has continuously been updated by new image acquisition techniques, 3D visualization and frameless stereotaxy. 1. PET, MRI and angiography (DSA, MRA) have been used in addition to CT. Moreover, such images of different modalities can automatically be fused without using a stereotactic frame or other artificial fiducials. 2. 3D images for surgery planning have become available, a feature that has proved to be very useful for the cerebral blood vessels. 3. The stereotactic frame can be replaced by a robot arm or an optical guidance system. Registration of the instruments with the patient is then performed by using markers in the cranial bone or on the scalp, or by means of intra-operative images such as radiographs or video images. Recently the use of registered video images has resulted in a number of experiments with "improved reality" and telesurgery. The same working principle has shown to be useful for bone and bone related surgery. As in brain surgery, the prerequisites of rigidity and immobility with respect to a reference are satisfied. Because bone structures are rigid and can easily be outlined in CT images, 3D graphical as well as stereolithographic representations can be produced for the purpose of planning and even for navigation. Unfortunately, for most organs or soft tissue the above conditions of inflexibility and fixed position with respect to a reference are not fulfilled. Real time imaging, partly due to the introduction of the "open" MR scanner, may offer a solution. It can be expected that interventional diagnostic imaging will become increasingly important in the future, also for neurosurgery.