Umeå University Medical Dissertations

Background The birth of stereotactic functional neurosurgery in 1947 was to a great extent dependent on the development of ventriculography. The last decades have witnessed a renaissance of functional stereotactic neurosurgery in the treatment of patients with movement disorders. Initially, these procedures were largely based on the same imaging technique that had been used since the birth of this technique, and that is still used in some centres. The introduction of new imaging modalities such as Computed Tomography (CT) and Magnetic Resonance Imaging (MRI) provided new potentials, but also new challenges for accurate identification and visualisation of the targets in the basal ganglia and the thalamus with an urge to thoroughly evaluate and optimize the stereotactic targeting technique, as well as evaluate accurately in stereotactic space the location and extent of stereotactic Radiofrequency (RF) lesions and the exact position of deep brain stimulation (DBS) electrodes. Aims To study the differences between CT and MRI regarding indirect atlas coordinates in thalamic and pallidal procedures and to evaluate and validate visualisation of the pallidum and the subthalamic nucleus in view of direct targeting irrespective of atlas-derived coordinates. Furthermore, to evaluate the contribution of various RF parameters to the size of stereotactic lesions, as well as the impact of size and location on clinical outcome. Method The surgical coordinates defined in relation to the landmarks of the 3 ventricle, of the targets in the pallidum and ventrolateral thalamus were compared between CT and MRI in 34 patients. In another 48 other patients direct visualization of the pallidum was evaluated and compared to indirect atlas based targeting. The possibility and versatility of visualizing the subthalamic nucleus (STN) on short acquisition T2-weighted MRI were evaluated in a multicentre study, and the use of alternative anatomical landmarks in identification of the STN was demonstrated in another study. In 46 patients CT and MRI were compared with respect to the volume of the visible RF lesions in the thalamus and the pallidum. The volume was analysed with regard to the parameters of the RF coagulation. The location and size of the lesions were further evaluated in relation to clinical outcome. Results Minor deviations, mostly frame dependent, were seen between MRI and CT coordinates of brain targets. The rostro-caudal direction of these deviations were such that they would be easily accounted for during surgery. MRI using a proton density sequence provided detailed images of the pallidal structures, which demonstrated considerable inter-individual variations in relation to the landmarks of the 3 ventricle. By using a direct visualization of the target, each patient will act as his or her own atlas, avoiding the uncertainties of atlas-based targeting. The STN could be visualized on various brands of MRI machines in 8 centres in 6 countries with good discrimination and with a short acquisition time, allowing direct visual targeting. The same scanning technique could be used for postoperative localization of the implanted electrodes. In cases where the lateral and inferior borders of the STN cannot be easily distinguished on MRI, the Sukeroku sign and the dent internal-capsule-sign signs might be useful. The volume of a stereotactic RF lesion could be as accurately assessed by CT as by MRI. The lesion ́s size was most strongly influenced by the temperature of coagulation. The lesions ́ volumes were however rather scattered and difficult to predict in the individual patient based solely on the RF coagulation parameters. For thalamotomy, the results on tremor was not related to the lesion ́s volume. For pallidotomy, larger and more posterior-ventral lesions had better effect on akinesia while effects on tremor and dyskinesias were not related to size or location of the lesions.. Conclusions The minor deviations of MRI from CT coordinates can be accounted for during surgery, why MRI can obviate the need of CT in some procedures. Direct visualized targeting on MRI of the pallidum is superior to atlas based targeting. The targets in the pallidum and the STN, as well as the location of the electrodes, can be well visualized with short acquisition MRI. When borders of the STN are poorly defined on MRI the Sukeroku sign and the dent internal-capsule-sign signs proved to be useful. The volumes of RF lesions can be accurately assessed by both stereotactic thin slice CT and MRI. The size of these lesions is most strongly influenced by the temperature of coagulation, but difficult to predict in the individual patient based on the coagulation parameters. Within certain limits, there were no clear relationships between lesions ́ volume and location and clinical effects of thalamotomies and pallidotomies.