Improving DBS targeting using 3D visualization of intraoperative stimulation tests

PosterInternational audienceBackground: In the past three decades, over 100,000 movement disorder patients like Parkinson’s disease (PD) have been treated by deep brain stimulation (DBS). Despite an increasing use of DBS, the fundamental mechanisms underlying therapeutic and adverse effects as well as the optimal stimulation site remain largely unknown. Among other techniques, computational simulations of the distribution of electric entities have been used to analyze long term chronic stimulation results in relation to the anatomy surrounding the stimulating contact. To our knowledge such methods have never been applied to study clinical results obtained during intraoperative stimulation tests. Therapeutic effects of stimulation are in general visually evaluated based on subjective clinical rating scales which are known for their inter- and intra-rater variability. While very few research groups have attempted intraoperative quantitative tremor evaluation, no research group has used computational simulations of the distribution of electrical entities during such stimulation tests. This study presents a method to correlate simulations of the electric field distribution during intraoperative stimulation tests with quantitatively evaluated symptom improvement and patient specific anatomy to get more informationregarding mechanisms of action and in turn optimize DBS target selection.Method: During DBS surgery of 3 essential tremor patient at the University Hospital in Clermont-Ferrand, France, a previously developed accelerometer based quantitative tremor evaluation technique was used [1]. The ventro-intermediate nucleus (VIM) and its anatomic neighbors were manually outlined based on spontaneous MRI contrasts and using a high field (4.7 Tesla) atlas. For each paitient, two parallel trajectories were planned per hemisphere with 7–8 stimulation test positions per trajectory spanning the region of interest. During the intraoperative stimulation tests, accelerometer data were recorded in sync with the stimulation current amplitude. Tremor improvement was postoperatively quantified compared to baseline tremor. For two stimulation amplitudes (low and high improvement) per position the effect of intraoperative stimulation tests in relation to the patient’s anatomy was studied along with the Department of Biomedical Engineering at Linkoping University. A computational model of the intraoperatively used exploration electrode was developed to simulate electric-field isosurface (0.2 V/mm) in the brain for the previously identified stimulation current amplitudes at the different test positions. Due to the large number of simulations, each voxel in theregion of interest may be part of several isosurfaces-each surface depicting one amplitude responsible for one improvement in tremor. To simplify visualization and interpretation, a maximum improvement map was generated, where each voxel was assigned to the isosurface representing the maximum improvement. Anatomical images, delineated structures, trajectories and improvement maps were visualized together in Paraview (Vtk based visualization software). The resulting visualization was evaluated by clinicians.Results: The software allowed 3D visualization as well as orthographic slices parallel to the trajectory. Clinicians confirmed that it enables the identification of the most effective stimulation areas with respect to the anatomy. A visual analysis of the improvement map for all the patients indicate that the highest improvement in tremor is observed when the region inferior, posterior and medial to the VIM is stimulated, i.e. the region where prelemniscal radiations merge with the VIM.Discussion and Conclusion: The proposed concept based on quantitative tremor evaluation, electric field simulations and patient specific anatomical data proposed provides a unique way to visualize a multitude of information in an interactive and adaptable way. The application of the method to 3 patients shows that the region where prelemniscal radiations merge with the VIM may be optimal for reducing tremor. This is also observed by other researchers. By applying this new method on more patients, the analysis of a high amount of intraoperative data might help to elucidate the mechanism of action of DBS