Improved magnetic resonance imaging (MRI) techniques, higher field strength, quality enhancing postprocessing and automatic segmentation approaches, the visualization of relevant target structures such as the subthalamic nucleus (STN) in deep brain stimulation (DBS), has become a valuable tool during preoperative planning to optimize lead positioning on an individual level, or to enhance postoperative programming by visualizing target structures in relation to leads and volume of activated tissue (VTA). Traditionally performed monomodal, multimodal automatic segmentation approaches allow for the integration of multiple MRI submodalities that are considered for atlas mapping and refinement. However, its effect on the visualization of the target structure needs to be investigated.

Ten patients diagnosed with Parkinson"s Disease (PD) who underwent STN-DBS were included. Preoperative MRI data acquired at a 3T MRI (Trio, Siemens, Erlangen, Germany) included a T1-weighted (T1w), T2-weighted (T2w), fluid-attenuated inversion recovery (FLAIR) and a susceptibility weighted (SWI) data set. In addition, intraoperatively, CT data sets were acquired after microelectrode placement (tip at target level) in clinical routine. For analysis regarding the effect of multimodality, for each patient eight MRI subsets were created (only T1w, T1w in combination with one further, two further or all three other data sets). For all eighty data subsets, rigid Image Fusion (Brainlab, Munich, Germany) was performed with the T1w as root data set, followed by automatic segmentation using the Anatomical Mapping Element (Brainlab, Munich, Germany). Within the CT data the microelectrodes were localized manually using the Trajectory Element (Brainlab, Munich, Germany) and entry and exit along the segmented STN was evaluated.

In total, 160 STN segmentations (80 left, 80 right) in relation to 25 detected trajectories were analyzed. In case of six STN outlines, the trajectory did not intersect with the segmentation. Overall the variability regarding the STN entry along the trajectory was 1.40 ± 0.74 mm, regarding the STN exit 1.43 ± 0.63 mm.

The spatial location of automatically generated STN segmentations varies depending on the included imaging data, and thereof can affect preoperative planning of optimal DBS leads location, intraoperative visualization in relation to electrophysiological measurements and postoperative image guided programming.