EP 77. Target identification in deep brain stimulation for Parkinson’s disease: The role of probabilistic tractography

Introduction Deep brain stimulation (DBS) is a widely accepted therapeutic choice in the management of Parkinson’s disease. Optimal location of the implantable DBS lead should be chosen according to intraoperative electrophysiology, and macrostimulation, while preoperative targeting is still heavily influenced by standard stereotactic coordinates. Previous electrophysiological and in vitro studies were able to identify functional subgroups within the subthalamic nucleus (STN), but noninvasive methods are yet to be developed to achieve the same distinction. Methods 7 patients suffering from Parkinson’s disease, who underwent bilateral STN DBS, were enrolled in this study. Preoperative 3D T2 (isotropic, 2 × 2 × 2 mm), 3D T1 (isotropic, 1 × 1 × 1 mm), and DTI (isotropic, 2 × 2 × 2 mm, 32 and 60 directions) sequences were acquired from each participant using a 3T Phillips Achieva and a 1.5 T GE Signa Excite scanner in two different centers. Postoperative CT scans were acquired at least 6 weeks after implantation to rule out undesired MR-CT registration artifacts. Given as the result of cortical and subcortical parcellation (Freesurfer 5.3), 8 target regions (limbic system (LS), dorsal prefrontal cortex (DPFC), premotor cortex (PM), primary motor cortex (M1), primary sensory cortex, parietal lobe, occipital lobe, temporal lobe) were selected for probabilistic tractography. Segmentation of the STN, by determining the probability of its connections to the aforementioned cortical regions, was carried out using FSL 5.0.9 (Oxford University, FMRIB Software Library). Postoperative CT images were registered to the preoperative T1 images using linear registration with 6 degrees of freedom (Flirt, FSL 5.0.9). DBS lead position was correlated with the results of probabilistic tractography, intraoperative microelectrode recordings and macrostimulation. Results Probabilistic tractography resulted in significant connections between the STN and the LS, the DPFC, the PM, and the M1. Higher intersubject variability was found in limbic connectivity. Remaining cortical targets were excluded from further analysis due to significantly lower probability. LS portion of the STN was situated in the ventral inferomedial region in each patient followed by portions connecting to the DPFC, PM, M1 respectively. Overlapping connections were found in the LS-DPFC, and in the PM-M1 regions. Intraoperative macrostimulation resulted in significant decrease of symptoms at levels where the DBS leads entered the PM-M1 portion of the nucleus in all patients. Intraoperative assessment of tremor and rigidity showed decreased symptoms on both sides (mean symptom reduction on the left side: 71.43%, right side: 67.86%). Mean distance between the most distal point of the first stimulation contact and the weighted average of connectivity probabilities in the PM-M1 portion of the STN are as follows: Right side: x : 2.2 mm y :1.2 mm z : 3.4 mm; Left side: x : 1.4 mm y : 0.3 mm z : 4 mm. Conclusion Probabilistic tractography can reveal distinct functional subgroups within the STN, resulting in that this method might be a suitable alternative to invasive intraoperative electrophysiology providing an optimal target for DBS lead placement. Levels of leads penetrating the PMM1 portion of the STN on postoperative scans were in accordance with levels of macrosimulation electrodes and decrease of clinical signs during intraoperative neurological examination. KTIA_NAP_13-1-2013-0001.