Clinically applicable delineation of the pallidal sensorimotor region in patients with advanced Parkinson’s disease: study of probabilistic and deterministic tractography

Abstract

OBJECTIVEDeep brain stimulation (DBS) is an effective procedure in improving motor symptoms for patients with advanced Parkinson’s disease (PD) through the use of high-frequency stimulation. Although one of the most commonly used target sites for DBS, sensorimotor regions of the globus pallidus interna (GPi) have yet to be thoroughly described with advanced neuroimaging analysis in vivo for human subjects. Furthermore, many imaging studies to date have been performed in a research setting and bring into question the feasibility of their applications in a clinical setting, such as for surgical planning. This study compares two different tractography methods applied to clinically feasible acquisition sequences in identifying sensorimotor regions of the GPi and the subthalamic nucleus (STN) in patients with advanced PD selected to undergo DBS.METHODSSeven patients with refractory PD selected for DBS were examined by MRI. Diffusion images were acquired with an average acquisition time of 15 minutes. Probabilistic and deterministic tractography methods were applied to each diffusion-weighted data set using FSL and MRtrix, respectively. Fiber assignment was performed using combined sensorimotor areas as initiation seeds and the STN and GPi, separately, as inclusion masks. Corticospinal tracts were excluded by setting the cerebral peduncles as exclusion masks. Variability between proposed techniques was shown using center of gravity (CoG) coordinates.RESULTSDeterministic and probabilistic corticopallidal and corticosubthalamic pathways were successfully reconstructed for all subjects across all target sites (bilaterally). Both techniques displayed large connections between the sensorimotor cortex with the posterolateral aspect of the ipsilateral GPi and the posterosuperolateral aspect of the ipsilateral STN. The average variability was 2.67 mm, with the probabilistic method identifying the CoG consistently more posterior and more lateral than the deterministic method.CONCLUSIONSSuccessful delineation of the sensorimotor regions in both the GPi and STN is achievable within a clinically reasonable timeframe. The techniques described in this paper may enhance presurgical planning with increased accuracy and improvement of patient outcomes in patients undergoing DBS. The variability found between tracking techniques warrants the use of the probabilistic tractography method over the deterministic method for presurgical planning. Probabilistic tractography was found to have an advantage over deterministic tractography in its sensitivity, in accurately describing previously described tracts, and in its ability to detect a larger number of fibers.

Associations Between White Matter Microstructures and Cognitive Functioning in 8-Year-Old Children: A Track-Weighted Imaging Study

“…The diffusion-weighted MRI data used in this study was acquired using a single shell with basic DWI acquisition. Multiple previous studies 23,[47][48][49][50][51] involving low b-values such as 800 and 1000 and with gradient directions in the range of 15 to 30 have used constrained spherical deconvolutionbased tractography and demonstrated consistent tractography results. For example, Tournier et al 26 discussed about the feasibility of using various acquisition protocols for constrained spherical deconvolution model, in which DWI acquired in a 1.5-T scanner with b-value of 1000 s/mm 2 and 20 directions was still shown to provide better results in resolving multiple fiber orientations.…”

Should the Globus Pallidus Targeting Be Refined in Dystonia?

“…Those fibers can be mapped to cortical areas, as a presurgical estimate of the cortico-pallidal circuits that will be stimulated during the therapy. 66 A combination of these techniques may allow the surgeon to refine the lead placement, aiming for tracts that are connected to the sensorimotor cortex. Besides the detectability of theta oscillations during surgery may help in targeting as described below.…”

Imaging Modalities Used for Frameless and Fiducial-Less Deep Brain Stimulation: A Single Centre Exploratory Study among Parkinson’s Disease Cases

“…Multi-compartment Ball-and-Stick model is then fitted on artefacts-corrected dMRI data using the 4edpost [ 11 ] function. Initially, STN is segmented using a standard anatomical atlas [ 13 ] complemented with manual edits [ 14 , 15 , 16 ] if necessary, exploiting T 2 -weighted and SWI images linearly co-registered with high-resolution T 1 -weighted image using the FLIRT tool [ 17 ]. Next, the neurosurgeon creates virtual manual plans of electrode placement for both patient groups (while blinded to group allocation).…”

“…In contrast, BA 25, 30, 34, and 35 serve as seed areas for the pathway reaching the limbic region of the STN [ 18 ]. Using the probtrackx [ 11 , 16 , 18 ] tool, probabilistic tractography is performed from cortical seed areas to STN to delineate the two STN subdivisions. Time-demanding estimation of the Ball-and-Sticks model and running of probabilistic tractography is parallelised using the HTCondor parallelisation tool over multiple CPUs.…”