4721
Segmentation and probabilistic tractography of GPi, GPe, STN and RN using Lead-DBS and FSL1Department of Health Sciences and Technology, GAIHST, Gachon University, Incheon, Korea, Republic of, 2Gachon University, Incheon, Korea, Republic of
Synopsis
Lead-DBS toolbox is used to segment globus pallidus internal, globus pallidus external, subthalamic nucleus and red nucleus, all of which are structures not automatically segmented by popular toolboxes such as FSL and Freesurfer. In addition, FSL's diffusion toolbox was used to generate probabilistic tractography between each segmented structure as well as compare the level of connectivity between each segmented structure.
Introduction
Deep brain structures have shown to be effective targets of deep brain stimulation (DBS) in treatment for Parkinson's Disease1. Therefore, segmentation of deep brain structures is important in observing the effects of Parkinson's, as well as providing accurate location of deep brain stimulation targets. However, common automatic segmentation toolboxes such as the FMRIB Software Library (FSL) and Freesurfer are unable to segment certain deep brain structures involved in Parkinson's disease and treatment. In this study, we utilize Lead-DBS toolbox2 to segment globus pallidus internal (GPi), globus pallidus external (GPe), subthalamic nucleus (STN) and red nucleus (RN). In addition, we use PROBTRACKX from the FMRIB software library (FSL) diffusion toolkit to generate and compare probabilistic tractography between each segmented structure3.Methods
For Lead-DBS structural segmentation, we used 3T MRI 3D-T1 weighted (TR = 2400 ms, TE = 2.14 ms, Flip Angle = 8 deg, Voxel Size = 0.7 mm isotropic, BW = 210 Hz/Px, Acquisition Time = 7 min 40 sec), 3D-T2 weighted (TR = 3200 ms, TE = 565 ms, Voxel Size = 0.7 mm isotropic, BW = 744 Hz/Px, Acquisition Time = 8 min 24 sec) and diffusion MRI images (TR = 5520 ms, TE = 89.5 ms, Flip Angle = 78 deg, Slice Thickness = 1.25 mm, 111 slices, Multiband = 3, Echo Spacing = 0.78 ms, BW = 1488 Hz/Px, b-values = 1000, 2000, and 3000 s/mm2) of 4 females with the age range 26~35, mean age 30.5 years (std = 3.2 years), obtained from Human Connectome Project (HCP), 'WU-Minn HCP Data – 1200 subjects'4. All T1, T2 and diffusion data were preprocessed using the HCP preprocessing pipeline5. Lead-DBS segmentation was done using MAGeT Brain-like segmentation approach6, which began with the segmentation of STN, GPi, GPe, RN in each subject using ANTs7. Subsequently, each warps and segmentations generated from previous segmentation were aggregated to provide additional information on the anatomical variance of segmented subjects, which were used to re-segment each subject. The segmented structures were overlaid on top of T2 images, where intensities of segmentation targets are visible, to check for accuracy as shown in Figure 1A. Each segmented structure was registered to its respective subject's diffusion MRI data, which were already preprocessed using eddy correction8 and BEDPOSTX9. Then probabilistic tractography of each registered structure and between registered structures were generated through PROBTRACKX10. The number of fiber tracks generated by each subject were averaged, and percentages were generated by dividing the totals of fibers that stemmed from each seed structure.Results
Figure 1B shows the 4 regions segmented by Lead-DBS. Each segmented structure that were overlaid on top of T2 images for validation matched the intensities of segmentation targets, confirming that segmentation was mainly successful. Table 1 shows the mean volumes of the segmented structures (A) and the results of waypoint connectivity mapping between registered structures (B and C). We found that most of the fibers that were generated were between GPe and GPi, with the least amount being generated between GPi and RN. The number of fibers generated between left and left, left and right, right and left, right and right segmented structures were used to generate a connectivity matrix as shown in Figure 2. 3D and 2D representations of fibers generated by PROBTRACKX are shown in Figure 3 and Figure 4.Discussion
Using Lead-DBS, we were able to successfully segment globus pallidus internal, globus pallidus external, subthalamic nucleus and red nucleus, and validate each segmentation through visual inspection. As such, we predict that there will be an improved consistency when preparing DBS treatment or comparing results of DBS treatment on the segmented structures by using Lead-DBS to segment structures. Previous studies have shown that deep brain stimulation of subthalamic nucleus had an increase in connectivity strength of surrounding white matter tracts11. Other studies have shown that direct deep brain stimulation of the tractography linking the basal ganglia have shown improvements in treatment of Parkinson's disease12. Therefore, we believe that the use of Lead-DBS to generate masks needed to study the effects of DBS on tractography that pass through target structures should improve the consistency of results.Conclusion
Our study shows that Lead-DBS was able to accurately segment the four structures in each one of our HCP patients. In addition, we showed that we can successfully generate probabilistic tractography between each segmented structure using FSL's PROBTRACKX. Results show that there is potential in using Lead-DBS to study effects of Parkinson's and Parkinson's treatment on GPe, GPi, RN and STN, as well as the white matter tracts that pass through them.Acknowledgements
This research was supported by Brain Research Program through the National Research Foundation of Korea (NRF) funded by the Ministry of Science and ICT (NRF-2017M3C7A1044367).References
1. Krause M, Foge, W, Heck A, et al. Deep brain stimulation for the treatment of Parkinson's disease: subthalamic nucleus versus globus pallidus internus. Journal of Neurology, Neurosurgery & Psychiatry 2001;70(4): 464-470.
2. Horn A and Kühn AA. Lead-DBS: a toolbox for deep brain stimulation electrode localizations and visualizations. Neuroimage 2015;107:27-135.
3. Behrens TE, Berg HJ, Jbabdi S, et al. Probabilistic diffusion tractography with multiple fibre orientations: What can we gain?. Neuroimage 2007;34(1):144-155.
4. Van Essen DC, Smith SM, Barch DM, et al. The WU-Minn human connectome project: an overview. Neuroimage 2013;80:62-79.
5. Glasser MF, Sotiropoulos SN, Wilson JA, et al. The minimal preprocessing pipelines for the Human Connectome Project. Neuroimage 2013;80:105-124.
6. Chakravarty MM, Steadman P, van Eede MC, et al. Performing label‐fusion‐based segmentation using multiple automatically generated templates. Human brain mapping 2013;34(10):2635-2654.
7. Avants BB, Tustison N, Song G. Advanced normalization tools (ANTS). Insight j 2009;2:1-35.
8. Andersson JL, Sotiropoulos SN. An integrated approach to correction for off-resonance effects and subject movement in diffusion MR imaging. Neuroimage 2016;125:1063-1078.
9. Jbabdi S, Sotiropoulos SN, Savio AM, et al. Model‐based analysis of multishell diffusion MR data for tractography: How to get over fitting problems. Magnetic Resonance in Medicine. 2012;68(6):1846-1855.
10. Behrens TE, Woolrich MW, Jenkinson M, et al. Characterization and propagation of uncertainty in diffusion‐weighted MR imaging. Magnetic Resonance in Medicine: An Official Journal of the International Society for Magnetic Resonance in Medicine. 2003;50(5):1077-1088.
11. Vanegas-Arroyave N, Lauro PM, Huang L, et al. Tractography patterns of subthalamic nucleus deep brain stimulation. Brain 2016;139(4):1200-1210.
12. Sweet JA, Walter BL, Gunalan K, et al. Fiber tractography of the axonal pathways linking the basal ganglia and cerebellum in Parkinson disease: implications for targeting in deep brain stimulation. Journal of neurosurgery 2014;120(4):988-996.
Figures
