Oren Solomon1, Tara Palnitkar1,2, Rémi Patriat1, Henry Braun1, Joshua Aman2, Michael C Park2,3, Guillermo Sapiro4, Jerrold Vitek2, and Noam Harel1,3
1Center for Magnetic Resonance Research, Department of Radiology, University of Minnesota, Minneapolis, MN, United States, 2Department of Neurology, University of Minnesota, Minneapolis, MN, United States, 3Department of Neurosurgery, University of Minnesota, Minneapolis, MN, United States, 4Department of Electrical and Computer Engineering, Department of Biomedical Engineering, Department of Computer Science, Department of Mathematics, Duke University, Durham, NC, United States
1Center for Magnetic Resonance Research, Department of Radiology, University of Minnesota, Minneapolis, MN, United States, 2Department of Neurology, University of Minnesota, Minneapolis, MN, United States, 3Department of Neurosurgery, University of Minnesota, Minneapolis, MN, United States, 4Department of Electrical and Computer Engineering, Department of Biomedical Engineering, Department of Computer Science, Department of Mathematics, Duke University, Durham, NC, United States
Deep-learning
based segmentation of subcortical structures from 7 Tesla MRI allows for
patient-specific, robust and accurate deep brain stimulation surgery planning
and post-operative lead location assessment.
Figure 3: (A) 3D reconstruction of a DBS electrode
placement with respect to the manually delineated GPe (green)/GPi (yellow) for
a specific PD patient. (B) 3D reconstruction of the same patient’s data showing
the DBS electrode with respect to GP-net’s segmentation of the GPe (orange)/GPi
(blue). Green arrow – anterior (A), blue arrow – superior (S) and red arrow –
right (R).
Figure 4: (A) 3D reconstruction
of a DBS electrode placement with respect to the manually delineated STN for a
specific PD patient. (B) 3D reconstruction of the same patient’s data showing
the DBS electrode with respect to STN-net’s segmentation of the STN. Green arrow
– anterior (A), blue arrow – superior (S) and red arrow – right (R).
