Refining Computer-Assisted Stereo-Electroencephalography (SEEG) Planning Using Spatial Prior Trajectories – A Retrospective Validation
Abstract number :
1.318
Submission category :
9. Surgery / 9A. Adult
Year :
2021
Submission ID :
1826159
Source :
www.aesnet.org
Presentation date :
12/4/2021 12:00:00 PM
Published date :
Nov 22, 2021, 06:52 AM
Authors :
Debayan Dasgupta, BA MB BS MA MRCS (Eng.) - UCL Queen Square Institute of Neurology and National Hospital for Neurology and Neurosurgery; Cameron Elliott, MD, PhD, FRCSC - Senior Clinical Fellow, Victor Horsley Department of Neurosurgery, National Hospital for Neurology and Neurosurgery, London, UK; Rachel Sparks, PhD - School of Biomedical Engineering and Imaging Sciences, King’s College London, London, UK; Roman Rodionov - National Hospital for Neurology and Neurosurgery, London, UK; Aidan O'Keeffe - Dept of Statistical Science - University College London, UK; Andrew McEvoy - Consultant Neurosurgeon, Victor Horsley Department of Neurosurgery, National Hospital for Neurology and Neurosurgery, London, UK; Anna Miserocchi - Consultant Neurosurgeon, Victor Horsley Department of Neurosurgery, National Hospital for Neurology and Neurosurgery, London, UK; Sebastian Ourselin - School of Biomedical Engineering and Imaging Sciences, King's College London, UK; John Duncan - Professor of Neurology, Department of Clinical and Experimental Epilepsy, National Hospital for Neurology and Neurosurgery, London, UK
Rationale: Computer-assisted planning (CAP) gives faster SEEG planning and improved safety and efficiency metrics using EpiNav planning software, particularly with the use of accurate vascular models1,2. Preliminary work suggests that reference to prior SEEG trajectories may enhance CAP planning3.
Methods: A SEEG priors library was generated using 763 trajectories (n = 98) to define 31 targets and 51 entry zones from consecutive implantations, 2015-19. The potential benefit of refining CAP by the addition of SEEG priors (“CAP+Priors”) was compared to CAP alone by comparing planning time, expert rater implantability rates and resultant safety metrics in a separate dataset of 11 adult SEEG cases at NHNN (159 trajectories; 2016 – 2020). Vascular models for CAP were generated using digital subtraction angiography. Each case was planned by 2 fellows in a randomized order across the 22 plans guided by clinical implantation plans recording time for initial computer-generated plan output (T1), user-edited final plan (T2), time spent on each individual electrode, and what proportion of electrodes required manual planning (having to significantly alter entry and target points from the CAP/CAP+Prior output) vs minor adjustments following CAP output. The clinical feasibility of each planning type was assessed by blinded review of each trajectory by 2 epilepsy neurosurgeons.
Results: Qualitatively, priors allowed easer placement of multiple trajectories spaced through large gyri, particularly in the superior frontal gyrus and the superior parietal lobule, and aided in visualization of alternative trajectories where manual planning was required (particularly in plans with multiple electrodes in close proximity).
Median (IQR) T1 for CAP alone was 276s (51), vs CAP+Priors was 379s (154) (p=.03), a negligible difference in clinical planning time, and there was no difference in T2: CAP median (IQR) was 105 minutes (22), CAP+Priors was 96 minutes (68) (p=.92).
Proportion of trajectories requiring fully manual planning in CAP alone was 44.8%, CAP+Priors 38.1% (p=.30). Preliminary analysis reveals that in certain subregions CAP+Priors may decrease the proportion of computer-generated plans that require fully manual planning.
Conclusions: This retrospective validation of a comprehensive prior SEEG trajectory library demonstrates that this approach adds to the armory of CAP-SEEG planning, and allows for more granularity of trajectory planning – particularly in large gyri and hippocampus, and adds a method of including center-specific experience. This technique also allows easier standardization of planning, and allows for the future incorporation of experience and expertise from multiple expert centers, all without significantly increasing planning time, and potentially decreasing the proportion of trajectories that require manual planning.
Funding: Please list any funding that was received in support of this abstract.: Wellcome Trust 218380/Z/19/Z.
References:
1. Nowell M et al. J Neurosurg. 2016:124:1820-8
2. Vakharia VN et al. J Neurosurg. 2018:1-10
3. Vakharia VN et al. Front Neurol. 2020:11:706
Surgery