Abstracts

Rationale: Digital subtraction angiography (DSA) is a historically accepted gold standard to show intracranial vasculature when planning safe SEEG implantation. Current methods of SEEG implantation offer submillimetric accuracy which contributes to safer surgery. The risks of morbidity from angiography and SEEG electrode implantation are of a similar order. This explains the interest to find an alternative to DSA for planning SEEG trajectories. We quantified the safety of planned trajectories, comparing plans made using post-contrast T1, optimised MR-venography and their combination, with DSA. Methods: Post-Gadolinium T1 MRI, optimised 3D Phase contrast MRI (MRV) and DSA were acquired as part of the imaging protocol for SEEG implantation. Three patients were randomly selected from a cohort of 10 cases in whom all vascular imaging was acquired. The vascular image processing was performed using EpiNavTM software (UCL, London, UK): applying Sato filter, masking out extra-cranial information. The Computer Assisted Planning (CAP) algorithm was used to plan trajectories according to clinical need to sample specific brain areas, with a total of 25 trajectories. Three sets of trajectories were generated - to avoid blood vessels segmented from (a) postGAD T1, (b) combination of postGAD T1 and MRV, (c) DSA. The parameters characterising trajectories (length, angle to skull, Cumulative Risk along the trajectory (CR), Minimal Distance to a blood vessel along the course of the trajectory (MD)) were calculated for each set of trajectories. CR and MD were calculated for the trajectory sets (a) and (b) when using vascular model derived from DSA. Results: There was a significant increase of MD metric when postGAD or its combination with MRV were used, as opposed to DSA. The CR metric was statistically significantly higher when using vascular information from DSA. This is due to significantly richer vasculature visualisation on DSA. The parameter “angle to skull” was higher when using vascular 3D model from DSA, as the rich vasculature on DSA left fewer options for safe trajectories which were easy to implement. The trajectories from sets (a,b) hit the blood vessels of the 3D model built using DSA: case 1 – 3/9 trajectories from set (a), 3/9 trajectories from set (b); case 2 – 4/8 from (a), 5/8 from (b); case 3 –4/8 from (a), 2/8 from (b). The reasons for these trajectories to appear unsafe on DSA are multiple and range from significant inter-patient variability of quality of vascular MR images to inability of segmentation algorithms to segment vascular information from vascular MRI. Conclusions: We set up a framework for quantitative description of implanted SEEG electrodes and demonstrated the use of this framework for quantifying and comparing different vascular imaging modalities for planning safe and accurate SEEG implantation. At present, DSA appears necessary for optimal display of vasculature when using CAP. Funding: The Health Innovation Challenge Fund (HICF-T4-275, WT097914, WT106882) which is a parallel funding partnership between the Wellcome Trust and the Department of Health.