Objectives Automated planning of stereoelectroencephalography (SEEG) electrode trajectories is dependent on vessel segmentation.¹ We quantify imaging protocols ability to measure trajectory-to-vessel distance.

Design Retrospective analysis.

Subjects Ten consecutive patients were selected whom had SEEG implantation (95 electrodes) and Digital Catheter Subtraction Angiography (DSA) with catheterization of carotid or vertebral arteries, post-gadolinium T1-weighted (GAD), phase-contrast MR angiography and MR venography (MR) acquired.

Methods SEEG trajectories were planned manually with DSA. Minimum distance to vessels and risk¹ were computed for each trajectory using vessel segmentation from GAD, MR, or DSA. Vessel size was considered by including DSA vessels diameters above 1, 2, 3, or 4 mm.

Results Minimum distance to a vessel was 6.2±3.9 mm (GAD), 2.5±1.6 mm (MR), and 1.5±1.2 mm (DSA). Based on DSA vessel size minimum distances were 2.0±1.5 mm (DSA >1 mm), 3.4±2.6 (DSA >2 mm), 6.6±4.6 mm (DSA >3 mm), and 11.8±7.9 mm (DSA >4 mm). Risk was 0.4±0.4 (GAD), 0.8±0.4 (MR), and 1.1±0.2 (all DSA), 1.0±0.2 (DSA >1 mm), 0.7±0.4 (DSA >2 mm), 0.4±0.5 (DSA >3 mm), and 0.2±0.3 (DSA >4 mm).

Conclusions DSA is best able to segment vessels. MR has metrics similar to DSA vessels above 2 mm. GAD has metrics similar to DSA vessels above 3 mm.