Implantationsgenauigkeit und -dauer der Stereoelektroenzephalographie in einem Schädelphantom: Vergleich eines modularen Robotiksystems mit einem traditionellen stereotaktischen Rahmen

In drug-resistant epilepsy patients, the epileptogenic zone can be identified safely and effectively through Stereoelectroencephalography (SEEG). Traditionally, stereotactic frames are used, but interest in robotic systems is growing to increase operational efficiency. Our study aimed to compare the accuracy and performance of a table-mounted robotic platform (Cirq®, Brainlab AG, Germany) to a Leksell Coordinate Frame G (Elekta, Sweden) on phantoms. Cirq is already used in spinal procedures and cranial biopsies and recently obtained the CE mark for SEEG.

Using publicly available MRIs (IXI database), 1 bilateral and 2 unilateral SEEG cases with 10 realistic trajectories each were planned in Brainlab Elements. 6 skull models (Sawbones, USA) were coated with epoxide gel to mimic bone layers and filled with 1.6% agar. For implantation, 2 sets of 10 SEEG electrodes and anchor bolts (Ad-Tech Medical, USA) were used. Each SEEG case was performed twice by P.R. in identical phantoms positioned laterally with a Leksell head ring. For Cirq, Automatic Image Registration (AIR, Brainlab) was performed using cone-beam computed tomography (CBCT, Loop-X Mobile Imaging Robot, Brainlab). CBCT was also used for stereotactic frame localization and for post-op imaging in both workflows. Accuracy metrics were calculated for all 60 electrodes as shown in Fig. 1. For both workflows implantation time per electrode was measured. Results between the two groups were compared using the Mann-Whitney U-Test.

In the robotic workflow, the mean radial entry error was 0.62 ± 0.40 mm, and in the frame workflow 0.61 ± 0.38 mm (p=0.982). For the target point, the mean Euclidean error was 1.60 ± 0.94 mm with the robot and 0.99 ± 0.36 mm with the frame (p<0.0001). The mean radial error was 1.03 ± 0.53 mm vs. 0.51 ± 0.30 mm (p<0.0001) and the absolute depth error 1.03 ± 1.02 mm vs. 0.75 ± 0.45 mm (p=0.266). With the robotic system, mean implantation time per electrode was shorter than with the frame, at 5.5 ± 1.3 min vs. 8.0 ± 1.9 min (p < 0.0001).

In our phantom setup, robot-assisted SEEG with automatic CBCT based image registration proved simpler and faster compared to the Leksell frame, while maintaining clinically acceptable accuracy. Its versatility across procedures suggests potential for wider clinical availability. Further research involving patients is warranted.