Robot-Assisted Minimally Invasive Asleep Single-Stage Deep Brain Stimulation Surgery: Operative Technique and Systematic Review

Abstract

BACKGROUND AND OBJECTIVES: Robotic assistance has garnered increased use in neurosurgery. Recently, this has expanded to include deep brain stimulation (DBS). Several studies have reported increased accuracy and improved efficiency with robotic assistance, but these are limited to individual robotic platforms with smaller sample sizes or are broader studies on robotics not specific to DBS. Our objectives are to report our technique for robot-assisted, minimally invasive, asleep, single-stage DBS surgery and to perform a meta-analysis comparing techniques from previous studies. METHODS: We performed a single-center retrospective review of DBS procedures using a floor-mounted robot with a frameless transient fiducial array registration. We compiled accuracy data (radial entry error, radial target error, and 3-dimensional target error) and efficiency data (operative time, setup time, and total procedure time). We then performed a meta-analysis of previous studies and compared these metrics. RESULTS: We analyzed 315 electrodes implanted in 160 patients. The mean radial target error was 0.9 ± 0.5 mm, mean target 3-dimensional error was 1.3 ± 0.7 mm, and mean radial entry error was 1.1 ± 0.8 mm. The mean procedure time (including pulse generator placement) was 182.4 ± 47.8 minutes, and the mean setup time was 132.9 ± 32.0 minutes. The overall complication rate was 8.8% (2.5% hemorrhagic/ischemic, 2.5% infectious, and 0.6% revision). Our meta-analysis showed increased accuracy with floor-mounted over skull-mounted robotic platforms and with fiducial-based registrations over optical registrations. CONCLUSION: Our technique for robot-assisted, minimally invasive, asleep, single-stage DBS surgery is safe, accurate, and efficient. Our data, combined with a meta-analysis of previous studies, demonstrate that robotic assistance can provide similar or increased accuracy and improved efficiency compared with traditional frame-based techniques. Our analysis also suggests that floor-mounted robots and fiducial-based registration methods may be more accurate.