Article
Accuracy of stereotactic robot-assisted surgery depends on preoperative imaging modality
Abhängigkeit der Präzision der roboterassistierten, stereotaktischen Chirurgie von der präoperativen Bildgebungsmodalität
Search Medline for
Authors
Published: | June 26, 2020 |
---|
Outline
Text
Objective: The accuracy of stereotactic robot-assisted surgery is still under debate. As every neuronavigation system, the accuracy of the robot-assisted technology depends on the registration process. Modern stereotactic robots allow several different registration methods: i) laser-based surfaces registration (LSR) and ii) (bone) fiducial-based registration (BFR). While LSR depends on the quality of surface representation of the preoperative imaging, BFR has been shown to be very accurate but invasive. The present study aimed to determine the impact of these registration methods on the accuracy of robot-assisted stereotactic procedures.
Methods: This prospective study enrolled 29 patients who underwent a robotic-assisted stereotactic procedure (e.g. SEEG implantation, intracranial biopsy). Every patient received a preoperative T1 MR image (resolution 1 mm, isovoxel) for trajectory planning. In 7 cases this image was used for intraoperative MRI-LSR. Another 7 case received an additional CT (resolution 1mm) for intraoperative CT-LSR. Finally, 15 patients received a preoperative CT (resolution 1mm) after bone fiducial implantation for intraoperative BFR. Surgical approach was performed with a 2.1 mm drill. Entry point error (EPE) was assessed by the position of the drill hole on the postoperative CT scan.
Results: The mean EPE was 1.0 ± 0.9 mm [range: 0-3.4]. There was a significant main effect of the registration process on the EPE (X²=6.61, p=0.037; Kruskal-Wallis). Follow-up test revealed no statistical difference between CT-LSR (0.9±0.8 [0-2.1] mm) and BFR (0.7±0.3 [0.2-1.1] mm) (X²=0.40, p=0.525; Kruskal-Wallis). In contrast, both CT-LSR and BFR had a smaller EPE than MRI-LSR (2.2±1.1 [0.8-3.4] mm) (X²=4.34, p=0.037 and X²=5.78, p=0.016; Kruskal-Wallis).
Conclusion: The present study proves that the accuracy of robot-assisted stereotactic procedures depends on the applied registration process. T1-based MRI-LSR led to a higher EPE than CT-LSR and BFR. If MRI-LSR is necessary, sequences (e.g. T2) with better surface depiction should be used.