Patient-to-robot registration: The fate of robot-assisted stereotaxy

Background Robot-assisted stereotaxy (RAS) promises higher stereotactic accuracy (SA) and time efficiency (TE) than frame-based stereotaxy. However, both aspects are attributed to the problem of patient-to-robot registration. Objective To examine different registration techniques regarding their SA and TE. Methods This study enrolled 57 patients undergoing RAS with bone fiducial registration (BFR) or laser surface registration (LSR). SA was measured by the entry point error (EPE). Additionally, predictors of SA (registration error [RegE], distance-to-registration plane [DTC]) and TE (imaging, skin-to-skin) were assessed. Results The mean SA was 1.0 +/- 0.8 mm. BFR increased SA by reducing RegE and DTC. In LSR, EPE depended on DTC (face and forehead) with highest accuracy for DTC <= 100 mm. CT-based LSR exerted a higher SA than MR-based LSR. In BFR, TE was confined by the additional imaging. Conclusion Every registration technique counteracts one of the promises of RAS. New solutions are needed to increase the acceptance of RAS in neurosurgery.