Mitigating radiation in robotic-assisted sacropelvic screw placement using 3D magnetic resonance imaging

To evaluate the feasibility and accuracy of robotic-assisted (RA) sacropelvic screw placement (SPSP) using three-dimensional magnetic resonance imaging (3D MRI) for preoperative planning and robotic registration. SPSP is a critical component of spinal fixation but is subject to trajectory deviations, and radiation exposure with fluoroscopic or computed tomography (CT)-based navigation. CT-like 3D MRI-based navigation provides radiation-free preoperative imaging, but its integration with RA systems remains unexplored in SPSP.

Eight human cadavers underwent RA SPSP using MRI-based preoperative planning, including the placement of S1 pedicle screws, S1-alar-iliac (AI) screws, and S2-AI screws bilaterally. Accuracy was measured as deviation between planned vs. achieved trajectories by postoperative CT analysis. This was quantified through direct measurements and subsequently categorized using the Gertzbein-Robbins Scale (GRS).

A total of 48 screws were placed using 3D MRI-based RA navigation. Postoperative CT confirmed that 46 screws (95.8%) were graded as GRS Grade A. Among the S1 pedicle screws (n = 16), the median deviation from the planned trajectory was 0.5 mm (IQR 0.3-0.8) in the axial plane and 0.8 mm (IQR 0.5-0.9) in the sagittal plane. For S1AI and S2AI screws (n = 30), the median deviation at the entry point was 0.7 mm (IQR 0.3-1.2), while the median axial deviation from the medial SI joint border measured 0.9 mm (IQR 0.6-1.7).

This first cadaveric validation of 3D MRI-guided RA sacropelvic fixation demonstrates the feasibility of radiation-free preoperative imaging and robotic registration for precise trajectory execution.