Feasibility and Accuracy of Robotic-Assisted Navigation for Thoracic Pedicle Screw Placement Using CT-Like 3D-MRI.

Abstract

Cadaveric study. To assess the feasibility and accuracy of robotic positioning of thoracic pedicle screws based on magnetic resonance imaging (MRI). Robotic-assisted navigation (RAN) has demonstrated enhanced precision in thoracic pedicle screw placement. Currently, preoperative computer tomography (CT) scanning is required, exposing patients to radiation-a particular concern for younger patients undergoing multilevel fusion surgery. CT-like three-dimensional (3D)-MRI sequences may offer a radiation-free alternative, but their application in thoracic pedicle screw placement remains unexplored. CT-like 3D-MRI scans of the thoracic spine were obtained in two human cadaveric specimens. A RAN system was employed to plan and guide 48 pedicle screws (bilateral screws in thoracic vertebrae T1-T12) using these MRI scans. Following the placement of the pedicle screws, post-procedure CT scans were obtained to evaluate the accuracy of screw positioning. Accuracy was assessed by comparing the actual placement to the pre-procedure plan (difference in millimetres) and via the Gertzbein-Robbins scale (GRS). A total of 48 thoracic pedicle screws were inserted robotically in two human specimens (T1-T12 bilaterally). Post-procedure CT scan evaluations revealed that all screws achieved an acceptable grade on the GRS (A or B). Specifically, 87.5% of the screws were classified as grade A and 12.5% were classified as grade B. The median deviations from the planned trajectory were 0.4 mm in the axial (IQR: 0.0; 1.2 mm) and 0.05 mm in the sagittal (IQR: -0.3; 0.3 mm) planes. This cadaveric study demonstrates that MRI-based RAN can accurately guide thoracic pedicle screw placement. The findings suggest MRI-based RAN could provide a radiation-free alternative for thoracic spine instrumentation, particularly beneficial for pediatric and adolescent patients.