Comparison of three imaging and navigation systems regarding accuracy of pedicle screw placement in a sawbone model

Abstract

3D-navigated pedicle screw placement is increasingly performed as the accuracy has been shown to be considerably higher compared to fluoroscopy-guidance. While different imaging and navigation devices can be used, there are few studies comparing these under similar conditions. Thus, the objective of this study was to compare the accuracy of two combinations most used in the literature for spinal navigation and a recently approved combination of imaging device and navigation system. With each combination of imaging system and navigation interface, 160 navigated screws were placed percutaneously in spine levels T11-S1 in ten artificial spine models. 470 screws were included in the final evaluation. Two blinded observers classified screw placement according to the Gertzbein Robbins grading system. Grades A and B were considered acceptable and Grades C-E unacceptable. Weighted kappa was used to calculate reliability between the observers. Mean accuracy was 94.9% (149/157) for iCT/Curve, 97.5% (154/158) for C-arm CBCT/Pulse and 89.0% for CBCT/StealthStation (138/155). The differences between the different combinations were not statistically significant except for the comparison of C-arm CBCT/Pulse and CBCT/StealthStation (p = 0.003). Relevant perforations of the medial pedicle wall were only seen in the CBCT group. Weighted interrater reliability was found to be 0.896 for iCT, 0.424 for C-arm CBCT and 0.709 for CBCT. Under quasi-identical conditions, higher screw accuracy was achieved with the combinations iCT/Curve and C-arm CBCT/Pulse compared with CBCT/StealthStation. However, the exact reasons for the difference in accuracy remain unclear. Weighted interrater reliability for Gertzbein Robbins grading was moderate for C-arm CBCT, substantial for CBCT and almost perfect for iCT.