Accurate placement of thoracolumbar pedicle screws using a handheld iOS-based navigation device: a prospective intra-patient agreement study.

Abstract

Accurate pedicle screw placement is a challenge with reported misplacement rates of 10% and higher. A handheld navigation device (HND) may provide accuracy equal to CT-based navigation (CT-Nav) but without the cost and complexity. To study the accuracy of a handheld navigation device for pedicle screw placement. This prospective cross-sectional study with consistently applied reference standard enrolled 20 patients undergoing 92 pedicle screw placements. Patients who underwent pedicle screw placement between May 2022 and September 2022. Pedicle screw placement accuracy per Gertzbein-Robbins. Once the screw pilot hole was established, the proposed trajectory of the HND was compared with that proposed by CT-Nav. Postoperatively, screw accuracy was graded according to Gertzbein-Robbins by a blinded radiologist based on CT scans. Accuracy was compared between the two systems and published control for fluoroscopy assisted and CT-Nav placement using Bayesian posterior distribution. The trajectory proposed by the HND and CT-Nav were in agreement in 98.9% (95% Exact CI; 94.09%-99.97%). The HND accuracy was 98.9% with 91 screws rated "A" and 1 rated "C". Noninferiority to fluoroscopic placement was achieved because the one-sided normal-approximation 95% CI Lower Bound (LB) of 95.3% is greater than the Performance Goal (PG) of 83.4%. Posthoc analysis demonstrated that the probability of superiority of the HND relative to the historical accuracy rate of 91.5% for fluoroscopy assisted procedures is >0.999 and that the HND's accuracy rate is within 4.5% of CT-Nav of 95.5% is >0.999. No adverse events or intra-operative complications associated with HND were observed. There was 1 (1.1%) intra-operative repositioning and no reoperations for any reason. The accuracy rate of the HND was 98.9%, and the proposed trajectory matched with CT-Nav in 98.9% of the time. This is superior to the historical published accuracy rate for fluoroscopy-assisted procedures and equivalent to the historical published accuracy rate for CT-Nav. Dutch trial register NL74268.058.20.