Learning curve analysis of 3D-fluoroscopy image-guided pedicle screw insertions in lumbar single-level fusion procedures.

Abstract

The implementation of 3D-navigation in the operating theater is reported to be complex, time consuming, and radiation intense. This prospective single-center cohort study was performed to objectify these assumptions by determining navigation-related learning curves in lumbar single-level posterior fusion procedures using 3D-fluoroscopy for real-time image-guided pedicle screw (PS) insertions. From August 2011 through July 2016, a total of 320 navigated PSs were inserted during 80 lumbar single-level posterior fusion procedures by a single surgeon without any prior experience in image-guided surgery. PS misplacements, navigation-related pre- and intraoperative time demand, and procedural 3D-radiation dose (dose-length-product, DLP) were prospectively recorded and congregated in 16 subgroups of five consecutive procedures to evaluate improving PS insertion accuracy, decreasing navigation-related time demand, and reduction of 3D-radiation dose. After PS insertion and intraoperative O-arm control scanning, 11 PS modifications were performed sporadically without showing "learning curve dependencies" (PS insertion accuracies in subgroups 96.6 ± 6.3%). Average navigation-related pre-surgical time from patient positioning on the operating table to skin incision decreased from 61 ± 6min (subgroup 1) to 28 ± 2min (subgroup 16, p < 0.00001). Average 3D-radiation dose per surgery declined from 919 ± 225mGycm (subgroup 1) to 66 ± 4mGycm (subgroup 16, p < 0.0001). In newly inaugurated O-arm based image-guidance, lumbar PS insertions can be performed at constantly high accuracy, even without prior experience in navigated techniques. Navigation-related time demand decreases considerably due to accelerating workflow preceding skin incision. Procedural 3D-radiation dose is reducible to a fraction (13.2%) of a lumbar diagnostic non-contrast-enhanced computed tomography scan's radiation dose.