gms | German Medical Science

Objective: This prospective cohort study evaluated the accuracy of pedicle screw insertion for dorsal instrumentation of the thoracolumbar spine comparing the freehand (FH) to the 3D-fluoroscopy-image guided navigation (3D-Fl) technique.

Method: From Jan 2011 to Jan 2012 313 consecutive patients received dorsal instrumentation with pedicle screw (PS) fixation mainly for degenerative, but also for metastatic, infectious or osteoporotic spine disease. 107 pats. (58m, 49f) underwent surgery in the FH technique (621 PS; 166 thoracic, 455 lumbar) and 206 pats. (98m, 108f) received surgery with the use of 3D-fluoroscopy-image guided navigation (BrainLAB) (1252 PS; 212 thoracic, 1040 lumbar). Accuracy of pedicle screws in postoperative CT-scans, intraoperative screw revisions and revision rates due to misplaced screws were analyzed. Accuracy of pedicle screws was classified into three categories by an independent blinded neuroradiologist: category I = optimal, PS completely within the pedicle, category II = suboptimal, minor pedicle breach, PS biomechanically acceptable/clinically asymptomatic and category III = revision needed, major pedicle breach, PS biomechanically unacceptable/clinically symptomatic.

Results: Overall accuracy of PS (category I+II) was 97.7% in the FH group (96.9% / 441 lumbar; 100% / 166 thoracic PS) and 98.6% in the 3D-Fl group (99.7% / 1037 lumbar; 92.9% / 198 thoracic PS).The number of intraoperative screw revisions was significantly higher (p<0.05) in the 3D-Fl group with 66 PS (17 thoracic, 49 lumbar; total 5.3%) in 37 pats compared to the FH group with 3 PS in 3 pats. (1 thoracic, 2 lumbar; 0.5%). Revision surgeries due to misplaced screws was significantly higher (p<0.05) in the FH group with 10 pats. (1 thoracic, 17 lumbar PS; total 2.9%) compared 5 pats. in the 3D-Fl group (4 thoracic, 2 lumbar PS; total 0.5%).

Conclusions: Overall PS accuracy was very high and comparable in both groups. There were significantly more intraoperative screw revisions in the 3D-Fl group compared to the FH group and significantly more revision surgeries in the FH group compared to the 3D-Fl group. Partly these findings may be attributed to the used techniques: in contrast to the freehand cases all navigated cases underwent intraoperative 3D-scans after PS insertion with the possibility of instant screw revision. Interestingly in our series lumbar PS were more accurately placed with navigation, whereas thoracic PS were more accurate in the freehand technique.