This retrospective single-center study evaluated the accuracy of open posterior thoracic, lumbar, and sacral pedicle screw placement in 180 consecutive patients operated on for different spinal diseases between January 2017 and December 2024 at our institution. Two experienced surgeons performed all surgeries. Ninety-two of the patients were males, and their mean age was 57.9 years (15–83 years). Posterior pedicle instrumentation was performed in 33 patients (18.3%) with primary or secondary neoplasms, 41 (22.8%) with traumatic fractures, 69 patients (38.3%) with degenerative spinal disease (stenosis and/or scoliosis), 21 patients (11.7%) with spondylolisthesis, 14 patients (7.8%) with inflammatory conditions, and 2 patients (1.1%) with parasitic diseases. These data are summarized in Table 1.

The inclusion criterion was the placement of navigated pedicle screws in the thoracic, lumbar, and sacral regions of the spine with the assistance of O-arm for different spinal pathologies. The exclusion criteria were a) placement of pedicle screws outside the mentioned spine regions; b) placement of pedicle screws without the O-arm navigation system; and c) patients who refused to sign the informed consent form. Data collected encompassed demographic information; preoperative and postoperative neurological status; the level and type of spinal pathology; the number of screws placed; clinical outcomes and complications; scores on the Modified McCormick Scale; and imaging data from the preoperative, intraoperative, and postoperative phases. The study was approved by the local Ethics Committee. Informed consent was obtained from all patients.

All patients underwent intraoperative O-arm imaging to assess screw placement accuracy. This included scans performed before and after screw placement and, if necessary, after intraoperative correction of malpositioned screws. The intraoperative assessment of accuracy was performed by a leading neurosurgeon and experienced radiologist using the Gertzbein-Robbins Grading Scale, the results of which are presented in Table 2.^[2] In addition, an early postoperative CT scan with a slice thickness of 1 mm was performed in all cases. The final accuracy of screw placement after the conventional CT scanning was also evaluated by an experienced radiologist and the leading neurosurgeon using the same scale. None of the patients needed a revision surgery due to screw misplacement.

For the purposes of this study, Grade A was considered a perfect position, and Grade B was considered a clinically acceptable position. For grades C, D, or E, immediate intraoperative repositioning was performed to achieve a clinically acceptable screw placement, which was confirmed by repeat intraoperative O-arm scanning.

All data were analyzed using IBM SPSS, version 23. Descriptive statistical methods were applied in order to assess the distribution of both qualitative and quantitative variables. For the purpose of making comparisons between two independent groups, the non-parametric Mann-Whitney U test was utilized; in the case of comparisons involving more than two groups, the Kruskal-Wallis test was employed. The associations between variables were evaluated using Spearman’s rank correlation coefficient (Spearman’s rho). The Wilson method was used to calculate the 95% confidence intervals (CI) for relative proportions. A p-value of <0.05 was considered to be statistically significant.

A total of 1286 navigated screws were implanted. Of these, only the 1257 pedicle screws placed in the thoracic, lumbar, and sacral regions of the spine were included in the analysis.

Intraoperative O-arm scans revealed 44 (3.5%) malpositioned screws out of a total of 1257. Of the 336 screws placed in the thoracic spine, 35 (10.4%) had pedicle breaches, with 16 screws (4.8%) graded as Grade B, 10 screws (3.0%) as Grade C, 5 screws (1.5%) as Grade D, and 4 screws (1.2%) as Grade E. Lateral pedicle breaches were observed in 30 cases (8.9%), while medial breaches were seen in 5 cases (1.5%).

There was a significant difference in the number of malpositioned pedicle screws between the upper and lower thoracic spine. Ninety screws were placed in the upper thoracic region (T1-T6), with 10 (11.1%) of them requiring intraoperative revision (grades C-E). At levels T7–T12, 246 screws were implanted, of which 9 (3.6%) were malpositioned.

A total of 748 screws were placed in the lumbar spine, and only 5 (0.7%) were found to be misplaced: 2 screws (0.3%) were Grade B, 1 screw (0.1%) was Grade C, and 2 screws (0.3%) were Grade E. All misplaced screws were found to be lateral in position.

In the sacral region (S1-S2), 173 screws were implanted, and 4 (2.3%) were found to be malpositioned: 3 screws (1.7%) were Grade B, and 1 screw (0.6%) was Grade D. The misplaced screws in this region were with ventral penetration.

As evident from Tables 3 and 4, 1213 pedicle screws (96.5%) had a perfect position (Grade A). On the other hand, 1234 screws (98.2%) in total had clinically acceptable positions (Grade A + Grade B) with no need for intraoperative revision. A total of 19 screws (1.8%) had an unacceptable position (Grade C, D, or E) and underwent intraoperative repositioning in the thoracic, lumbar, and sacral regions of the spine.

We used the Spearman’s rank correlation coefficient to investigate the uneven distribution of malpositioned pedicle screws across different spinal segments (Table 5).

The analysis using Spearman’s rank correlation coefficient revealed statistically significant associations between the number of pedicle screws placed in different spinal regions and the severity of their malposition according to the Gertzbein-Robbins Grading Scale.^[2] The strongest positive correlation was observed between the number of screws placed in the lumbar region and Grade A (ρ=0.643, p<0.001), indicating that screw implantation in the lumbar spine achieves significantly more accurate positioning. Negative correlation was found between the number of lumbar screws and Grade C (ρ=−0.256, p<0.05) and Grade D (ρ=−0.175, p<0.05), further emphasizing that lumbar localization is associated with easier and more precise screw placement.

For screws localized in the sacral region (S1/S2), a moderate positive correlation was found between their number and Grade A (ρ=0.356, p<0.001), along with a weaker negative correlation for Grade C (ρ=−0.169, p=0.023).

The thoracic region was associated with a significant increase in cases of pedicle screw malposition: Grade C (ρ=0.315, p<0.001) and Grade D (ρ=0.191, p=0.010).

All screws with breach grades (Grade C to Grade E) were intraoperatively revised following analysis of the intraoperative imaging obtained by the O-arm (Figs 1, 2).

The preoperative neurological status was assessed using the Modified McCormick Scale. It was 2.97 (2–5) for all patients, varying from 2.67 to 3.72. The average postoperative score for all patients was 2 (range 1–5), with variation from 1.6 to 2.96 depending on the specific spinal disease. None of the patients developed additional neurological deficits related to the surgery itself.