Precision in Cervical Spine Surgery: A Systematic Review and Comparative Meta-analysis of Navigated Guides for Safe and Effective Pedicle Screw Fixation

1. Literature Search Strategy

PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-analyses) [16] criteria (Figure 1) were followed in conducting this systematic review. The protocol for this review was published in the PROSPERO international prospective register of systematic reviews (CRD42024592914). This study was approved by the Ethics Committee and Institutional Review Board: R141h/67. A literature search was performed using the PubMed search engine to collect articles published in PubMed between 2019 and 2024 using the MeSH (medical subject headings) terms on September 22nd, 2024. The search strategy for inclusion in this study was ("Cervical spine" OR "Navigation assisted system") OR ("3D printing template" OR "computer-assisted system"). Additional manual checks of the reference lists were also accomplished. Only articles written in English were considered for inclusion.

Figure 1.

PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-analyses) flow diagram of study selection. *Records were identified through electronic database searches conducted in PubMed.

2. Inclusion and Exclusion Criteria

Randomized controlled trials, observational studies, and case control were included in this systematic review. There were no restrictions on the type of study, e.g., retrospective or prospective; however, all studies had to be published in English. The search did not include case reports, cases review, case studies or technical notes. All of the identified articles' titles and abstracts were evaluated and reviewed. All articles and journals were thoroughly studied, and essential details were recorded. Studies were included if they assessed the performance of these technologies in terms of accuracy, safety, radiation exposure, and operative time in human subjects. Only studies with more than 10 patients were included. Animal or cadaveric studies were excluded unless they were highly relevant for the specific purpose of developing new techniques. Studies were only included in the network meta-analysis if they included a consecutive group of patients, assessed accuracy of placement on intraoperative or postoperative CT.

3. Data Extraction

The following data on the studies were recorded: (1) author, year, and country of publication, (2) type of study design, (3) number of participants, (4) level of fixation, (5) intervention, (6) comparison, (7) performance of technique, (8) sensitivity and specificity, (9) error rate, (10) key finding, and (11) conclusion.

4. Assessment of Risk of Bias of This Systematic Review

The Cochrane Collaboration's suggested RevMan 5.4 (Review Manager, Cochrane Collaboration, UK) to determine the risk of bias for individual randomized controlled trials was used to evaluate each trial's risk of bias [17]. The evaluation included selection bias, performance bias, attrition bias, bias in detection, and bias in reporting, each of which were classified as high risk, low risk, or uncertain risk. Two investigators rated the included studies' levels of bias separately before comparing their results. Consensus-based decisions were made in cases of disagreement, and a third author's assessment was sought if necessary. A third reviewer also resolved any remaining disagreements regarding evaluation of the retrieved data. The publication bias was evaluated using a funnel plot.

5. Statistical Analyses

We conducted a network meta-analysis (NMA) using RevMan 5 (Review Manager) and R ver. 3.6.0 (R Foundation for Statistical Computing, Austria). Binary efficacy and safety outcomes were summarized as pooled risk differences (95% confidence interval [CI]) between conventional surgery and 3D printing and computer-assisted navigation. According to the rank order of the treatment method in each iteration of the Markov chain, each outcome was assessed with the probability of which is the best (superior to all other interventions), second best, and third best. Clinical outcome data were reported in term of binary result in accuracy of screw insertion. And they were reported in mean and SD for operative time and blood loss. A random-effects model (DerSimonian-Laird method) was applied to account for expected clinical and methodological heterogeneity. We used the study that reported outcomes in sufficient detail to derive a correlation coefficient, which we used to impute an SD of the change from the baseline for the other studies.

1. Data Analysis

Data on the studies, including author, year and country of publication, type of study design, number of participants, level of fixation, intervention, comparison, performance of technique, sensitivity and specificity, error rate, key finding, and conclusion were reviewed and analyzed (Tables 1 and 2).

Table 1.

Summary of 3-dimensional (3D)-printing guides for cervical spine surgery

Table 2.

Summary of computer-assisted navigation systems for cervical spine surgery

2. Risk of Bias Analysis

A summary of the nonrandomized controlled trial risk of bias is shown in Figure 2. Two studies had a moderate risk of bias in the deviations from intended intervention and missing data, and other had a low risk of confounding variables. All studies had a low risk of confounding, selection of participants, classification of interventions, measurement of outcomes and selection in reported result and funnel plot (Figures 2–4).

Figure 2.

Risk of bias assessment across included studies. This figure presents domain-level and overall risk-of-bias judgments for the 8 included studies, evaluated using the ROBINS-I (Risk of Bias in Non-randomized Studies of Interventions) tool. The 7 assessed domains are D1 (bias due to confounding), D2 (bias in selection of participants), D3 (bias in classification of interventions), D4 (bias due to deviations from intended interventions), D5 (bias due to missing data), D6 (bias in measurement of outcomes), and D7 (bias in selection of the reported result). Each cell displays the reviewer’s judgment for the corresponding domain, with green circles indicating low risk of bias and yellow circles indicating moderate risk. The final column summarizes the overall risk of bias for each study.

Figure 3.

The aggregated risk of bias across studies for each domain and for the overall assessment.

Figure 4.

Funnel plot of study precision versus effect size for navigation versus conventional techniques. Each point represents an individual study; the vertical line indicates the pooled log odds ratio, and the dashed lines represent the pseudo–95% confidence limits.

3. Study Design and Publication Information

In this review, we discovered 5 retrospective studies about 3D printing. For the computer-assisted navigation system, 2 investigations were retrospective, with only one being prospective.

4. Nationality

The number of publications on cervical pedicle screw fixation navigation increased between 2019 and 2024 as the trend toward this surgical technique grew. The 3D printing studies included 5 articles from China. The studies of computer-assisted navigation systems included 2 articles from Germany and 1 article from Japan.

5. Samples Size

In 3D printing studies, a 2021 prospective study conducted in China by Wu et al. [18] had the largest sample size with 71 patients, whereas a retrospective review by Niu et al. (2022) [19] had the smallest sample size of 23 patients. In computer-assisted navigation studies, a prospective study in Germany by Bertram et al. [20] had the largest sample size of 157 patients.

6. Level of Fixation

The level of fixation in cervical spine surgeries varies depending on the study and the specific surgical approach. Most of 3D printing navigation was studied at the atlantoaxial (C1–2) levels, where Wu et al. [18] focused on C1 pedicle screw template-assisted navigation. Niu et al. [19] and Yuan et al. [21] explored atlantoaxial (C1–2) levels, employing 3D-printed drill guides for screw placement. Another Wu et al. [22] studied C2 pedicle screw. However, only Wu et al. [23] focused on lower cervical levels.

For the computer-assisted navigation reported, one study was conducted at the atlantoaxial (C1–2) levels by Gierse et al. [24], while others did not identify the level.

7. Navigation Techniques, Interventions, and Evaluated Classification Usage

The surgical technique for computer-assisted cervical screw insertion begins with preoperative planning, where a high-resolution CT scan is used to generate a 3D model of the patient’s cervical spine. This model helps identify anatomical landmarks, potential obstacles, and the ideal trajectory for screw placement. The CT data is integrated into a navigation system that will guide the surgeon in real-time during the operation. The patient is positioned in the prone position on the operating table, with the head securely fixed to prevent movement. After positioning, a midline or paramedian incision is made to expose the posterior elements of the cervical spine, such as the lamina, lateral masses, or pedicles, depending on the screw type being used. The navigation system is then calibrated, and the spine is registered to the navigation system by identifying key anatomical landmarks. Once this registration is complete, the system provides real-time feedback to guide the surgeon in placing screws with precision. The system allows the surgeon to visualize the optimal path for screw insertion, minimizing the risk of damaging surrounding structures like nerves or blood vessels. Finally, after the screws are inserted, the wound is closed, and the patient is monitored postoperatively [25-27].

Takamatsu et al. [28], Gierse et al. [24], and Bertram et al. [20] compared the computer-assisted navigation to the conventional fluoroscopy-guided technique grading according to the modified Gertzbein and Robbins classification [29], Neo classification [30], and Bredow classifications [31].

Four publications were examined in this study between the traditional freehand method and the 3D navigation technology. The base of the navigation plate was developed by Wu et al. [18] to extend outward to a distance of 3 mm from the screw entrance point and inward to the posterior tubercle. The direction of the virtual screw was used to design a 10 mm hollow navigation tube. The medial diameter of the navigation tube measured 2.1 mm evaluated by modified Gertzbein and Robbins classification. Niu et al. [19] created a guide tube with the following dimensions: 0.4 mm for the inner diameter, 8 mm for the outer diameter, and 15–20 mm for the length using Kawaguchi method evaluation [32]. Wu et al. [22] created the base templates for C2 pedicle screw insertion. Wu et al. [22] stretched the extracted surface 3 mm towards the spinous process to form the navigation template's base. A hollow navigation tube 10 mm in length along the direction of the virtual screw and evaluated by Hlubek classification [33]. Then, Yuan et al. [21] presented construction of the “Pointing-Drilling” double template which is a cylinder with a diameter of 3.5–4 mm was used as a substitute for a screw to directly observe the potential perforation of the C1–2 cortical bone on each plane in the process of 2-dimensional and 3D reconstruction related to freehand technology described by Resnick and Benzel [34], and Tan et al. [35] which using Miyamoto and Uno [36] and Yukawa et al. [7] in evaluation.

Three-dimensional model of the patient’s unique anatomy begins with detailed preoperative imaging, usually a CT scan of the patient’s cervical spine. From this model, patient-specific guides are designed, which indicate the optimal entry points and trajectories for screw placement which fits precisely onto the specific anatomical landmarks of the patient’s spine. Once the screws are inserted through the guide, it is removed, and the surgical team proceeds with closing the incision [37,38].

Wu et al. [23] designed the 3D-printed flexible drill guiding templates for the lower cervical spine (C3–7) compare with the 3D-printed traditional drill guiding templates, which covers the lamina around the screw entry points and the spinous process. The base was set with holes with a diameter of 2.7 mm for 2.5 mm screws. They used a 5 mm long navigation tube with inner diameters of 1.3 and 1.6 mm evaluated by Hlubek classification [33].

8. Performance, Accuracy, and Error Rate

The accuracy of 3D-printed guides and navigation systems in cervical spine surgery has been consistently high, contributing to improved outcomes and reduced complications. Accuracy in screw placement is crucial for preventing neurovascular injuries and ensuring the stability of the spine, particularly in complex cervical levels like C1 and C2. Wu et al. [18] reported that navigation template-assisted C1 pedicle screw placement showed significant accuracy improvements over freehand techniques, with fewer deviations and a "low" error rate, though exact figures were not provided. Niu et al. [19] reported a 95.7% accuracy rate with 3D-printed navigation templates for atlantoaxial screw placement, with a 4.3% error rate, significantly outperforming the 80% accuracy and 20% error rate in the freehand group. Wu et al. [23] showed that lower cervical pedicle screw placement using flexible drill guides had a high accuracy, with a 2.4% error rate, compared to 15.1% in the traditional group. In another study of Wu et al. [22], C2 screw placement achieved 95.8% accuracy and a low error rate, outperforming the 77.5% accuracy in the traditional group. Yuan et al. [21] reported 94.4% accuracy with 3D-printed guide templates for C1–2 screws, with a 5.6% error rate, compared to 87.1% accuracy and 12.9% error in the freehand group.

In the computer-assisted group, Bertram et al. [20] found 92.8% accuracy with intraoperative computed tomography (iCT)-assisted navigation for pedicle screws, compared to 67% with fluoroscopic guidance, with a significant reduction in misplacement rates. Takamatsu et al. [28] showed higher accuracy up to 96.2% in the O-arm group compared to 87.5% in the conventional group. The final study by Gierse et al. [24] found higher accuracy for iCT at both the C1 and C2 levels.

The forest plot displays the odds ratios (ORs) and CIs for individual studies and overall outcomes. Each study listed compares outcomes between navigation-assisted which were 3D printing and computer-assisted navigation and conventional screw placement. Most studies favor navigation, with ORs above 1, indicating improved accuracy with navigation. For instance, studies by Wu et al. [22] and Bertram et al. [20] report highly significant ORs of 6.68 and 6.50, respectively, suggesting a strong benefit from navigation techniques. The overall OR, represented by the diamond at the bottom, is 3.86 (95% CI, 2.56–5.82), showing that navigation-assisted methods significantly improve accuracy over conventional methods. The heterogeneity of the studies (I²=46%) suggests moderate variability among them, but the test for overall effect (Z=6.43, p<0.00001) confirms that navigation is significantly more accurate than conventional approaches (Figure 5).

Figure 5.

Forest plot comparing the odds of success between navigation and conventional techniques. M-H, Mantel-Haenszel; CI, confidence interval; df, degrees of freedom.

We compared 3D printing and computer-assisted navigation accuracy using NMA. For 3D printing versus freehand, 5 studies are considered, with no heterogeneity (I²=0%). Both direct and network estimates indicate that 3D printing has a statistically significant improvement in accuracy, with a relative risk of 1.15 (95% CI, 1.05–1.26; p<0.01). In contrast, the comparison of computer-assisted techniques with freehand is based on 3 studies and shows substantial heterogeneity (I²=85%). Both direct and network estimates indicate a smaller, non- significant improvement in accuracy, with a relative risk of 1.07 (95% CI, 0.94–1.22; p=0.28). The 3D printing navigation showed more accuracy than computer-assisted navigation (Figure 6).

Figure 6.

Network meta-analysis comparing 3-dimensional (3D) printing, computer-assisted, and freehand surgical techniques. The diagram on the left illustrates the 3 interventions. The panel on the right shows the results of direct comparisons between 3D printing versus freehand and computer-assisted versus freehand techniques. RR, risk ratio; CI, confidence interval.

9. Operative Time and Blood Loss

The studies provide detailed comparisons of operative time and blood loss between 3D-printed navigation templates and conventional methods. In Wu et al. [18], the operative time for C1 pedicle screw placement was significantly reduced to 76.47 minutes in the navigation group compared to 125.63 minutes in the freehand group. Niu et al. [19] also reported shorter operative times for C1–2 pedicle screw placement using 3D templates (110 minutes) versus freehand techniques (173.8 minutes), along with lower blood loss (159.6 mL vs. 304 mL). Wu et al. [22] demonstrated a similar reduction in operative time for lower cervical (C3–7) pedicle screws, with the navigation group averaging 95.7 minutes compared to 132.6 minutes, and blood loss was also lower (128.9 mL vs. 253 mL). Yuan et al. [21] reported a slight reduction in operative time (121 minutes vs. 126.3 minutes) and blood loss (235.4 mL vs. 276.6 mL) for C1–2 screws.

For computer-assisted navigation, the studies focused less on blood loss but still provided insight into operative times. Gierse et al. [24] reported reduced operative time with iCT-based navigation for atlantoaxial screw placement, averaging 135 minutes compared to 158 minutes for fluoroscopy-guided methods. However, Takamatsu et al. [28] found that O-arm navigation increased operative time (277 minutes).

10. Key Finding and Conclusion

The key findings and conclusions from both tables highlight the superiority of advanced navigation techniques in cervical spine surgery. Studies using 3D-printed navigation templates consistently demonstrated improved accuracy, reduced operative times, and lower blood loss compared to conventional freehand methods. Wu et al. [18] showed that C1 pedicle screw placement using navigation templates led to shorter operative times and less radiation exposure. Niu et al. [19] found that atlantoaxial screw placement with 3D-printed templates resulted in higher accuracy, reduced blood loss, and fewer fluoroscopic exposures compared to freehand techniques. Wu et al. [23] showed that flexible drill guiding templates for lower cervical screws achieved more precise placement with shorter surgical times and less blood loss. Similarly, Yuan et al. [21] reported improved accuracy and reduced radiation exposure with the "pointing-drilling" guide template for C1–2 screws.

For computer-assisted navigation, the key findings focus on increased accuracy, especially in complex cases. Gierse et al. [24] demonstrated that iCT-based navigation significantly reduced operative times and improved accuracy in atlantoaxial screw placement compared to fluoroscopy. Bertram et al. [20] showed that iCT navigation enhanced the accuracy of pedicle screw placement, particularly in trauma and degenerative cases, though they did not provide specific results for operative time or blood loss. Takamatsu et al. [28] found that while O-arm navigation improved accuracy, it resulted in longer operative times.

In conclusion, both 3D-printed navigation templates and computer-assisted systems such as iCT and O-arm demonstrate significant improvements in surgical accuracy and safety, particularly at the C1–2 levels. The 3D-printed navigation templates also show clear benefits in reducing operative time and blood loss, making them a promising tool for improving surgical outcomes.