Evaluating robotic pedicle screw placement against conventional modalities: a systematic review and network meta-analysis

View More View Less
  • 1 Carle Illinois College of Medicine, University of Illinois Urbana-Champaign, Champaign; and
  • | 2 Department of Neurosurgery, Carle Neuroscience Institute, Carle Foundation Hospital, Urbana, Illinois
Free access

OBJECTIVE

Several approaches have been studied for internal fixation of the spine using pedicle screws (PSs), including CT navigation, 2D and 3D fluoroscopy, freehand, and robotic assistance. Robot-assisted PS placement has been controversial because training requirements, cost, and previously unclear benefits. This meta-analysis compares screw placement accuracy, operative time, intraoperative blood loss, and overall complications of PS insertion using traditional freehand, navigated, and robot-assisted methods.

METHODS

A systematic review was performed of peer-reviewed articles indexed in several databases between January 2000 and August 2021 comparing ≥ 2 PS insertion methods with ≥ 10 screws per treatment arm. Data were extracted for patient outcomes, including PS placement, misplacement, and accuracy; operative time, overall complications, intraoperative blood loss, postoperative hospital length of stay, postoperative Oswestry Disability Index (ODI) score, and postoperative visual analog scale (VAS) score for back pain. Risk of bias was assessed using the Newcastle-Ottawa score and Cochrane tool. A network meta-analysis (NMA) was performed to estimate PS placement accuracy as the primary outcome.

RESULTS

Overall, 78 studies consisting of 6262 patients and > 31,909 PSs were included. NMA results showed that robot-assisted and 3D-fluoroscopy PS insertion had the greatest accuracy compared with freehand (p < 0.01 and p < 0.001, respectively), CT navigation (p = 0.02 and p = 0.04, respectively), and 2D fluoroscopy (p < 0.01 and p < 0.01, respectively). The surface under the cumulative ranking (SUCRA) curve method further demonstrated that robot-assisted PS insertion accuracy was superior (S = 0.937). Optimal screw placement was greatest in robot-assisted (S = 0.995) placement, and misplacement was greatest with freehand (S = 0.069) approaches. Robot-assisted placement was favorable for minimizing complications (S = 0.876), while freehand placement had greater odds of complication than robot-assisted (OR 2.49, p < 0.01) and CT-navigation (OR 2.15, p = 0.03) placement.

CONCLUSIONS

The results of this NMA suggest that robot-assisted PS insertion has advantages, including improved accuracy, optimal placement, and minimized surgical complications, compared with other PS insertion methods. Limitations included overgeneralization of categories and time-dependent effects.

ABBREVIATIONS

GRS = Gertzbein-Robbins scale; LOS = length of stay; NMA = network meta-analysis; ODI = Oswestry Disability Index; PS = pedicle screw; RCT = randomized controlled trial; SMD = standardized mean difference; SUCRA = surface under the cumulative ranking; VAS = visual analog scale.

OBJECTIVE

Several approaches have been studied for internal fixation of the spine using pedicle screws (PSs), including CT navigation, 2D and 3D fluoroscopy, freehand, and robotic assistance. Robot-assisted PS placement has been controversial because training requirements, cost, and previously unclear benefits. This meta-analysis compares screw placement accuracy, operative time, intraoperative blood loss, and overall complications of PS insertion using traditional freehand, navigated, and robot-assisted methods.

METHODS

A systematic review was performed of peer-reviewed articles indexed in several databases between January 2000 and August 2021 comparing ≥ 2 PS insertion methods with ≥ 10 screws per treatment arm. Data were extracted for patient outcomes, including PS placement, misplacement, and accuracy; operative time, overall complications, intraoperative blood loss, postoperative hospital length of stay, postoperative Oswestry Disability Index (ODI) score, and postoperative visual analog scale (VAS) score for back pain. Risk of bias was assessed using the Newcastle-Ottawa score and Cochrane tool. A network meta-analysis (NMA) was performed to estimate PS placement accuracy as the primary outcome.

RESULTS

Overall, 78 studies consisting of 6262 patients and > 31,909 PSs were included. NMA results showed that robot-assisted and 3D-fluoroscopy PS insertion had the greatest accuracy compared with freehand (p < 0.01 and p < 0.001, respectively), CT navigation (p = 0.02 and p = 0.04, respectively), and 2D fluoroscopy (p < 0.01 and p < 0.01, respectively). The surface under the cumulative ranking (SUCRA) curve method further demonstrated that robot-assisted PS insertion accuracy was superior (S = 0.937). Optimal screw placement was greatest in robot-assisted (S = 0.995) placement, and misplacement was greatest with freehand (S = 0.069) approaches. Robot-assisted placement was favorable for minimizing complications (S = 0.876), while freehand placement had greater odds of complication than robot-assisted (OR 2.49, p < 0.01) and CT-navigation (OR 2.15, p = 0.03) placement.

CONCLUSIONS

The results of this NMA suggest that robot-assisted PS insertion has advantages, including improved accuracy, optimal placement, and minimized surgical complications, compared with other PS insertion methods. Limitations included overgeneralization of categories and time-dependent effects.

Since the 1970s, transpedicular screws, or pedicle screws (PS), have been utilized for spondylodesis, which is the internal fixation of 2 or more vertebral segments to stabilize the spine in a wide array of treatment of spinal conditions, including degenerative disc disease, vertebral trauma, spinal neoplasia, and infection.1,2 Screw placement was originally achieved with radiograph-based plans and anatomical landmarks using a freehand intraoperative approach.1 While major innovations over the past several decades have improved the field of interbody fusion since its inception, the risks of all approaches are still high, with failed back surgery syndrome occurring in an estimated 10% to 46% of cases.3 Complications can be severely debilitating, although rarely life-threatening, when they occur.4

Based on the freehand approach to spondylodesis, surgeons found improvement in patient outcomes when they paired the anatomical freehand approaches with fluoroscopy. These early surgical navigation concepts evolved into the various forms of surgical navigation systems used today. Modern, popular platforms for operative navigation include the Airo mobile intraoperative CT-based spinal navigation (Brainlab), Stryker spinal navigation with the SpineMask tracker and SpineMap software (Stryker), StealthStation surgical navigation and the O-arm imaging system (Medtronic), and Ziehm Vision FD Vario 3D with NaviPort integration (Ziehm Imaging).5,6

With the approval of the Mazor SpineAssist in 2004, robots began to find use in PS placement for fusion procedures. To date, there are 3 robots available in the US that are approved by the FDA for PS placement: Mazor X Stealth Edition (Medtronic), ExcelsiusGPS (Globus Medical), and ROSA (Zimmer Biomet).7 Mazor led the market with the first and most extensively used robotic spine surgery platform.8 ROSA aims at operating as a more universal neurosurgery robot; it was previously used for cranial procedures and has recently acquired FDA approval for spine surgery with the release of the ROSA ONE Spine platform.9 ExcelsiusGPS expanded treatment options to more complicated procedures in cervical, thoracic, and sacral regions while significantly reducing the radiation dose with an innovative intraoperative metric for screw placement accuracy.10

Recent studies have demonstrated that placement of PSs with robotic assistance is controversial, with limited benefits associated with significant end-user experience variability, longer operating room times, longer training times, additional training platforms, and a long learning curve compared with traditional freehand approaches.11,12 While some systematic reviews and trials demonstrated that other methodologies have improved outcomes when compared with robotic surgery,13,14 there are meta-analyses that have posited the superiority of robotic approaches to PS placement.11,15 The largest meta-analysis to date concluded that CT navigation was the optimal approach for maximizing PS placement accuracy;16 however, it only compared single-cohort studies of indirect comparisons. In this study, we present a comprehensive network meta-analysis (NMA) of 6262 patients and > 31,909 screws across 78 studies to understand the comparative accuracy, rate of misplacement, and operative and postoperative patient outcomes of various PS placement modalities, including traditional freehand, navigated, and robotic guidance approaches.

Methods

Search Strategy

A systematic search following PRISMA guidelines was performed. This review was not registered. An a priori search protocol can be obtained on request from the corresponding author. Searches for publications of randomized controlled trials and observational studies from January 2000 to August 2021 were performed using Cochrane Library, Ovid, EMBASE, Google Scholar, MEDLINE, Scopus, PubMed, and Web of Knowledge. No language restrictions were imposed. The search was restricted to January 2000 to prevent era-associated biases. Search terms included: robot, robotic, robotics, robotic surgery, robot-assisted, robotic surgical procedures, spine surgery, spinal surgery, spine, thoracic vertebrae, lumbar vertebrae, cervical vertebrae, zygapophyseal joint, facet joint, computer-assisted, neuronavigation, navigation, frameless stereotaxis, artificial intelligence, augmented reality, fluoroscopic images, two-dimensional images, computed tomography-based/CT navigation, two-dimensional/2D fluoroscopy-based navigation, three-dimensional/3D fluoroscopy-based navigation, pedicle screw, pedicle screws, pedicle screw insertion, bone screws, traditional trajectory screws, freehand, C-arm, O-arm, accuracy. After removing duplicates, full-text articles that provide direct comparison against PS insertion with a spine robot were included. The included studies also provided sufficient data to allow meaningful comparison (> 10 PSs in each treatment arm). Articles were excluded that did not provide a treatment comparison. Publications such as review articles, systematic reviews, meta-analyses, technical reports, case reports, or experimental studies were also excluded.

Data Extraction

Prior to quantitative analysis, authors (A.N., A.D.S., D.T.K., and A.S.) extracted data for patient outcomes such as screw placement accuracy, optimal screw insertion, screw misplacement, operative time, overall complications, intraoperative blood loss, postoperative hospital length of stay (LOS), postoperative Oswestry Disability Index (ODI) score, and postoperative visual analog scale (VAS) score for back pain.

Accurate placement of PSs was defined as grade A and B placement according to the Gertzbein-Robbins scale (GRS),17 grade I and II from the Ravi scale,18 and grade A and B based on the Rampersaud criteria.19 Optimal placement was defined as a subset of accurate placement and only included grade 0 placement on the Wiesner scale,20 grade A on the GRS and Rampersaud scale, and grade I on the Ravi scale. Previous studies have demonstrated the equivalency and effectiveness of the scales to describe accurate PS placement.21 Screw misplacements were defined as screws that required replacement intraoperatively. Overall complications were pooled to represent any postoperative complication present. These findings were initially substratified; however, due to the heterogeneous categorizations, they were aggregated into one outcome. Intraoperative blood loss was defined in milliliters, and operative time was uniformly defined as the incision-to-close surgical procedure time. Postoperative LOS was defined by the number of days a patient remained in the hospital until discharge. A limited number of studies reported postoperative ODI, and VAS back outcomes.22,23 These measures provide information on functional patient outcomes and, thus, were included in the analysis. Higher ODI scores suggested increased functionality postoperatively. Higher VAS scores were suggestive of worse pain postoperatively. All extracted variables were cross-validated independently by at least one other author. Conflicts were resolved by unanimous consensus on discussion with senior authors. Data were tabulated on a spreadsheet.

Assessment of Quality

Quality assessment was performed independently by 3 authors (A.N., D.T.K., and C.M.M.), followed by unanimous cross-validation. For retrospective and prospective nonrandomized cohort trials, the Newcastle-Ottawa scale was used to generate a quality score. For randomized controlled trials (RCTs), the Cochrane risk-of-bias tool was used. From these results, the Agency for Healthcare Research and Quality standardized reporting categories were applied, whereby studies were labeled as having poor, sufficient, or good quality.

Statistical Methodology

Given the presence of substantial evidence for multiple navigation modalities, a frequentist network meta-analysis (NMA), pooling odds ratios (OR [95% CI]) and standardized mean differences (SMD [95% CI]), was performed to provide a statistical comparison. To determine the usage of a fixed- or random-effects model, internal and external heterogeneity was assessed using the significance of the Cochrane’s Q statistic. Significance was determined to be p < 0.05. From the findings of the NMA, ordering of navigation strategies for the outcome under consideration was performed via the surface under the cumulative ranking (SUCRA) curve method, which provides an aggregated and weighted score for each node of the NMA from 0 to 1. Publication bias was assessed using funnel plots, comparing the odds ratio and standard error of each outcome. Significance of publication bias was assessed using the Egger’s test for symmetry. All analyses were conducted using the R package netmeta (The R Project).

Results

Our search strategy yielded 78 studies that matched the defined inclusion and exclusion criteria (Supplemental Table 1). Quality assessment outcomes for observational studies using the Newcastle-Ottawa scale are included in Supplemental Table 1, and RCTs using the Cochrane’s risk-of-bias tool are included in Supplemental Table 2. The included studies spanned 5 different PS placement strategies, including robot-assisted placement, 2D fluoroscopic navigation, 3D fluoroscopic navigation, CT navigation, and freehand screw placement. Across all cohorts, 6262 patients and > 31,909 total screws placed were included in the analysis (Fig. 1).

FIG. 1.
FIG. 1.

PRISMA diagram for studies included in the meta-analysis.

Placement Accuracy

We performed a frequentist NMA of 53 studies providing 60 direct comparisons for the PS placement accuracy between 5 navigation strategies (Fig. 2A). No studies compared freehand screw placement with 2D fluoroscopic navigation and CT navigation; thus, only indirect comparisons were available for these edges. Heterogeneity of comparisons revealed significant heterogeneity within designs (Q = 90.26 [df 52], p < 0.01) and between designs (Q = 31.88 [df 4], p < 0.01). For this reason, a random-effects model was used for the pooling of outcomes. Figure 2B demonstrates the outcomes of the NMA. Robotic placement and 3D-fluoroscopic navigation had significantly higher accuracy compared with freehand placement (p < 0.01 and p < 0.001, respectively), CT navigation (p = 0.02 and p = 0.04, respectively), and 2D-fluoroscopic navigation (p < 0.01 and p < 0.01, respectively). There was no significant difference between robotic and 3D-fluoroscopic navigation when compared with each other. SUCRA hierarchical ranking for PS placement accuracy demonstrated the superiority of robotic placement (S = 0.937), followed by 3D-fluoroscopic navigation (S = 0.807) (Fig. 3). CT navigation (S = 0.386) and freehand placement (S = 0.267) were ranked higher than 2D-fluoroscopic navigation (S = 0.104).

FIG. 2.
FIG. 2.

Comparison of PS placement accuracy. A: Network diagram of direct evidence. B: Results of the NMA comparing navigation modalities.

FIG. 3.
FIG. 3.

SUCRA hierarchical ranking of outcomes with a favorable outcome as the maximum score.

Optimal Placement and Screw Misplacement

Optimal placement of PSs was assessed as a secondary outcome by an NMA of 36 studies providing 37 direct comparisons across all 5 navigation modalities (Fig. 4A). No direct comparisons were present comparing freehand placement with CT navigation and 2D-fluoroscopic navigation, in addition to CT navigation compared with 3D-fluoroscopic navigation and 2D-fluoroscopic navigation. Figure 4B demonstrates the findings of the NMA for optimal placement of PSs. Heterogeneity was found to be significant within (Q = 90.55 [df 31], p < 0.001) and between (Q = 24.35 [df 2], p < 0.001) designs. For this reason, a a random-effects model was used. Robotic placement had significantly improved optimal placement compared with freehand placement (OR 2.03 [95% CI 1.57–2.62], p < 0.01), 3D-fluoroscopic placement (OR 1.6 [95% CI 1.14–2.26], p = 0.002), 2D-fluoroscopic navigation (OR 3.36 [95% CI 1.88–6.02], p < 0.01), and CT navigation (OR 2.48 [95% CI 1.06–5.80], p = 0.03). Additionally, 3D-fluoroscopic navigation was significantly better than 2D-fluoroscopic navigation (p < 0.01). SUCRA hierarchical ranking demonstrated the superiority of robotic navigation (S = 0.995), followed by 3D-fluoroscopic navigation (S = 0.690) (Fig. 3).

FIG. 4.
FIG. 4.

Comparison of optimal screw placement, screw misplacement, and complications. A, C, and E: Network diagram of direct evidence. B, D, and F: Results of the NMA comparing navigation modalities. p < 0.05 was considered significant.

Screw misplacement was assessed with an NMA of 23 studies with 23 direct comparisons across 5 navigation strategies (Fig. 4C). Four indirect-only comparisons were required for analysis, including CT navigation compared with 2D-fluoroscopic navigation, 3D-fluoroscopic navigation, and freehand placement. Compared with freehand, 3D-fluoroscopic and CT navigation were found to have significantly decreased odds of misplacement (OR 0.26 [95% CI 0.12–0.54] and OR 0.36 [95% CI 0.15–0.83], respectively) (Fig. 4D). SUCRA hierarchical ranking (Fig. 3) was performed such that the lowest rates of misplacement were associated with a high SUCRA score. Therefore, 3D-fluoroscopic navigation was associated with the highest rank (S = 0.909), followed by CT navigation (S = 0.720). Freehand navigation was ranked lowest and was associated with the highest rates of misplaced screws (S = 0.069).

Overall Complications

Rates of complications were assessed as an outcome from 23 studies with 23 direct comparisons across 5 navigation approaches. Four comparisons necessitated indirect-only comparisons in the NMA (Fig. 4E). Total heterogeneity in this comparison was found to be nonsignificant (Q = 7.48 [df 19, p = 0.99), thus, a fixed-effects model was used. Our analysis, shown in Fig. 4F, demonstrated that freehand placement had higher reported odds of complications than robotic placement (OR 2.49 [95% CI 1.48–4.18], p < 0.01) and CT navigation (OR 2.15 [95% CI 1.03–4.49], p = 0.03). All other comparisons demonstrated no significant differences in complication rates. An analysis of subclassifications of complications was not done due to an inability to meet the necessary comparisons for the NMA from the available evidence. SUCRA hierarchical ranking demonstrated the favorability of robotic placement of screws to minimize complications (S = 0.876), followed by CT navigation (S = 0.754) (Fig. 3).

Operative Time, Intraoperative Blood Loss, and Postoperative LOS

Operative time and intraoperative blood loss were assessed as additional secondary surgical outcomes for PS placement. Results from 35 studies were compared to provide 38 direct comparisons for a NMA, as demonstrated by the network diagram in Fig. 5A. Three indirect-only comparisons were additionally required for analysis. An SMD > 0 indicated longer operative times. CT navigation and freehand surgery were found to have the longest reported operative times when compared with robotic placement (SMD 0.91, p < 0.01 and SMD 0.81, p < 0.01, respectively) (Fig. 5B). Robotic placement had no difference in operative time when compared with 2D- and 3D-fluoroscopic navigation. Intraoperative blood loss was assessed by reported volumes in milliliters from 17 studies, forming 17 direct comparisons. Three indirect-only comparisons, evident in the network diagram in Fig. 5C, were requisite for analysis. An SMD > 0 was defined to indicate less blood loss, favoring the comparative intervention. Compared with freehand placement, robotic placement, 3D-fluoroscopic navigation, and CT navigation were associated with significantly less blood loss (p < 0.01, p < 0.01, and p = 0.02, respectively). Supporting this finding, SUCRA scores for freehand PS placement were the highest (S = 0.99) (Fig. 3). There were 17 studies included in the NMA to compare postoperative hospital LOS, a secondary outcome measured in our analysis (Fig. 5E). Compared with freehand placement, robotic navigation was associated with a decreased postoperative LOS (SMD 0.57 [95% CI 0.16–0.97], p = 0.01). No other significant relationships were observed. SUCRA ranking confirmed these findings, demonstrating the superiority of robotic surgery with the favorable outcome of a reduced postoperative hospital LOS (S = 0.91), followed by 3D-fluoroscopic navigation (S = 0.69) (Fig. 3).

FIG. 5.
FIG. 5.

Comparison of operative time (minutes), intraoperative blood loss (mL), and postoperative hospital stay (days). A, C, and E: Network diagram of direct evidence. B, D, and F: Results of the NMA comparing navigation modalities. p < 0.05 was considered significant.

Postoperative Functional Outcomes

Postoperative functional outcomes as assessed by the ODI and VAS back were done excluding CT navigation due to the lack of reported comparisons. Eight studies reported both outcomes; the network diagrams are shown in Fig. 6A and C. Only indirect comparisons were available for analysis between 2D- and 3D-fluoroscopic navigation. Figure 6B and D shows pairwise comparisons. For the ODI, a significant SMD was observed with freehand placement compared with 3D-fluoroscopic navigation (SMD −1.18 [95% CI −1.97 to −0.39], p = 0.01). The VAS showed similar comparisons. The SUCRA ranking for both postoperative variables demonstrates that 2D-fluoroscopic navigation and freehand placement had the most optimal postoperative functional outcomes for patients (Fig. 6E).

FIG. 6.
FIG. 6.

Comparison of postoperative functional outcomes. A and C: Network diagrams of direct evidence between navigation approaches. B and D: Results of the NMA comparing navigation modalities. p < 0.05 was considered significant. E: SUCRA hierarchical ranking analysis using a random-effects model.

Publication Bias

Funnel plots for outcomes are shown in Fig. 7. No significant publication bias was observed for any outcomes tested (p > 0.05, Egger’s test).

FIG. 7.
FIG. 7.

Funnel plots showing publication bias. Significance assessed with the Egger’s test. p < 0.05 was considered significant.

Discussion

Robotic guidance in the placement of PSs has remained controversial. We attempted to build on the work of previous studies, including large randomized controlled trials and meta-analyses, to help answer how robot-guided PS placement compares with other modalities such as fluoroscopy and CT navigation. In a comprehensive multivariate meta-analysis by Perdomo-Pantoja et al.,16 78 single-cohort retrospective and prospective studies were pooled to determine the efficacy among navigation modalities. While the analysis was able to present a comparison between the pooled accuracy and revision rate, the reliance on indirect-only analysis presented an important limitation on the generalizability of the comparative outcomes presented. In our study, we only included comparative trials and utilized an NMA approach because of the statistical strength and variant efficiency.24,25

In our NMA, we found that robotic surgery has significant advantages as it pertains to the accuracy and optimal placement of PSs, operative time, and minimization of complications. This finding builds on the outcomes presented by other meta-analyses. Two meta-analyses of randomized controlled trials corroborate some of our expanded findings, reporting the superiority of robotic guidance over freehand screw placement.26,27 Additionally, in an attempt to compare robotic screw placement with computer-assisted navigation modalities, a meta-analysis of 6 studies demonstrated the superiority of robotic guidance.28 Because CT navigation and 3D fluoroscopy were integrated under computer-assisted navigation, our analysis added to the previously presented findings. We additionally demonstrated that while robotic guidance is better than CT navigation, it is not significantly better than 3D-fluoroscopic navigation using technologies such as O-arms. The superiority of 3D fluoroscopy over freehand placement has been demonstrated in previous meta-analyses.29

Notably, Ringel et al. demonstrated a significantly lower accuracy rate of 58.2% for robotic placement of screws than other studies included in our study.30 That paper, published in 2012, used the SpineAssist robot for all procedures. This observation was explored by another trial using the SpineAssist system, which argued that the percutaneous cannula for screw insertion was potentially interacting with bony surfaces, leading to decreased accuracy.15 Since then, robotics systems have adapted to create better drill guides and improved haptic feedback for surgeons.31,32 This would explain our observation that the accuracy of screw placement has generally improved over time as technologies have improved to address such challenges.

Additionally, we demonstrate that robotic surgery is comparable with other technologies when considering screw misplacement, intraoperative blood loss, and postoperative functional outcomes. Previous studies have demonstrated that while screw misplacement is observed in robotic placement, these instances have not yielded a clinical impact to the patient.32 Additionally, postoperative functional outcomes, specifically those assessed with the ODI and VAS for back pain, have demonstrated similar outcomes across cohorts; however, the findings reported are incomplete, as they do not consider preoperative functionality, which was infrequently reported in cohort studies and RCTs. Additionally, other robust measures should also be considered in future studies, including the Progressive Isoinertial Lifting Evaluation (PILE), which relies on movement capacity in addition to patient reporting.33 To address the insufficient data in this domain, a robust prospective clinical trial is evaluating the long-term need of revision surgery, and other functional outcomes, associated with the use of robotics.34

Due to insufficient homogeneously reported evidence, certain outcomes were excluded from analysis, including intraoperative radiation. This was done due to the inconvertibility of the reported radiation units. Compared with freehand and fluoroscope-guided PS placement, previous trials have demonstrated the utility of robotic placement in reducing intraoperative radiation.26 This is due to fluoroscopic guidance requiring near-continuous irradiation of the field of view, whereas robotic placement relies on preoperative imaging.35

As with any meta-analysis, there are limitations to this study. Given the heterogeneity of reporting among studies, it was necessary to place navigation approaches in larger categories. This included the consolidation of different types of robots and different types of fluoroscopic technologies. This methodology was applied to simplify the analysis and include as many retrospective studies as possible. Many retrospective studies and clinical trials did not specify the exact robot or technology used (e.g., O-arm or C-arm) for each individual patient. For this reason, aggregation and broad categorization were essential for analysis. This aggregation and consolidation, however, may have led to decreased effect sizes for the specific technologies in question. Additionally, different pathologies requiring PS placement were not uniquely identified. It is possible that robotic placement of PSs may not be the optimal approach for patients with certain pathologies, such as scoliosis. Additionally, robotic placement of screws may be more optimal for specific spinal locations, such as was demonstrated with thoracic placement.15 Another limitation arises from the inclusion of observational studies in addition to RCTs. While this increased the power of our analysis, this inclusion additionally increased the heterogeneity of our data.

Although this paper demonstrates that robotic surgical placement of PSs has many advantages, it also demonstrates the need for improvement. Technologies that work to minimize blood loss and costs for patients are actively being investigated. With the advent of new robots that demonstrate improved patient outcomes, it may be possible for robotic spine surgery to become more ubiquitous and more widely accepted.

Conclusions

We performed an NMA of clinical studies evaluating the efficacy of robot-guided PS placement compared with other conventional modalities. The results of our analysis show the superiority of robot guidance in screw placement accuracy compared with other approaches, and its equivalency in terms of postoperative functional outcomes. Robotic assistance in spine surgery is a rapidly emerging approach and it should be continually evaluated for its clinical utility and safety, and improvement of patient outcomes.

Disclosures

The authors report no conflict of interest concerning the materials or methods used in this study or the findings specified in this paper.

Author Contributions

Conception and design: Naik, Smith, Krist. Acquisition of data: Naik, Smith, Shaffer. Analysis and interpretation of data: Naik, Smith, Shaffer, Krist, Moawad. Drafting the article: Arnold, Naik, Smith, Krist, Moawad, MacInnis. Critically revising the article: Arnold, Naik, Smith, Moawad, MacInnis, Teal, Hassaneen. Reviewed submitted version of manuscript: all authors. Approved the final version of the manuscript on behalf of all authors: Arnold. Statistical analysis: Naik, Krist. Administrative/technical/material support: Arnold, MacInnis, Teal, Hassaneen. Study supervision: Arnold, Naik, Teal, Hassaneen.

Supplemental Information

Online-Only Content

Supplemental material is available online.

References

  • 1

    Kabins MB, Weinstein JN. The history of vertebral screw and pedicle screw fixation. Iowa Orthop J. 1991;11:127136.

  • 2

    Mobbs RJ, Phan K, Malham G, Seex K, Rao PJ. Lumbar interbody fusion: techniques, indications and comparison of interbody fusion options including PLIF, TLIF, MI-TLIF, OLIF/ATP, LLIF and ALIF. J Spine Surg. 2015;1(1):218.

    • PubMed
    • Search Google Scholar
    • Export Citation
  • 3

    Burton CV. Failed back surgery patients: the alarm bells are ringing. Surg Neurol. 2006;65(1):56.

  • 4

    Reis RC, de Oliveira MF, Rotta JM, Botelho RV. Risk of complications in spine surgery: a prospective study. Open Orthop J. 2015;9:2025.

  • 5

    Overley SC, Cho SK, Mehta AI, Arnold PM. Navigation and robotics in spinal surgery: where are we now? Neurosurgery. 2017;80(3S):S86S99.

  • 6

    Rawicki N, Dowdell JE, Sandhu HS. Current state of navigation in spine surgery. Ann Transl Med. 2021;9(1):85.

  • 7

    Vo CD, Jiang B, Azad TD, Crawford NR, Bydon A, Theodore N. Robotic spine surgery: current state in minimally invasive surgery. Global Spine J. 2020;10(2)(suppl):34S40S.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 8

    D’Souza M, Gendreau J, Feng A, Kim LH, Ho AL, Veeravagu A. Robotic-assisted spine surgery: history, efficacy, cost, and future trends. Robot Surg. 2019;6:923.

    • PubMed
    • Search Google Scholar
    • Export Citation
  • 9

    Zimmer Biomet. ROSA ONE Spine. Accessed November 10, 2021. https://www.zimmerbiomet.com/en/products-and-solutions/zb-edge/robotics/rosa-spine.html

    • Search Google Scholar
    • Export Citation
  • 10

    Jiang B, Karim Ahmed A, Zygourakis CC, et al. Pedicle screw accuracy assessment in ExcelsiusGPS® robotic spine surgery: evaluation of deviation from pre-planned trajectory. Chin Neurosurg J. 2018;4(1):23.

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 11

    Fan Y, Du JP, Liu JJ, et al. Accuracy of pedicle screw placement comparing robot-assisted technology and the free-hand with fluoroscopy-guided method in spine surgery: an updated meta-analysis. Medicine (Baltimore). 2018;97(22):e10970.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 12

    McDonnell JM, Ahern DP, Ó Doinn T, et al. Surgeon proficiency in robot-assisted spine surgery. Bone Joint J. 2020;102-B(5):568572.

  • 13

    Peng YN, Tsai LC, Hsu HC, Kao CH. Accuracy of robot-assisted versus conventional freehand pedicle screw placement in spine surgery: a systematic review and meta-analysis of randomized controlled trials. Ann Transl Med. 2020;8(13):824.

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 14

    Hyun SJ, Kim KJ, Jahng TA, Kim HJ. Minimally invasive robotic versus open fluoroscopic-guided spinal instrumented fusions: a randomized controlled trial. Spine (Phila Pa 1976).2017;42(6):353358.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 15

    Molliqaj G, Schatlo B, Alaid A, et al. Accuracy of robot-guided versus freehand fluoroscopy-assisted pedicle screw insertion in thoracolumbar spinal surgery. Neurosurg Focus. 2017;42(5):E14.

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 16

    Perdomo-Pantoja A, Ishida W, Zygourakis C, et al. Accuracy of current techniques for placement of pedicle screws in the spine: a comprehensive systematic review and meta-analysis of 51,161 screws. World Neurosurg. 2019;126:664678.e3.

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 17

    Gertzbein SD, Robbins SE. Accuracy of pedicular screw placement in vivo. Spine (Phila Pa 1976).1990;15(1):1114.

  • 18

    Ravi B, Zahrai A, Rampersaud R. Clinical accuracy of computer-assisted two-dimensional fluoroscopy for the percutaneous placement of lumbosacral pedicle screws. Spine (Phila Pa 1976).2011;36(1):8491.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 19

    Rampersaud YR, Pik JHT, Salonen D, Farooq S. Clinical accuracy of fluoroscopic computer-assisted pedicle screw fixation: a CT analysis. Spine (Phila Pa 1976).2005;30(7):E183E190.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 20

    Wiesner L, Kothe R, Schulitz KP, Rüther W. Clinical evaluation and computed tomography scan analysis of screw tracts after percutaneous insertion of pedicle screws in the lumbar spine. Spine (Phila Pa 1976).2000;25(5):615621.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 21

    McKenzie DM, Westrup AM, O’Neal CM, et al. Robotics in spine surgery: a systematic review. J Clin Neurosci. 2021;89:17.

  • 22

    Fairbank JCT. Oswestry Disability Index. Letter. J Neurosurg Spine. 2014;20(2):239242.

  • 23

    Bijur PE, Silver W, Gallagher EJ. Reliability of the visual analog scale for measurement of acute pain. Acad Emerg Med. 2001;8(12):11531157.

  • 24

    Faltinsen EG, Storebø OJ, Jakobsen JC, Boesen K, Lange T, Gluud C. Network meta-analysis: the highest level of medical evidence? BMJ Evid Based Med. 2018;23(2):5659.

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 25

    Riley RD, Jackson D, Salanti G, et al. Multivariate and network meta-analysis of multiple outcomes and multiple treatments: rationale, concepts, and examples. BMJ. 2017;358:j3932.

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 26

    Li HM, Zhang RJ, Shen CL. Accuracy of pedicle screw placement and clinical outcomes of robot-assisted technique versus conventional freehand technique in spine surgery from nine randomized controlled trials: a meta-analysis. Spine (Phila Pa 1976).2020;45(2):E111E119.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 27

    Fatima N, Massaad E, Hadzipasic M, Shankar GM, Shin JH. Safety and accuracy of robot-assisted placement of pedicle screws compared to conventional free-hand technique: a systematic review and meta-analysis. Spine J. 2021;21(2):181192.

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 28

    Zhou LP, Zhang RJ, Sun YW, Zhang L, Shen CL. Accuracy of pedicle screw placement and four other clinical outcomes of robotic guidance technique versus computer-assisted navigation in thoracolumbar surgery: a meta-analysis. World Neurosurg. 2021;146:e139e150.

    • PubMed
    • Search Google Scholar
    • Export Citation
  • 29

    Sun J, Wu D, Wang Q, Wei Y, Yuan F. Pedicle screw insertion: is o-arm-based navigation superior to the conventional freehand technique? A systematic review and meta-analysis. World Neurosurg. 2020;144:e87e99.

    • Search Google Scholar
    • Export Citation
  • 30

    Ringel F, Stüer C, Reinke A, et al. Accuracy of robot-assisted placement of lumbar and sacral pedicle screws: a prospective randomized comparison to conventional freehand screw implantation. Spine (Phila Pa 1976).2012;37(8):E496E501.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 31

    Jiang B, Pennington Z, Azad T, et al. Robot-assisted versus freehand instrumentation in short-segment lumbar fusion: experience with real-time image-guided spinal robot. World Neurosurg. 2020;136:e635e645.

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 32

    Khan A, Meyers JE, Yavorek S, et al. Comparing next-generation robotic technology with 3-dimensional computed tomography navigation technology for the insertion of posterior pedicle screws. World Neurosurg. 2019;123:e474e481.

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 33

    Smeets R, Köke A, Lin CW, Ferreira M, Demoulin C. Measures of function in low back pain/disorders: Low Back Pain Rating Scale (LBPRS), Oswestry Disability Index (ODI), Progressive Isoinertial Lifting Evaluation (PILE), Quebec Back Pain Disability Scale (QBPDS), and Roland-Morris Disability Questionnaire (RDQ). Arthritis Care Res (Hoboken). 2011;63(suppl 11):S158S173.

    • Search Google Scholar
    • Export Citation
  • 34

    Staartjes VE, Molliqaj G, van Kampen PM, et al. The European Robotic Spinal Instrumentation (EUROSPIN) study: protocol for a multicentre prospective observational study of pedicle screw revision surgery after robot-guided, navigated and freehand thoracolumbar spinal fusion. BMJ Open. 2019;9(9):e030389.

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 35

    Lieberman IH, Kisinde S, Hesselbacher S. Robotic-assisted pedicle screw placement during spine surgery. JBJS Essential Surg Tech. 2020;10(2):e0020.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • View in gallery

    PRISMA diagram for studies included in the meta-analysis.

  • View in gallery

    Comparison of PS placement accuracy. A: Network diagram of direct evidence. B: Results of the NMA comparing navigation modalities.

  • View in gallery

    SUCRA hierarchical ranking of outcomes with a favorable outcome as the maximum score.

  • View in gallery

    Comparison of optimal screw placement, screw misplacement, and complications. A, C, and E: Network diagram of direct evidence. B, D, and F: Results of the NMA comparing navigation modalities. p < 0.05 was considered significant.

  • View in gallery

    Comparison of operative time (minutes), intraoperative blood loss (mL), and postoperative hospital stay (days). A, C, and E: Network diagram of direct evidence. B, D, and F: Results of the NMA comparing navigation modalities. p < 0.05 was considered significant.

  • View in gallery

    Comparison of postoperative functional outcomes. A and C: Network diagrams of direct evidence between navigation approaches. B and D: Results of the NMA comparing navigation modalities. p < 0.05 was considered significant. E: SUCRA hierarchical ranking analysis using a random-effects model.

  • View in gallery

    Funnel plots showing publication bias. Significance assessed with the Egger’s test. p < 0.05 was considered significant.

  • 1

    Kabins MB, Weinstein JN. The history of vertebral screw and pedicle screw fixation. Iowa Orthop J. 1991;11:127136.

  • 2

    Mobbs RJ, Phan K, Malham G, Seex K, Rao PJ. Lumbar interbody fusion: techniques, indications and comparison of interbody fusion options including PLIF, TLIF, MI-TLIF, OLIF/ATP, LLIF and ALIF. J Spine Surg. 2015;1(1):218.

    • PubMed
    • Search Google Scholar
    • Export Citation
  • 3

    Burton CV. Failed back surgery patients: the alarm bells are ringing. Surg Neurol. 2006;65(1):56.

  • 4

    Reis RC, de Oliveira MF, Rotta JM, Botelho RV. Risk of complications in spine surgery: a prospective study. Open Orthop J. 2015;9:2025.

  • 5

    Overley SC, Cho SK, Mehta AI, Arnold PM. Navigation and robotics in spinal surgery: where are we now? Neurosurgery. 2017;80(3S):S86S99.

  • 6

    Rawicki N, Dowdell JE, Sandhu HS. Current state of navigation in spine surgery. Ann Transl Med. 2021;9(1):85.

  • 7

    Vo CD, Jiang B, Azad TD, Crawford NR, Bydon A, Theodore N. Robotic spine surgery: current state in minimally invasive surgery. Global Spine J. 2020;10(2)(suppl):34S40S.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 8

    D’Souza M, Gendreau J, Feng A, Kim LH, Ho AL, Veeravagu A. Robotic-assisted spine surgery: history, efficacy, cost, and future trends. Robot Surg. 2019;6:923.

    • PubMed
    • Search Google Scholar
    • Export Citation
  • 9

    Zimmer Biomet. ROSA ONE Spine. Accessed November 10, 2021. https://www.zimmerbiomet.com/en/products-and-solutions/zb-edge/robotics/rosa-spine.html

    • Search Google Scholar
    • Export Citation
  • 10

    Jiang B, Karim Ahmed A, Zygourakis CC, et al. Pedicle screw accuracy assessment in ExcelsiusGPS® robotic spine surgery: evaluation of deviation from pre-planned trajectory. Chin Neurosurg J. 2018;4(1):23.

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 11

    Fan Y, Du JP, Liu JJ, et al. Accuracy of pedicle screw placement comparing robot-assisted technology and the free-hand with fluoroscopy-guided method in spine surgery: an updated meta-analysis. Medicine (Baltimore). 2018;97(22):e10970.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 12

    McDonnell JM, Ahern DP, Ó Doinn T, et al. Surgeon proficiency in robot-assisted spine surgery. Bone Joint J. 2020;102-B(5):568572.

  • 13

    Peng YN, Tsai LC, Hsu HC, Kao CH. Accuracy of robot-assisted versus conventional freehand pedicle screw placement in spine surgery: a systematic review and meta-analysis of randomized controlled trials. Ann Transl Med. 2020;8(13):824.

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 14

    Hyun SJ, Kim KJ, Jahng TA, Kim HJ. Minimally invasive robotic versus open fluoroscopic-guided spinal instrumented fusions: a randomized controlled trial. Spine (Phila Pa 1976).2017;42(6):353358.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 15

    Molliqaj G, Schatlo B, Alaid A, et al. Accuracy of robot-guided versus freehand fluoroscopy-assisted pedicle screw insertion in thoracolumbar spinal surgery. Neurosurg Focus. 2017;42(5):E14.

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 16

    Perdomo-Pantoja A, Ishida W, Zygourakis C, et al. Accuracy of current techniques for placement of pedicle screws in the spine: a comprehensive systematic review and meta-analysis of 51,161 screws. World Neurosurg. 2019;126:664678.e3.

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 17

    Gertzbein SD, Robbins SE. Accuracy of pedicular screw placement in vivo. Spine (Phila Pa 1976).1990;15(1):1114.

  • 18

    Ravi B, Zahrai A, Rampersaud R. Clinical accuracy of computer-assisted two-dimensional fluoroscopy for the percutaneous placement of lumbosacral pedicle screws. Spine (Phila Pa 1976).2011;36(1):8491.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 19

    Rampersaud YR, Pik JHT, Salonen D, Farooq S. Clinical accuracy of fluoroscopic computer-assisted pedicle screw fixation: a CT analysis. Spine (Phila Pa 1976).2005;30(7):E183E190.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 20

    Wiesner L, Kothe R, Schulitz KP, Rüther W. Clinical evaluation and computed tomography scan analysis of screw tracts after percutaneous insertion of pedicle screws in the lumbar spine. Spine (Phila Pa 1976).2000;25(5):615621.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 21

    McKenzie DM, Westrup AM, O’Neal CM, et al. Robotics in spine surgery: a systematic review. J Clin Neurosci. 2021;89:17.

  • 22

    Fairbank JCT. Oswestry Disability Index. Letter. J Neurosurg Spine. 2014;20(2):239242.

  • 23

    Bijur PE, Silver W, Gallagher EJ. Reliability of the visual analog scale for measurement of acute pain. Acad Emerg Med. 2001;8(12):11531157.

  • 24

    Faltinsen EG, Storebø OJ, Jakobsen JC, Boesen K, Lange T, Gluud C. Network meta-analysis: the highest level of medical evidence? BMJ Evid Based Med. 2018;23(2):5659.

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 25

    Riley RD, Jackson D, Salanti G, et al. Multivariate and network meta-analysis of multiple outcomes and multiple treatments: rationale, concepts, and examples. BMJ. 2017;358:j3932.

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 26

    Li HM, Zhang RJ, Shen CL. Accuracy of pedicle screw placement and clinical outcomes of robot-assisted technique versus conventional freehand technique in spine surgery from nine randomized controlled trials: a meta-analysis. Spine (Phila Pa 1976).2020;45(2):E111E119.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 27

    Fatima N, Massaad E, Hadzipasic M, Shankar GM, Shin JH. Safety and accuracy of robot-assisted placement of pedicle screws compared to conventional free-hand technique: a systematic review and meta-analysis. Spine J. 2021;21(2):181192.

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 28

    Zhou LP, Zhang RJ, Sun YW, Zhang L, Shen CL. Accuracy of pedicle screw placement and four other clinical outcomes of robotic guidance technique versus computer-assisted navigation in thoracolumbar surgery: a meta-analysis. World Neurosurg. 2021;146:e139e150.

    • PubMed
    • Search Google Scholar
    • Export Citation
  • 29

    Sun J, Wu D, Wang Q, Wei Y, Yuan F. Pedicle screw insertion: is o-arm-based navigation superior to the conventional freehand technique? A systematic review and meta-analysis. World Neurosurg. 2020;144:e87e99.

    • Search Google Scholar
    • Export Citation
  • 30

    Ringel F, Stüer C, Reinke A, et al. Accuracy of robot-assisted placement of lumbar and sacral pedicle screws: a prospective randomized comparison to conventional freehand screw implantation. Spine (Phila Pa 1976).2012;37(8):E496E501.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 31

    Jiang B, Pennington Z, Azad T, et al. Robot-assisted versus freehand instrumentation in short-segment lumbar fusion: experience with real-time image-guided spinal robot. World Neurosurg. 2020;136:e635e645.

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 32

    Khan A, Meyers JE, Yavorek S, et al. Comparing next-generation robotic technology with 3-dimensional computed tomography navigation technology for the insertion of posterior pedicle screws. World Neurosurg. 2019;123:e474e481.

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 33

    Smeets R, Köke A, Lin CW, Ferreira M, Demoulin C. Measures of function in low back pain/disorders: Low Back Pain Rating Scale (LBPRS), Oswestry Disability Index (ODI), Progressive Isoinertial Lifting Evaluation (PILE), Quebec Back Pain Disability Scale (QBPDS), and Roland-Morris Disability Questionnaire (RDQ). Arthritis Care Res (Hoboken). 2011;63(suppl 11):S158S173.

    • Search Google Scholar
    • Export Citation
  • 34

    Staartjes VE, Molliqaj G, van Kampen PM, et al. The European Robotic Spinal Instrumentation (EUROSPIN) study: protocol for a multicentre prospective observational study of pedicle screw revision surgery after robot-guided, navigated and freehand thoracolumbar spinal fusion. BMJ Open. 2019;9(9):e030389.

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 35

    Lieberman IH, Kisinde S, Hesselbacher S. Robotic-assisted pedicle screw placement during spine surgery. JBJS Essential Surg Tech. 2020;10(2):e0020.

    • Crossref
    • Search Google Scholar
    • Export Citation

Metrics

All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 272 272 272
PDF Downloads 317 317 317
EPUB Downloads 0 0 0