- Email: [email protected]
Accuracy of Pedicle Screw Insertion Among 3 Image-Guided Navigation Systems: Systematic Review and Meta-Analysis
Accepted Manuscript Accuracy of Pedicle Screw Insertion among Three Image-Guided Navigation Systems: A Systematic Review and Meta-Analysis Jin Peng Du, Yong Fan, Qi Ning Wu, Dai Hua Wang, Jing Zhang, Ding Jun Hao PII:
S1878-8750(17)31254-8
DOI:
10.1016/j.wneu.2017.07.154
Reference:
WNEU 6216
To appear in:
World Neurosurgery
Received Date: 31 March 2017 Revised Date:
24 July 2017
Accepted Date: 26 July 2017
Please cite this article as: Du JP, Fan Y, Wu QN, Wang DH, Zhang J, Hao DJ, Accuracy of Pedicle Screw Insertion among Three Image-Guided Navigation Systems: A Systematic Review and MetaAnalysis, World Neurosurgery (2017), doi: 10.1016/j.wneu.2017.07.154. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
1
ACCEPTED MANUSCRIPT
Accuracy of Pedicle Screw Insertion among Three Image-Guided Navigation Systems: A Systematic Review and Meta-Analysis
RI PT
Jin Peng Du1,4 ▪ Yong Fan1,3 ▪ Qi Ning Wu3 ▪ Dai Hua Wang4 ▪ Jing Zhang2,4 ▪ Ding Jun Hao2,3
These authors contributed equally to this study and should be considered co-first
SC
1
2
The co-first authors work for two principal investigators, who should be considered
co-corresponding authors.
Department of Spine Surgery, Xi'an Jiaotong University-Affiliated Honghui
TE D
3
M AN U
authors.
Hospital, Youyidong Road, Xi’an City 710000, China. Email: [email protected],
Medical College, Yan'an University, No 38 Guanghua Road, Yan’an City 716000,
AC C
4
EP
Tel/Fax: 0086-29-87800002. ORCID: 0000-0002-0946-6080
Shanxi Province, China. Email: [email protected], Tel/Fax: 0086-911-2412308
Author Contributions Statement: J.P.D. and Y.F. extracted and analysed the data. D.J.H. and Q.N.W. conceived of and designed the study. D.H.W., J.Z. and D.J.H. participated in the critical revision of the manuscript. J.P.D. and Y.F. drafted the paper.
2
ACCEPTED MANUSCRIPT All authors have read and approved the final submitted manuscript.
Accuracy of Pedicle Screw Insertion among Three Image-Guided Navigation Systems: A Systematic Review and
RI PT
Meta-Analysis
Abstract
SC
Background: Many retrospective studies of pedicle screw placement have revealed that intraoperative navigation systems provide higher accuracy rates and safety than
M AN U
do free-hand techniques. The accuracy of various image-guided navigation systems has been studied; however, differences have not been well defined due to the lack of adequate evidence-based comparative studies.
Objective: A meta-analysis was conducted to focus on the variation in pedicle screw
TE D
insertion among three navigation systems: a three-dimensional fluoroscopy-based navigation system (3D FluoroNav), a two-dimensional fluoroscopy-based navigation
(CT Nav).
EP
system (2D FluoroNav) and a conventional computed tomography navigation system
AC C
Methods: We screened for comparative studies on different pedicle screw insertion navigation systems published through January 2017 using the Cochrane Library, Ovid, Web of Science, PubMed, and EMBASE databases. Results: From 125 papers that were identified, 10 articles were finally chosen. The present comparative study included 8 retrospective clinical studies, 1 prospective clinical trial and 1 randomized controlled cadaveric study. The prevalence rate of pedicle violation in the 3D FluoroNav group was significantly lower than the rates of
3
ACCEPTED MANUSCRIPT the 2D FluoroNav group (RR 95% CI: 0.16-0.61, P<0.01) and the CT Nav group (RR 95% CI: 0.42-0.90, P=0.01), and the rate of the CT Nav group was significantly lower than that of the 2D FluoroNav group (RR 95% CI: 0.29-0.81, P<0.01).
RI PT
Conclusion: There are significant differences among CT Nav, 3D FluoroNav and 2D FluoroNav. Our review suggests that 3D FluoroNav may be superior to the other two methods in reducing pedicle violation and that clinicians should consider 3D
SC
FluoroNav as a better choice.
M AN U
Keywords image-guided navigation, CT navigation, 3D fluoroscopy-based navigation, 2D fluoroscopy-based navigation, pedicle screw.
Introduction
TE D
The introduction of the pedicle screw technique for the treatment of spinal diseases has shown great progress, and this technique has been widely used by spine
1
EP
surgeons. However, the traditional free-hand technique is prone to pedicle violation . Consequently, various navigation systems have been invented to assist screw
AC C
insertion and reduce the incidence of pedicle violation. Three main intraoperative navigation systems are used to assist in pedicle screw placement: a three-dimensional fluoroscopy-based navigation system (3D FluoroNav) (SIREMOBIL Iso-C, Siemens, Forchheim, Germany), a two-dimensional fluoroscopy-based navigation system (2D FluoroNav) (Vector Vision, Brain LAB GmbH, Germany) and a conventional computed tomography navigation system (CT Nav) (Siemens, Forchheim, Germany). However, it was previously unclear which
4
ACCEPTED MANUSCRIPT image-guided navigation system could provide the highest accuracy. Many researchers have suggested that CT Nav is more accurate 2-5, while other scholars have concluded that 3D FluoroNav 6-8 and 2D FluoroNav 9 are more
RI PT
accurate. The results of the existing studies are inconsistent, and the sample sizes were small. Tian NF et al 10 published a meta-analysis demonstrating that these three navigation systems are significantly more accurate than the free-hand method but
SC
could not draw a definitive conclusion regarding the most accurate navigation
M AN U
system due to insufficient sample sizes and comparative research data. Therefore, the present meta-analysis was conducted to help spine surgeons determine which system should be chosen to reduce pedicle breaches.
TE D
Materials and methods Data sources and searches
EP
The Cochrane Library, Ovid, Web of Science, PubMed and EMBASE databases were searched for studies published from 2002 through January 2017 that focused on
AC C
the accuracy of pedicle screw insertion. We explored abstracts, titles and subtitles by searching for the term pedicle screw combined with the following keywords: computed tomography-based/CT navigation , two-dimensional/2D fluoroscopy-based navigation , three-dimensional/3D fluoroscopy-based navigation , and accuracy . The resulting abstracts were inspected to identify articles that should be excluded. Subsequently, the remaining full-text articles were reviewed to select publications that were suitable for inclusion. The study selection process
5
ACCEPTED MANUSCRIPT conformed to the Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) guidelines. The following inclusion criteria were established: (1) publications written in English
RI PT
and (2) articles involving pedicle screw insertion with different types of navigation assistance. The exclusion criteria were as follows: (1) articles not involving a
comparative study, (2) duplicate publications, (3) articles involving pedicle screw
SC
insertion using an image-guided navigation system compared with other methods,
M AN U
and (4) articles with no postoperative CT scanning accuracy assessment. Data extraction
Two reviewers (D.J.P. and F.Y.) independently screened and extracted data by applying the same standards; discrepancies were resolved through discussion until a
TE D
consensus was reached. The following basic information was extracted: first author's name, year of publication, study design, patient characteristics, sample sizes,
EP
interventions, assessment tool and key results. Studies using uniform measurement standards were extracted for further analysis. Specifically, the accuracy of the screw
AC C
in piercing the pedicle, as determined using postoperative CT scanning, was classified into three categories 7: (1) excellent (screw completely within pedicle), (2) acceptable (a portion of the screw outside the pedicle ≤1 mm), and (3) poor (a portion of the screw outside the pedicle >1 mm). The poor grade was considered unacceptable pedicle violation in the analysis.
Data synthesis and statistical analysis
6
ACCEPTED MANUSCRIPT Risk ratios (RRs) with 95% confidence intervals (CIs) were summarized for pedicle violation. The level of significance was set at P<0.05. A forest plot was generated to compare the violation of pedicle screw placement between the experimental and
RI PT
control groups, the point estimate on the left of the vertical line means that the experimental group was significantly more accurate than the control group.
Heterogeneity among pedicle violation in various systems was assessed using the χ2
SC
test. For P>0.1 or I2<50%, no clear heterogeneity of results was assumed.
M AN U
Heterogeneity stems from a variety of different variations among several studies in systematic reviews, which may affect the synergistic effects of several studies and the explanation of the results of the meta-analysis; a strict evaluation is generally required. Although it is impossible to completely prevent, it can be controlled by
TE D
including high-quality studies that all have the same study objective. We selected a fixed-effect model for all analyses unless there was significant heterogeneity. To
EP
examine the stability of the outcomes and identify important sources of slight heterogeneity, influence analyses were conducted by consecutively removing one or
AC C
all subgroup reviews. Statistical analyses were performed using Review Manager Version 5.0 (Cochrane Collaboration, Software Update, Oxford, UK).
Results Literature search The primary search strategy resulted in 580 articles, of which 455 were excluded due to having an irrelevant title or abstract. The remaining 125 papers underwent
7
ACCEPTED MANUSCRIPT full-text evaluation, and 61 were excluded for not meeting the inclusion criteria. Of the remaining papers, 54 were removed per the exclusion criteria. Ultimately, 10 studies were included in this meta-analysis; no additional studies were identified
RI PT
from the references of published results. The present meta-analysis included 1 randomized controlled cadaveric trial5, 2 prospective cohort studies7,17 and 7
retrospective cohort studies2-4,6-9,11. A flow diagram of the report selection is shown
SC
in Fig. 1.
M AN U
Basic characteristics
Detailed characteristics of the involved patients are presented in Table 1. In these trials, the sample size ranged from 24 to 220, including a total of 648 patients and 40 adult cadavers. Altogether, approximately 3255 pedicle screws were inserted in
TE D
patients with various spinal diseases requiring stability reconstruction, and 240 screws were inserted in cadavers. CT scanning was performed for the postoperative
EP
assessment of pedicle violation in all trials. Comparison of pedicle violation
AC C
3D FluoroNav vs CT Nav
As shown in Fig. 2, the first comparative group included 1387 screws inserted using 3D FluoroNav and 887 screws using CT Nav. The 3D FluoroNav group was significantly more accurate (RR 95% CI: 0.42-0.90, P=0.01) than the CT Nav group in guiding pedicle screw fixation. There was no significant heterogeneity (χ2=2.66, P=0.45, I2=0%) between the two groups; therefore, it was not necessary to identify sources of heterogeneity. In addition, Tian et al 5 found the pedicle violation rates of
8
ACCEPTED MANUSCRIPT the two groups to be equal to 0%; therefore, the RR could not be estimated (as shown in Fig. 2), which may have introduced little bias to the results. 3D FluoroNav vs 2D FluoroNav
RI PT
The second comparative group included 474 screws inserted using 3D FluoroNav and 481 screws using 2D FluoroNav. The 3D FluoroNav group was significantly
more accurate (RR 95% CI: 0.16-0.61, P<0.01) than the 2D FluoroNav group in
SC
aiding pedicle screw placement (Fig. 3). There was slight heterogeneity (χ2=3.63,
M AN U
P=0.30, I2=17%) between the two groups. The influence analysis results shown in Table 2 revealed that when Tian et al 5 was removed, there was no significant change in the positive results; the 3D FluoroNav group was still more accurate (RR 95% CI: 0.21-0.89, P=0.02) than the 2D FluoroNav group. However, the slight
TE D
heterogeneity disappeared (χ2=1.17, P=0.56, I2=0%) with this exclusion. When Lekovic et al 11 was excluded, no significant change in the positive results was
EP
shown, and the 3D FluoroNav group was still more accurate (RR 95% CI: 0.06-0.47, P<0.01) than the 2D FluoroNav group; however, the slight heterogeneity
AC C
vanished (χ2=1.46, P=0.48, I2=0%). CT Nav vs 2D FluoroNav The third comparative group included 417 screws inserted using CT Nav and 423 screws using 2D FluoroNav. The CT Nav group was significantly more accurate (RR 95% CI: 0.29-0.81, P<0.01) than the 2D FluoroNav group in assisting pedicle screw placement (Fig. 4). There was slight heterogeneity (χ2=5.79, P=0.22, I2=31%) between the two groups. When Gruetzner et al was rejected, no remarkable variation
9
ACCEPTED MANUSCRIPT in the final outcome was found (Table 2). The CT Nav group was still more accurate (RR 95% CI: 0.22-0.70, P=0.002) than the 2D FluoroNav group, and the heterogeneity decreased (χ2=3.39, P=0.34, I2=11%) with this exclusion. In contrast,
RI PT
the exclusion of Tian et al resulted in extreme alteration (RR 95% CI: 0.37-1.10, P=0.10), and the heterogeneity decreased greatly (χ2=2.33, P=0.51, I2=0%).
SC
Discussion
M AN U
Pedicle screws are widely applied to treat conditions such as spinal degenerative disease, fracture, and deformity, as three-column fixations provide the most rigid form of posterior stabilization. However, pedicle screw malposition remains an inevitable problem. A variety of navigation systems have been introduced in
TE D
response to the desire of spine surgeons to reduce pedicle violation. There are three important, clinically used image-guided navigation methods based on intraoperative
EP
guidance theory: CT Nav, 3D FluoroNav and 2D FluoroNav. The first step in CT Nav is preoperative CT scanning to target the spine level; subsequently, the captured
AC C
preoperative anatomical information is transferred to a navigation platform via matching and registration to assist pedicle screw insertion 12. In contrast to CT Nav, 2D FluoroNav depends on intraoperative two-dimensional images and provides real-time visualization of pedicle anatomy without requiring a complicated registration procedure. With the development of 3D FluoroNav, image-guided navigation systems are rapidly developing and increasing in popularity in the medical field. Because 3D FluoroNav combines the advantages of 2D FluoroNav
10
ACCEPTED MANUSCRIPT and CT Nav without complex matching and single-level or multi-level registration 10
, it provides real-time three-dimensional images, and its quality is comparable to
Meta-analysis of comparisons of pedicle screw accuracy
RI PT
that of CT Nav.
Tian NF et al conducted a systemic review and meta-analysis comparing three
SC
image-guided navigation systems to assist with pedicle screw insertion. These
M AN U
navigation systems provided higher accuracy than conventional methods. The superiority of the navigation systems was apparent when they were applied to abnormal spinal structure10. This study suggested that the significant differences between the three systems are likely due to the lack of comparative studies. In our
TE D
review, a meta-analysis was conducted by identifying additional studies to determine whether there are differences among the systems and which system is the most
EP
accurate.
In the present analysis, the pooled RR values indicate that CT Nav is slightly more
AC C
accurate than 2D FluoroNav (CT Nav vs 2D FluoroNav: RR=0.49, P<0.05) but less accurate than 3D FluoroNav (3D FluoroNav vs 2D FluoroNav: RR=0.31, P<0.05). The bar graph in Fig. 5 shows the overall collective accuracy of each of the three modalities, with pedicle violation classified as excellent (0 mm), acceptable (≤1 mm) or poor (>1 mm). In conclusion, there are significant differences among the three navigation systems, and our study suggests that the 3D FluoroNav system may
11
ACCEPTED MANUSCRIPT be superior in reducing pedicle violation. This conclusion agrees with the in vitro results of Tian NF et al but not with their in vivo results.
RI PT
CT Nav vs 3D FluoroNav
Both CT Nav and 3D FluoroNav can increase the accuracy of screw fixation
SC
compared with the traditional free-hand method. Although CT Nav provides
M AN U
high-quality, real-time intraoperative images, there are some apparent defects. The actual 3D spatial relations during surgery may differ from the CT data obtained preoperatively if the position of the patient changes, possibly leading to errors 7. Moreover, the process of point-to-point registration prolongs the operation time and
TE D
reduces accuracy. However, 3D FluoroNav automatically allows for the collection of 3D-reconstructed images and then transfers them to the navigation system.
EP
Automatic registration and matching can be conducted, as with virtual fluoroscopy navigation. Moreover, 3D FluoroNav combines the advantages of CT Nav and 2D
AC C
FluoroNav 2,8-9,11 and reduces extra preoperative preparation. Geerling et al 13 used a foam spine model to compare the accuracy of CT-based and Iso-C-based navigation. The measured accuracy of the markers was significantly different between CT-based and Iso-C-based navigation for one of six markers placed on the vertebra. Iso-C-based navigation demonstrated a lower mean deviation of 0.5 mm compared to 1 mm with CT-based navigation. Liu et al 7 conducted a prospective cohort study that compared the accuracy of cervical (C2–C7) pedicle screw fixation assisted by
12
ACCEPTED MANUSCRIPT X-ray fluoroscopy, CT Nav, or 3D FluoroNav performed in 82 consecutive patients. The results of CT Nav were poor compared to 3D FluoroNav, with pedicle violation rates of 2.5% and 0%, respectively. Which revealed that conventional CT-based
RI PT
navigation and 3D FluoroNav were similarly accurate. These two techniques are therefore preferable for high-risk cervical pedicle screw fixation. However, due to its
systems in future clinical applications.
SC
ease and convenience, 3D FluoroNav may replace conventional CT-based navigation
M AN U
One drawback of CT Nav is increased radiation exposure due to preoperative CT scanning. However, Slomczykowski et al 14 concluded that the radiation exposure of patients mainly depended on the CT protocol; by determining the radiation dose for pedicle screw insertion, they recommended the spiral mode of CT scanning.
TE D
Nevertheless, that study also revealed that 3D FluoroNav produced a relative reduction in intraoperative radiation dosage compared with CT Nav.
EP
Another drawback of CT Nav is the steep learning curve for the registration procedure, although multilevel registration is more convenient than single-level
AC C
registration. Quick registration requires an intimate knowledge of surgical anatomy and cooperation between the surgeon and navigation system 15. Comparatively, the automatic registration principles of 3D FluoroNav result in easy learning and short operative time. Furthermore, CT Nav is perceived as the most expensive navigation method, and the large volume and weight of the machine itself may increase the difficulty of its clinical application 16. Overall, several benefits of 3D FluoroNav imply that it is a better choice in clinical practice. However, an upgraded version of
13
ACCEPTED MANUSCRIPT intraoperative CT imaging (O-arm-based navigation) was recently introduced that
offers some benefits compared with conventional CT-based navigation. O-arm-based navigation uses CT data acquired intraoperatively from patients and obtains updated
RI PT
data as required. It is not necessary to select data from a 3D model of the spine created by a computer system using preoperative CT. Surgeons are not dependent on preoperative CT, in which the patient’s position may vary from the intraoperative
SC
surgical position. The image quality of the O-arm system is almost comparable to
M AN U
that of the recent multidetector helical CT6. Therefore, further study is required to determine whether 3D FluoroNav will remain superior to O-arm-based navigation.
3D FluoroNav vs 2D FluoroNav
TE D
The 2D FluoroNav system provides assistance by simultaneously displaying the positions of different surgical instruments in flat images. However, 3D FluoroNav provides intraoperative real-time images and may reduce errors, as flat images can
EP
be reconstructed freely in the three-dimensional database. The advantages in
AC C
iconography make 3D FluoroNav more accurate than 2D FluoroNav. Recent work has suggested that lower rates of pedicle screw misplacement occur when 3D FluoroNav is used compared with 2D fluoroscopic image guidance 10; however, another study reported no differences between the 2D and 3D FluoroNav methods in the rate of pedicle screw misplacement 4. A systematic review by Marson et al 12 revealed that 3D fluoroscopy-based image guidance systems were significantly more
14
ACCEPTED MANUSCRIPT
accurate in pedicle screw placement than 2D FluoroNav methods, which agrees with our conclusion.
As shown in Table 2, there was slight heterogeneity in the results of a thorough
RI PT
influence analysis in which we recalled the main heterogeneity from Tian et al 5 and Lekovic et al 11. Clinical homogeneity was considered before setting the
SC
standard of inclusion, although a cadaveric study of human spinal navigation
differs from a clinical trial. The ultimate outcome selected by the researchers was
M AN U
pedicle violation, with no direct analysis of postoperative effects or patient prognosis. The accuracy of the navigation device and the surgeon’s experience were the only two factors that influenced pedicle violation. All the studies exhibited the same influence of spinal soft tissue on navigation; therefore,
TE D
cadaveric studies should be included. It is difficult to determine whether a cadaver has suffered spinal degeneration, spinal deformity or spinal fracture; furthermore,
EP
cadavers show slight differences in basic pathology. These differences may have produced slight bias in the present review, but the exclusion of this study did not
AC C
affect the final results of the influence test. As expected, Lekovic et al found that the only statistically significant predictor of screw breakout was pedicle diameter, as the cervical pedicle diameter was often smaller than the smallest available screw (4.5 mm) 11. Recognizing that screw breakout is still possible, we excluded unreliable data for accurate synthesis. This integration may lead to a slight risk of bias but does not significantly influence final outcomes.
15
ACCEPTED MANUSCRIPT The use of 3D FluoroNav allows the collection of stereoscopic spatial images, an inherent advantage over 2D FluoroNav in the challenge of dangerous cervical pedicle screw placement or minimally invasive percutaneous pedicle screw
RI PT
insertion. An RCT article by Tian W et al evaluating the accuracy of five different techniques for lower cervical pedicle screw placement showed a poor rate of 0% with 3D FluoroNav compared with 12.5% with 2D FluoroNav. That study also
SC
revealed that the Iso-C 3D navigation system significantly improved accuracy
M AN U
compared with virtual fluoroscopy navigation assistance 5. Furthermore, Wood et al published a retrospective comparative study indicating that the accuracy of minimally invasive lumbar pedicle screw placement using an image-guided system with intraoperative 3D fluoroscopy was superior to preoperative CT scan
TE D
images merged with intraoperative 2D fluoroscopy 17.
The radiation dosage of those two methods also concerns spine surgeons. In a
EP
prospective study conducted by Gebhard et al, 2D FluoroNav produced more radiation than did CT Nav. However, 3D FluoroNav had a further reduction of
AC C
intraoperative radiation dosage compared with the other navigation methods 18. However, 3D FluoroNav prolonged surgical time, mainly during the image collection process, by approximately 10 minutes compared with 2D FluoroNav. Nevertheless, the benefits of 3D FluoroNav far offset the increased time in the operating room. According to a case study analysing the surgical times of 248 patients using intraoperative 3D images, the additional time decreased from 16 to
16
ACCEPTED MANUSCRIPT 7 minutes over the course of 2 years as the learning curve changed19. Overall, these results suggest the obvious superiority of 3D FluoroNav.
RI PT
CT Nav vs 2D FluoroNav Although a graphical discrepancy of preoperative scan vs intraoperative collection was observed in CT Nav, high-quality real-time 3D images still provide low
SC
pedicle breach rates. There were three retrospective comparative studies of the accuracy of pedicle screw insertion of CT Nav vs 2D FluoroNav. They all
M AN U
concluded that CT Nav was significantly more accurate than 2D FluoroNav in assisting pedicle screw placement. However, a meta-analysis conducted by Tian NF et al showed that although the risk associated with CT Nav was lower than
TE D
with 2D FluoroNav, the difference was not statistically significant, indicating that 2D FluoroNav has similar potential to CT Nav in pedicle screw insertion assistance. In the present study, our conclusion disagrees with the results of Tian
EP
NF et al.
AC C
There was moderate heterogeneity (I2=31%) between the two navigation groups. Therefore, we designed an influence analysis (Table 2) to identify the source of this moderate heterogeneity. The results mainly showed bias risk from Tian NF et al and Gruetzner et al. We have discussed the bias risk of Tian N F et al; however, the bias risk of Gruetzner et al was considered to mainly result from unadvanced preoperative CT scanning devices, a learning curve or image drift.
17
ACCEPTED MANUSCRIPT The final results were derived from comparisons of any two of the three groups but not from a parallel comparison of CT Nav, 3D FluoroNav and 2D FluoroNav, which is considered a limitation of this review. However, within the present
RI PT
dataset, we clearly demonstrate that 3D fluoroscopic navigation techniques result in higher pedicle screw placement accuracy than conventional CT-based and 2D
fluoroscopic navigation techniques. Another limitation is that we failed to apply a
M AN U
the lack of sufficient studies in each analysis group.
SC
correction for multiple comparisons, such as a funnel plot or L’Abbé plot, due to
In conclusion, among the three navigation methods, cumulative analyses suggest that conventional CT Nav is slightly more accurate than 2D FluoroNav but less accurate than 3D FluoroNav. Ultimately, 3D FluoroNav may be the best choice for
Reference
EP
TE D
surgeons requiring a system to assist pedicle screw insertion.
AC C
1. Lau, D, Terman, SW, Patel, R, et al. 2013. Incidence of and risk factors for superior facet violation in minimally invasive versus open pedicle screw placement during transforaminal lumbar interbody fusion: a comparative analysis.
Journal of neurosurgery. Spine 18: 356-361. 2. Fu, TS, Wong, CB, Tsai, TT, et al. 2008. Pedicle screw insertion: computed tomography versus fluoroscopic image guidance. International orthopaedics 32: 517-521.
18
ACCEPTED MANUSCRIPT 3. Huang, Y, Kong, R, Fang, SY, et al. 2009. Cmoparison study between C-arm X-ray and 3D-CT in guiding thoracolumbar pedicle screw fixation. Shandong Medical Journal.
RI PT
4. Lee, GY, Massicotte, EM, Rampersaud, YR. 2007. Clinical accuracy of cervicothoracic pedicle screw placement: a comparison of the "open"
disorders & techniques 20: 25-32.
SC
lamino-foraminotomy and computer-assisted techniques. Journal of spinal
M AN U
5. Tian, W, Liu, Y, Zheng, S, Lv, Y. 2013. Accuracy of lower cervical pedicle screw placement with assistance of distinct navigation systems: a human cadaveric study. European spine journal : official publication of the European Spine Society, the European Spinal Deformity Society, and the European Section of the Cervical
TE D
Spine Research Society 22: 148-155.
6. Kotani, T, Akazawa, T, Sakuma, T, et al. 2014. Accuracy of Pedicle Screw
EP
Placement in Scoliosis Surgery: A Comparison between Conventional Computed Tomography-Based and O-Arm-Based Navigation Techniques. Asian spine
AC C
journal 8: 331-338.
7. Liu, YJ, Tian, W, Liu, B, et al. 2010. Comparison of the clinical accuracy of cervical (C2-C7) pedicle screw insertion assisted by fluoroscopy, computed tomography-based navigation, and intraoperative three-dimensional C-arm navigation. Chinese medical journal 123: 2995-2998. 8. Nottmeier, EW, Seemer, W, Young, PM. 2009. Placement of thoracolumbar pedicle screws using three-dimensional image guidance: experience in a large
19
ACCEPTED MANUSCRIPT patient cohort. Journal of neurosurgery. Spine 10: 33-39.
9. Gruetzner, P, Wentzensen, A, Waelti, H, et al. 2004. Navigation Using Fluoro-CT Technology: Concept and Clinical Experience in a New Method for Intraoperative
RI PT
Navigation. European Journal of Trauma 30: 161-170. 10. Tian, NF, Huang, QS, Zhou, P, et al. 2011. Pedicle screw insertion accuracy with different assisted methods: a systematic review and meta-analysis of comparative
SC
studies. European spine journal : official publication of the European Spine
M AN U
Society, the European Spinal Deformity Society, and the European Section of the Cervical Spine Research Society 20: 846-859.
11. Lekovic, GP, Potts, EA, Karahalios, DG, Hall, G. 2007. A comparison of two techniques in image-guided thoracic pedicle screw placement: a retrospective
393-398.
TE D
study of 37 patients and 277 pedicle screws. Journal of neurosurgery. Spine 7:
EP
12. Mason, A, Paulsen, R, Babuska, JM, et al. 2014. The accuracy of pedicle screw placement using intraoperative image guidance systems. Journal of neurosurgery.
AC C
Spine 20: 196-203.
13. Geerling, J, Gosling, T, Gosling, A, et al. 2008. Navigated pedicle screw placement: experimental comparison between CT- and 3D fluoroscopy-based techniques. Computer aided surgery : official journal of the International Society for Computer Aided Surgery 13: 157-166.
20
ACCEPTED MANUSCRIPT 14. Slomczykowski, M, Roberto, M, Schneeberger, P, et al. 1999. Radiation dose for pedicle screw insertion. Fluoroscopic method versus computer-assisted surgery. Spine 24:975–982
RI PT
15. Kendo, D, Citak, M, Hüfner, T, et al. 2007. Current concepts and applications of computer navigation in orthopedic trauma surgery. Central Eur J Med 2:392–403 16. Park, P, Foley, KT, Cowan, JA, Marca, FL. 2010. Minimally invasive pedicle
SC
screw fixation utilizing O-arm fluoroscopy with computer-assisted navigation:
M AN U
Feasibility, technique, and preliminary results. Surgical neurology international 1: 44.
17. Wood, M, Mannion, R. 2011. A comparison of CT-based navigation techniques for minimally invasive lumbar pedicle screw placement. Journal of spinal
TE D
disorders & techniques 24: E1-5.
18. Gebhard, FT, Kraus, MD, Schneider, E, et al. 2006. Does computer-assisted
EP
spine surgery reduce intraoperative radiation doses? Spine 31: 2024-2027; discussion 2028.
AC C
19. Kendoff, D, Citak, M, Gardner, MJ, et al. 2009. Intraoperative 3D imaging: value and consequences in 248 cases. The Journal of trauma 66: 232-238.
21
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
Figure legends
TE D
Fig. 1 Flow diagram of the selection of reports for this meta-analysis
EP
Fig. 2 Pedicle screw insertion accuracy: three-dimensional fluoroscopy-based
AC C
navigation (3D FluoroNav) vs computed tomography-based navigation (CT Nav)
Fig. 3 Pedicle screw insertion accuracy: three-dimensional fluoroscopy-based navigation (3D FluoroNav) vs two-dimensional fluoroscopy-based navigation (2D FluoroNav)
Fig. 4 Pedicle screw insertion accuracy: two-dimensional fluoroscopy-based navigation (2D FluoroNav) vs computed tomography-based navigation (CT Nav)
22
ACCEPTED MANUSCRIPT
Fig. 5 Bar graph showing the overall collective accuracy with each of the three
AC C
EP
TE D
M AN U
SC
RI PT
modalities of no breach, 1 mm or less breach, and >1 mm breach
23
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
24
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
ACCEPTED MANUSCRIPT
Description of included study
Study (year)
Study design (class
Participants of
Interventions
(spinal level)
3D FluNAV
evidence) Prospective
58 patients:40 male;18 female; age: 20-75 years
G1:3D FluoroNav n=29
2010
cohort study(II)
abnormal cervical curvature, instability trauma,
(140 screws)
spinal kyphosis, degenerative cervical
G2: CT Nav n=29
disease(C2-C7)
(159 screws)
Forty adult spine cadavers(C1–T1)
G1:3D FluoroNav n=8
2013
controlled
M AN U
Random
SC
Liu et al.
Tian et al.
Pedicle screws misplacement
RI PT
Table 1
2D FluNAV
assessment
Nav 4/159
CT
10/80
0/80
CT
13/98
6/104
CT
0/140
0/80
CT
Accuracy
(80 screws)
cadaveric study(I)
G2: CT Nav n=8 (80 screws)
G3:2D FluoroNav n=8
TE D
(80 screws)
Retrospective
60 patients: sex unstated; age: 21-91 years
G1: CT Nav n=9
2007
cohort study(III)
Cervical spondylotic myelopathy, trauma,
(45 screws)
deformity, infection, OPLL. RA, osteoporosis
G2: 2D FluoroNav n=19
(C7, T1, T2)
(63 screws)
et
al. 2007
Retrospective
37 patients:20 male;27 female; age:35-81 years
G1:3D FluoroNav n=12
study cohort(III)
trauma, degenerative disease, tumor.
(94 screws)
AC C
Lekovic
EP
Lee et al.
5/94
CT
16/183
G2:2D Nav n=25 (183 screws)
Nottmeier et
Retrospective
220 patients: sex and age unstated, spinal
G1:3DFluoroNav n=140
al. 2009
cohort study(III)
disease unstated(T1-S1).
(637 screws) G2: CT Nav n=80 (314 screws)
42/637
29/314
CT
ACCEPTED MANUSCRIPT
Continued
Study (year)
Study design (class
Participants of
RI PT
Table 1
Interventions
(spinal level)
Pedicle screws misplacement
3D
evidence)
Accuracy assessment
2D FluNAV
CT Nav
3/108
5/112
CT
11/222
CT
FluNAV
Retrospective
79 patients: 52male;27 female; age:16-76 years
G1:3D FluoroNav n=24
al. 2004
cohort study(III)
Injuries and degeneration changes to the spine
(114 screws)
(cervical, thoracic, lumbar)
G2: CT Nav n=27
1/114
SC
Gruetzner et
M AN U
(112 screws)
G3:2D FluoroNav n=28 (108 screws)
Kotani et al.
Retrospective
61 patients:11 male;50 female; age:12-31 years
G1:3D FluoroNav n=32
2014
cohort study(III)
scoliosis with a thoracic curve
(416 screws)
13/416
TE D
G2: CT Nav n=29 (222 screws)
Prospective
67 patients:30 male;37 female; age:22-78 years
G1:3D FluoroNav n=43
2011
cohort study(II)
lumbar radiculopathy, instability with
(186 screws)
spondylolisthesis, trauma, tumor.
G2: 3D FluoroNav n=24
EP
Wood et al.
CT
7/110
(110 screws)
Fu et al.
Retrospective
24 patients:9 male;15 female; age:19-79 years
2008
cohort study(III)
Fracture. Spondylolisthesis, tuberculous
(76 screws)
spondylitis. ankylosing spondylitis. revision
G2: 2D FluoroNav n=13
(below T8 Level)
(74 screws)
AC C
3/186
G1: CT Nav n=11
5/74
3/76
CT
ACCEPTED MANUSCRIPT
Retrospective
42 patients: 29 male;13 female; age: 24-64 years
G1: CT Nav n=21
2009
cohort study(III)
Fracture. spondylolisthesis, lumbar disc
(104 screws)
herniation (thoracic, lumbar)
G2: 2D FluoroNav n=21
AC C
EP
TE D
M AN U
SC
(98 screws)
RI PT
Huang et al.
13/98
6/104
CT
ACCEPTED MANUSCRIPT
Influence analyses were conducted by consecutively removing one or all subgroup reviews CI) M-H, Fixed 0.01
0.1
1
10
0.31(0.16-0.61)
3D Fluoro Nav vs. 2D fluoro Nav Consecutively removing following one
Heterogeneity
RI PT
Risk Ratios (95%
Influence analysis
P=0.0008, I²=17%
SC
Table 2
Tian 2013
0.43(0.21-0.89)
Gruetzner 2004
0.31(0.15-0.63)
Lekovic 2007
0.17(0.06-0.47)
wood 2011
0.33(0.17-0.73)
P=0.006, I²=40%
CT Nav vs. 2D Fluoro Nav
0.49(0.29-0.81)
P=0.006, I²=31%
Fu 2008
0.47(0.27-0.82)
P=0.008, I²=49%
Gruetzner 2004
0.4(0.22-0.7)
P=0.002, I²=11%
Huang 2009
0.51(0.27-0.95)
P=0.03, I²=45%
0.44(0.24-0.81)
P=0.009, I²=48%
0.64(0.37-1.1)
P=0.10, I²=0%
P=0.001, I²=45% P=0.0008, I²=0%
M AN U
EP AC C
Tian 2013
TE D
Consecutively removing following one
Lee 2007
P=0.02, I²=0%
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
ACCEPTED MANUSCRIPT Abbreviations list
3D Fluoro Nav: Three-dimension fluoroscopy-based navigation
AC C
EP
TE D
M AN U
SC
CT Nav: Computer tomography-based navigation
RI PT
2D Fluoro Nav: Two-dimension fluoroscopy-based navigation
S1878-8750(17)31254-8
DOI:
10.1016/j.wneu.2017.07.154
Reference:
WNEU 6216
To appear in:
World Neurosurgery
Received Date: 31 March 2017 Revised Date:
24 July 2017
Accepted Date: 26 July 2017
Please cite this article as: Du JP, Fan Y, Wu QN, Wang DH, Zhang J, Hao DJ, Accuracy of Pedicle Screw Insertion among Three Image-Guided Navigation Systems: A Systematic Review and MetaAnalysis, World Neurosurgery (2017), doi: 10.1016/j.wneu.2017.07.154. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
1
ACCEPTED MANUSCRIPT
Accuracy of Pedicle Screw Insertion among Three Image-Guided Navigation Systems: A Systematic Review and Meta-Analysis
RI PT
Jin Peng Du1,4 ▪ Yong Fan1,3 ▪ Qi Ning Wu3 ▪ Dai Hua Wang4 ▪ Jing Zhang2,4 ▪ Ding Jun Hao2,3
These authors contributed equally to this study and should be considered co-first
SC
1
2
The co-first authors work for two principal investigators, who should be considered
co-corresponding authors.
Department of Spine Surgery, Xi'an Jiaotong University-Affiliated Honghui
TE D
3
M AN U
authors.
Hospital, Youyidong Road, Xi’an City 710000, China. Email: [email protected],
Medical College, Yan'an University, No 38 Guanghua Road, Yan’an City 716000,
AC C
4
EP
Tel/Fax: 0086-29-87800002. ORCID: 0000-0002-0946-6080
Shanxi Province, China. Email: [email protected], Tel/Fax: 0086-911-2412308
Author Contributions Statement: J.P.D. and Y.F. extracted and analysed the data. D.J.H. and Q.N.W. conceived of and designed the study. D.H.W., J.Z. and D.J.H. participated in the critical revision of the manuscript. J.P.D. and Y.F. drafted the paper.
2
ACCEPTED MANUSCRIPT All authors have read and approved the final submitted manuscript.
Accuracy of Pedicle Screw Insertion among Three Image-Guided Navigation Systems: A Systematic Review and
RI PT
Meta-Analysis
Abstract
SC
Background: Many retrospective studies of pedicle screw placement have revealed that intraoperative navigation systems provide higher accuracy rates and safety than
M AN U
do free-hand techniques. The accuracy of various image-guided navigation systems has been studied; however, differences have not been well defined due to the lack of adequate evidence-based comparative studies.
Objective: A meta-analysis was conducted to focus on the variation in pedicle screw
TE D
insertion among three navigation systems: a three-dimensional fluoroscopy-based navigation system (3D FluoroNav), a two-dimensional fluoroscopy-based navigation
(CT Nav).
EP
system (2D FluoroNav) and a conventional computed tomography navigation system
AC C
Methods: We screened for comparative studies on different pedicle screw insertion navigation systems published through January 2017 using the Cochrane Library, Ovid, Web of Science, PubMed, and EMBASE databases. Results: From 125 papers that were identified, 10 articles were finally chosen. The present comparative study included 8 retrospective clinical studies, 1 prospective clinical trial and 1 randomized controlled cadaveric study. The prevalence rate of pedicle violation in the 3D FluoroNav group was significantly lower than the rates of
3
ACCEPTED MANUSCRIPT the 2D FluoroNav group (RR 95% CI: 0.16-0.61, P<0.01) and the CT Nav group (RR 95% CI: 0.42-0.90, P=0.01), and the rate of the CT Nav group was significantly lower than that of the 2D FluoroNav group (RR 95% CI: 0.29-0.81, P<0.01).
RI PT
Conclusion: There are significant differences among CT Nav, 3D FluoroNav and 2D FluoroNav. Our review suggests that 3D FluoroNav may be superior to the other two methods in reducing pedicle violation and that clinicians should consider 3D
SC
FluoroNav as a better choice.
M AN U
Keywords image-guided navigation, CT navigation, 3D fluoroscopy-based navigation, 2D fluoroscopy-based navigation, pedicle screw.
Introduction
TE D
The introduction of the pedicle screw technique for the treatment of spinal diseases has shown great progress, and this technique has been widely used by spine
1
EP
surgeons. However, the traditional free-hand technique is prone to pedicle violation . Consequently, various navigation systems have been invented to assist screw
AC C
insertion and reduce the incidence of pedicle violation. Three main intraoperative navigation systems are used to assist in pedicle screw placement: a three-dimensional fluoroscopy-based navigation system (3D FluoroNav) (SIREMOBIL Iso-C, Siemens, Forchheim, Germany), a two-dimensional fluoroscopy-based navigation system (2D FluoroNav) (Vector Vision, Brain LAB GmbH, Germany) and a conventional computed tomography navigation system (CT Nav) (Siemens, Forchheim, Germany). However, it was previously unclear which
4
ACCEPTED MANUSCRIPT image-guided navigation system could provide the highest accuracy. Many researchers have suggested that CT Nav is more accurate 2-5, while other scholars have concluded that 3D FluoroNav 6-8 and 2D FluoroNav 9 are more
RI PT
accurate. The results of the existing studies are inconsistent, and the sample sizes were small. Tian NF et al 10 published a meta-analysis demonstrating that these three navigation systems are significantly more accurate than the free-hand method but
SC
could not draw a definitive conclusion regarding the most accurate navigation
M AN U
system due to insufficient sample sizes and comparative research data. Therefore, the present meta-analysis was conducted to help spine surgeons determine which system should be chosen to reduce pedicle breaches.
TE D
Materials and methods Data sources and searches
EP
The Cochrane Library, Ovid, Web of Science, PubMed and EMBASE databases were searched for studies published from 2002 through January 2017 that focused on
AC C
the accuracy of pedicle screw insertion. We explored abstracts, titles and subtitles by searching for the term pedicle screw combined with the following keywords: computed tomography-based/CT navigation , two-dimensional/2D fluoroscopy-based navigation , three-dimensional/3D fluoroscopy-based navigation , and accuracy . The resulting abstracts were inspected to identify articles that should be excluded. Subsequently, the remaining full-text articles were reviewed to select publications that were suitable for inclusion. The study selection process
5
ACCEPTED MANUSCRIPT conformed to the Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) guidelines. The following inclusion criteria were established: (1) publications written in English
RI PT
and (2) articles involving pedicle screw insertion with different types of navigation assistance. The exclusion criteria were as follows: (1) articles not involving a
comparative study, (2) duplicate publications, (3) articles involving pedicle screw
SC
insertion using an image-guided navigation system compared with other methods,
M AN U
and (4) articles with no postoperative CT scanning accuracy assessment. Data extraction
Two reviewers (D.J.P. and F.Y.) independently screened and extracted data by applying the same standards; discrepancies were resolved through discussion until a
TE D
consensus was reached. The following basic information was extracted: first author's name, year of publication, study design, patient characteristics, sample sizes,
EP
interventions, assessment tool and key results. Studies using uniform measurement standards were extracted for further analysis. Specifically, the accuracy of the screw
AC C
in piercing the pedicle, as determined using postoperative CT scanning, was classified into three categories 7: (1) excellent (screw completely within pedicle), (2) acceptable (a portion of the screw outside the pedicle ≤1 mm), and (3) poor (a portion of the screw outside the pedicle >1 mm). The poor grade was considered unacceptable pedicle violation in the analysis.
Data synthesis and statistical analysis
6
ACCEPTED MANUSCRIPT Risk ratios (RRs) with 95% confidence intervals (CIs) were summarized for pedicle violation. The level of significance was set at P<0.05. A forest plot was generated to compare the violation of pedicle screw placement between the experimental and
RI PT
control groups, the point estimate on the left of the vertical line means that the experimental group was significantly more accurate than the control group.
Heterogeneity among pedicle violation in various systems was assessed using the χ2
SC
test. For P>0.1 or I2<50%, no clear heterogeneity of results was assumed.
M AN U
Heterogeneity stems from a variety of different variations among several studies in systematic reviews, which may affect the synergistic effects of several studies and the explanation of the results of the meta-analysis; a strict evaluation is generally required. Although it is impossible to completely prevent, it can be controlled by
TE D
including high-quality studies that all have the same study objective. We selected a fixed-effect model for all analyses unless there was significant heterogeneity. To
EP
examine the stability of the outcomes and identify important sources of slight heterogeneity, influence analyses were conducted by consecutively removing one or
AC C
all subgroup reviews. Statistical analyses were performed using Review Manager Version 5.0 (Cochrane Collaboration, Software Update, Oxford, UK).
Results Literature search The primary search strategy resulted in 580 articles, of which 455 were excluded due to having an irrelevant title or abstract. The remaining 125 papers underwent
7
ACCEPTED MANUSCRIPT full-text evaluation, and 61 were excluded for not meeting the inclusion criteria. Of the remaining papers, 54 were removed per the exclusion criteria. Ultimately, 10 studies were included in this meta-analysis; no additional studies were identified
RI PT
from the references of published results. The present meta-analysis included 1 randomized controlled cadaveric trial5, 2 prospective cohort studies7,17 and 7
retrospective cohort studies2-4,6-9,11. A flow diagram of the report selection is shown
SC
in Fig. 1.
M AN U
Basic characteristics
Detailed characteristics of the involved patients are presented in Table 1. In these trials, the sample size ranged from 24 to 220, including a total of 648 patients and 40 adult cadavers. Altogether, approximately 3255 pedicle screws were inserted in
TE D
patients with various spinal diseases requiring stability reconstruction, and 240 screws were inserted in cadavers. CT scanning was performed for the postoperative
EP
assessment of pedicle violation in all trials. Comparison of pedicle violation
AC C
3D FluoroNav vs CT Nav
As shown in Fig. 2, the first comparative group included 1387 screws inserted using 3D FluoroNav and 887 screws using CT Nav. The 3D FluoroNav group was significantly more accurate (RR 95% CI: 0.42-0.90, P=0.01) than the CT Nav group in guiding pedicle screw fixation. There was no significant heterogeneity (χ2=2.66, P=0.45, I2=0%) between the two groups; therefore, it was not necessary to identify sources of heterogeneity. In addition, Tian et al 5 found the pedicle violation rates of
8
ACCEPTED MANUSCRIPT the two groups to be equal to 0%; therefore, the RR could not be estimated (as shown in Fig. 2), which may have introduced little bias to the results. 3D FluoroNav vs 2D FluoroNav
RI PT
The second comparative group included 474 screws inserted using 3D FluoroNav and 481 screws using 2D FluoroNav. The 3D FluoroNav group was significantly
more accurate (RR 95% CI: 0.16-0.61, P<0.01) than the 2D FluoroNav group in
SC
aiding pedicle screw placement (Fig. 3). There was slight heterogeneity (χ2=3.63,
M AN U
P=0.30, I2=17%) between the two groups. The influence analysis results shown in Table 2 revealed that when Tian et al 5 was removed, there was no significant change in the positive results; the 3D FluoroNav group was still more accurate (RR 95% CI: 0.21-0.89, P=0.02) than the 2D FluoroNav group. However, the slight
TE D
heterogeneity disappeared (χ2=1.17, P=0.56, I2=0%) with this exclusion. When Lekovic et al 11 was excluded, no significant change in the positive results was
EP
shown, and the 3D FluoroNav group was still more accurate (RR 95% CI: 0.06-0.47, P<0.01) than the 2D FluoroNav group; however, the slight heterogeneity
AC C
vanished (χ2=1.46, P=0.48, I2=0%). CT Nav vs 2D FluoroNav The third comparative group included 417 screws inserted using CT Nav and 423 screws using 2D FluoroNav. The CT Nav group was significantly more accurate (RR 95% CI: 0.29-0.81, P<0.01) than the 2D FluoroNav group in assisting pedicle screw placement (Fig. 4). There was slight heterogeneity (χ2=5.79, P=0.22, I2=31%) between the two groups. When Gruetzner et al was rejected, no remarkable variation
9
ACCEPTED MANUSCRIPT in the final outcome was found (Table 2). The CT Nav group was still more accurate (RR 95% CI: 0.22-0.70, P=0.002) than the 2D FluoroNav group, and the heterogeneity decreased (χ2=3.39, P=0.34, I2=11%) with this exclusion. In contrast,
RI PT
the exclusion of Tian et al resulted in extreme alteration (RR 95% CI: 0.37-1.10, P=0.10), and the heterogeneity decreased greatly (χ2=2.33, P=0.51, I2=0%).
SC
Discussion
M AN U
Pedicle screws are widely applied to treat conditions such as spinal degenerative disease, fracture, and deformity, as three-column fixations provide the most rigid form of posterior stabilization. However, pedicle screw malposition remains an inevitable problem. A variety of navigation systems have been introduced in
TE D
response to the desire of spine surgeons to reduce pedicle violation. There are three important, clinically used image-guided navigation methods based on intraoperative
EP
guidance theory: CT Nav, 3D FluoroNav and 2D FluoroNav. The first step in CT Nav is preoperative CT scanning to target the spine level; subsequently, the captured
AC C
preoperative anatomical information is transferred to a navigation platform via matching and registration to assist pedicle screw insertion 12. In contrast to CT Nav, 2D FluoroNav depends on intraoperative two-dimensional images and provides real-time visualization of pedicle anatomy without requiring a complicated registration procedure. With the development of 3D FluoroNav, image-guided navigation systems are rapidly developing and increasing in popularity in the medical field. Because 3D FluoroNav combines the advantages of 2D FluoroNav
10
ACCEPTED MANUSCRIPT and CT Nav without complex matching and single-level or multi-level registration 10
, it provides real-time three-dimensional images, and its quality is comparable to
Meta-analysis of comparisons of pedicle screw accuracy
RI PT
that of CT Nav.
Tian NF et al conducted a systemic review and meta-analysis comparing three
SC
image-guided navigation systems to assist with pedicle screw insertion. These
M AN U
navigation systems provided higher accuracy than conventional methods. The superiority of the navigation systems was apparent when they were applied to abnormal spinal structure10. This study suggested that the significant differences between the three systems are likely due to the lack of comparative studies. In our
TE D
review, a meta-analysis was conducted by identifying additional studies to determine whether there are differences among the systems and which system is the most
EP
accurate.
In the present analysis, the pooled RR values indicate that CT Nav is slightly more
AC C
accurate than 2D FluoroNav (CT Nav vs 2D FluoroNav: RR=0.49, P<0.05) but less accurate than 3D FluoroNav (3D FluoroNav vs 2D FluoroNav: RR=0.31, P<0.05). The bar graph in Fig. 5 shows the overall collective accuracy of each of the three modalities, with pedicle violation classified as excellent (0 mm), acceptable (≤1 mm) or poor (>1 mm). In conclusion, there are significant differences among the three navigation systems, and our study suggests that the 3D FluoroNav system may
11
ACCEPTED MANUSCRIPT be superior in reducing pedicle violation. This conclusion agrees with the in vitro results of Tian NF et al but not with their in vivo results.
RI PT
CT Nav vs 3D FluoroNav
Both CT Nav and 3D FluoroNav can increase the accuracy of screw fixation
SC
compared with the traditional free-hand method. Although CT Nav provides
M AN U
high-quality, real-time intraoperative images, there are some apparent defects. The actual 3D spatial relations during surgery may differ from the CT data obtained preoperatively if the position of the patient changes, possibly leading to errors 7. Moreover, the process of point-to-point registration prolongs the operation time and
TE D
reduces accuracy. However, 3D FluoroNav automatically allows for the collection of 3D-reconstructed images and then transfers them to the navigation system.
EP
Automatic registration and matching can be conducted, as with virtual fluoroscopy navigation. Moreover, 3D FluoroNav combines the advantages of CT Nav and 2D
AC C
FluoroNav 2,8-9,11 and reduces extra preoperative preparation. Geerling et al 13 used a foam spine model to compare the accuracy of CT-based and Iso-C-based navigation. The measured accuracy of the markers was significantly different between CT-based and Iso-C-based navigation for one of six markers placed on the vertebra. Iso-C-based navigation demonstrated a lower mean deviation of 0.5 mm compared to 1 mm with CT-based navigation. Liu et al 7 conducted a prospective cohort study that compared the accuracy of cervical (C2–C7) pedicle screw fixation assisted by
12
ACCEPTED MANUSCRIPT X-ray fluoroscopy, CT Nav, or 3D FluoroNav performed in 82 consecutive patients. The results of CT Nav were poor compared to 3D FluoroNav, with pedicle violation rates of 2.5% and 0%, respectively. Which revealed that conventional CT-based
RI PT
navigation and 3D FluoroNav were similarly accurate. These two techniques are therefore preferable for high-risk cervical pedicle screw fixation. However, due to its
systems in future clinical applications.
SC
ease and convenience, 3D FluoroNav may replace conventional CT-based navigation
M AN U
One drawback of CT Nav is increased radiation exposure due to preoperative CT scanning. However, Slomczykowski et al 14 concluded that the radiation exposure of patients mainly depended on the CT protocol; by determining the radiation dose for pedicle screw insertion, they recommended the spiral mode of CT scanning.
TE D
Nevertheless, that study also revealed that 3D FluoroNav produced a relative reduction in intraoperative radiation dosage compared with CT Nav.
EP
Another drawback of CT Nav is the steep learning curve for the registration procedure, although multilevel registration is more convenient than single-level
AC C
registration. Quick registration requires an intimate knowledge of surgical anatomy and cooperation between the surgeon and navigation system 15. Comparatively, the automatic registration principles of 3D FluoroNav result in easy learning and short operative time. Furthermore, CT Nav is perceived as the most expensive navigation method, and the large volume and weight of the machine itself may increase the difficulty of its clinical application 16. Overall, several benefits of 3D FluoroNav imply that it is a better choice in clinical practice. However, an upgraded version of
13
ACCEPTED MANUSCRIPT intraoperative CT imaging (O-arm-based navigation) was recently introduced that
offers some benefits compared with conventional CT-based navigation. O-arm-based navigation uses CT data acquired intraoperatively from patients and obtains updated
RI PT
data as required. It is not necessary to select data from a 3D model of the spine created by a computer system using preoperative CT. Surgeons are not dependent on preoperative CT, in which the patient’s position may vary from the intraoperative
SC
surgical position. The image quality of the O-arm system is almost comparable to
M AN U
that of the recent multidetector helical CT6. Therefore, further study is required to determine whether 3D FluoroNav will remain superior to O-arm-based navigation.
3D FluoroNav vs 2D FluoroNav
TE D
The 2D FluoroNav system provides assistance by simultaneously displaying the positions of different surgical instruments in flat images. However, 3D FluoroNav provides intraoperative real-time images and may reduce errors, as flat images can
EP
be reconstructed freely in the three-dimensional database. The advantages in
AC C
iconography make 3D FluoroNav more accurate than 2D FluoroNav. Recent work has suggested that lower rates of pedicle screw misplacement occur when 3D FluoroNav is used compared with 2D fluoroscopic image guidance 10; however, another study reported no differences between the 2D and 3D FluoroNav methods in the rate of pedicle screw misplacement 4. A systematic review by Marson et al 12 revealed that 3D fluoroscopy-based image guidance systems were significantly more
14
ACCEPTED MANUSCRIPT
accurate in pedicle screw placement than 2D FluoroNav methods, which agrees with our conclusion.
As shown in Table 2, there was slight heterogeneity in the results of a thorough
RI PT
influence analysis in which we recalled the main heterogeneity from Tian et al 5 and Lekovic et al 11. Clinical homogeneity was considered before setting the
SC
standard of inclusion, although a cadaveric study of human spinal navigation
differs from a clinical trial. The ultimate outcome selected by the researchers was
M AN U
pedicle violation, with no direct analysis of postoperative effects or patient prognosis. The accuracy of the navigation device and the surgeon’s experience were the only two factors that influenced pedicle violation. All the studies exhibited the same influence of spinal soft tissue on navigation; therefore,
TE D
cadaveric studies should be included. It is difficult to determine whether a cadaver has suffered spinal degeneration, spinal deformity or spinal fracture; furthermore,
EP
cadavers show slight differences in basic pathology. These differences may have produced slight bias in the present review, but the exclusion of this study did not
AC C
affect the final results of the influence test. As expected, Lekovic et al found that the only statistically significant predictor of screw breakout was pedicle diameter, as the cervical pedicle diameter was often smaller than the smallest available screw (4.5 mm) 11. Recognizing that screw breakout is still possible, we excluded unreliable data for accurate synthesis. This integration may lead to a slight risk of bias but does not significantly influence final outcomes.
15
ACCEPTED MANUSCRIPT The use of 3D FluoroNav allows the collection of stereoscopic spatial images, an inherent advantage over 2D FluoroNav in the challenge of dangerous cervical pedicle screw placement or minimally invasive percutaneous pedicle screw
RI PT
insertion. An RCT article by Tian W et al evaluating the accuracy of five different techniques for lower cervical pedicle screw placement showed a poor rate of 0% with 3D FluoroNav compared with 12.5% with 2D FluoroNav. That study also
SC
revealed that the Iso-C 3D navigation system significantly improved accuracy
M AN U
compared with virtual fluoroscopy navigation assistance 5. Furthermore, Wood et al published a retrospective comparative study indicating that the accuracy of minimally invasive lumbar pedicle screw placement using an image-guided system with intraoperative 3D fluoroscopy was superior to preoperative CT scan
TE D
images merged with intraoperative 2D fluoroscopy 17.
The radiation dosage of those two methods also concerns spine surgeons. In a
EP
prospective study conducted by Gebhard et al, 2D FluoroNav produced more radiation than did CT Nav. However, 3D FluoroNav had a further reduction of
AC C
intraoperative radiation dosage compared with the other navigation methods 18. However, 3D FluoroNav prolonged surgical time, mainly during the image collection process, by approximately 10 minutes compared with 2D FluoroNav. Nevertheless, the benefits of 3D FluoroNav far offset the increased time in the operating room. According to a case study analysing the surgical times of 248 patients using intraoperative 3D images, the additional time decreased from 16 to
16
ACCEPTED MANUSCRIPT 7 minutes over the course of 2 years as the learning curve changed19. Overall, these results suggest the obvious superiority of 3D FluoroNav.
RI PT
CT Nav vs 2D FluoroNav Although a graphical discrepancy of preoperative scan vs intraoperative collection was observed in CT Nav, high-quality real-time 3D images still provide low
SC
pedicle breach rates. There were three retrospective comparative studies of the accuracy of pedicle screw insertion of CT Nav vs 2D FluoroNav. They all
M AN U
concluded that CT Nav was significantly more accurate than 2D FluoroNav in assisting pedicle screw placement. However, a meta-analysis conducted by Tian NF et al showed that although the risk associated with CT Nav was lower than
TE D
with 2D FluoroNav, the difference was not statistically significant, indicating that 2D FluoroNav has similar potential to CT Nav in pedicle screw insertion assistance. In the present study, our conclusion disagrees with the results of Tian
EP
NF et al.
AC C
There was moderate heterogeneity (I2=31%) between the two navigation groups. Therefore, we designed an influence analysis (Table 2) to identify the source of this moderate heterogeneity. The results mainly showed bias risk from Tian NF et al and Gruetzner et al. We have discussed the bias risk of Tian N F et al; however, the bias risk of Gruetzner et al was considered to mainly result from unadvanced preoperative CT scanning devices, a learning curve or image drift.
17
ACCEPTED MANUSCRIPT The final results were derived from comparisons of any two of the three groups but not from a parallel comparison of CT Nav, 3D FluoroNav and 2D FluoroNav, which is considered a limitation of this review. However, within the present
RI PT
dataset, we clearly demonstrate that 3D fluoroscopic navigation techniques result in higher pedicle screw placement accuracy than conventional CT-based and 2D
fluoroscopic navigation techniques. Another limitation is that we failed to apply a
M AN U
the lack of sufficient studies in each analysis group.
SC
correction for multiple comparisons, such as a funnel plot or L’Abbé plot, due to
In conclusion, among the three navigation methods, cumulative analyses suggest that conventional CT Nav is slightly more accurate than 2D FluoroNav but less accurate than 3D FluoroNav. Ultimately, 3D FluoroNav may be the best choice for
Reference
EP
TE D
surgeons requiring a system to assist pedicle screw insertion.
AC C
1. Lau, D, Terman, SW, Patel, R, et al. 2013. Incidence of and risk factors for superior facet violation in minimally invasive versus open pedicle screw placement during transforaminal lumbar interbody fusion: a comparative analysis.
Journal of neurosurgery. Spine 18: 356-361. 2. Fu, TS, Wong, CB, Tsai, TT, et al. 2008. Pedicle screw insertion: computed tomography versus fluoroscopic image guidance. International orthopaedics 32: 517-521.
18
ACCEPTED MANUSCRIPT 3. Huang, Y, Kong, R, Fang, SY, et al. 2009. Cmoparison study between C-arm X-ray and 3D-CT in guiding thoracolumbar pedicle screw fixation. Shandong Medical Journal.
RI PT
4. Lee, GY, Massicotte, EM, Rampersaud, YR. 2007. Clinical accuracy of cervicothoracic pedicle screw placement: a comparison of the "open"
disorders & techniques 20: 25-32.
SC
lamino-foraminotomy and computer-assisted techniques. Journal of spinal
M AN U
5. Tian, W, Liu, Y, Zheng, S, Lv, Y. 2013. Accuracy of lower cervical pedicle screw placement with assistance of distinct navigation systems: a human cadaveric study. European spine journal : official publication of the European Spine Society, the European Spinal Deformity Society, and the European Section of the Cervical
TE D
Spine Research Society 22: 148-155.
6. Kotani, T, Akazawa, T, Sakuma, T, et al. 2014. Accuracy of Pedicle Screw
EP
Placement in Scoliosis Surgery: A Comparison between Conventional Computed Tomography-Based and O-Arm-Based Navigation Techniques. Asian spine
AC C
journal 8: 331-338.
7. Liu, YJ, Tian, W, Liu, B, et al. 2010. Comparison of the clinical accuracy of cervical (C2-C7) pedicle screw insertion assisted by fluoroscopy, computed tomography-based navigation, and intraoperative three-dimensional C-arm navigation. Chinese medical journal 123: 2995-2998. 8. Nottmeier, EW, Seemer, W, Young, PM. 2009. Placement of thoracolumbar pedicle screws using three-dimensional image guidance: experience in a large
19
ACCEPTED MANUSCRIPT patient cohort. Journal of neurosurgery. Spine 10: 33-39.
9. Gruetzner, P, Wentzensen, A, Waelti, H, et al. 2004. Navigation Using Fluoro-CT Technology: Concept and Clinical Experience in a New Method for Intraoperative
RI PT
Navigation. European Journal of Trauma 30: 161-170. 10. Tian, NF, Huang, QS, Zhou, P, et al. 2011. Pedicle screw insertion accuracy with different assisted methods: a systematic review and meta-analysis of comparative
SC
studies. European spine journal : official publication of the European Spine
M AN U
Society, the European Spinal Deformity Society, and the European Section of the Cervical Spine Research Society 20: 846-859.
11. Lekovic, GP, Potts, EA, Karahalios, DG, Hall, G. 2007. A comparison of two techniques in image-guided thoracic pedicle screw placement: a retrospective
393-398.
TE D
study of 37 patients and 277 pedicle screws. Journal of neurosurgery. Spine 7:
EP
12. Mason, A, Paulsen, R, Babuska, JM, et al. 2014. The accuracy of pedicle screw placement using intraoperative image guidance systems. Journal of neurosurgery.
AC C
Spine 20: 196-203.
13. Geerling, J, Gosling, T, Gosling, A, et al. 2008. Navigated pedicle screw placement: experimental comparison between CT- and 3D fluoroscopy-based techniques. Computer aided surgery : official journal of the International Society for Computer Aided Surgery 13: 157-166.
20
ACCEPTED MANUSCRIPT 14. Slomczykowski, M, Roberto, M, Schneeberger, P, et al. 1999. Radiation dose for pedicle screw insertion. Fluoroscopic method versus computer-assisted surgery. Spine 24:975–982
RI PT
15. Kendo, D, Citak, M, Hüfner, T, et al. 2007. Current concepts and applications of computer navigation in orthopedic trauma surgery. Central Eur J Med 2:392–403 16. Park, P, Foley, KT, Cowan, JA, Marca, FL. 2010. Minimally invasive pedicle
SC
screw fixation utilizing O-arm fluoroscopy with computer-assisted navigation:
M AN U
Feasibility, technique, and preliminary results. Surgical neurology international 1: 44.
17. Wood, M, Mannion, R. 2011. A comparison of CT-based navigation techniques for minimally invasive lumbar pedicle screw placement. Journal of spinal
TE D
disorders & techniques 24: E1-5.
18. Gebhard, FT, Kraus, MD, Schneider, E, et al. 2006. Does computer-assisted
EP
spine surgery reduce intraoperative radiation doses? Spine 31: 2024-2027; discussion 2028.
AC C
19. Kendoff, D, Citak, M, Gardner, MJ, et al. 2009. Intraoperative 3D imaging: value and consequences in 248 cases. The Journal of trauma 66: 232-238.
21
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
Figure legends
TE D
Fig. 1 Flow diagram of the selection of reports for this meta-analysis
EP
Fig. 2 Pedicle screw insertion accuracy: three-dimensional fluoroscopy-based
AC C
navigation (3D FluoroNav) vs computed tomography-based navigation (CT Nav)
Fig. 3 Pedicle screw insertion accuracy: three-dimensional fluoroscopy-based navigation (3D FluoroNav) vs two-dimensional fluoroscopy-based navigation (2D FluoroNav)
Fig. 4 Pedicle screw insertion accuracy: two-dimensional fluoroscopy-based navigation (2D FluoroNav) vs computed tomography-based navigation (CT Nav)
22
ACCEPTED MANUSCRIPT
Fig. 5 Bar graph showing the overall collective accuracy with each of the three
AC C
EP
TE D
M AN U
SC
RI PT
modalities of no breach, 1 mm or less breach, and >1 mm breach
23
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
24
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
ACCEPTED MANUSCRIPT
Description of included study
Study (year)
Study design (class
Participants of
Interventions
(spinal level)
3D FluNAV
evidence) Prospective
58 patients:40 male;18 female; age: 20-75 years
G1:3D FluoroNav n=29
2010
cohort study(II)
abnormal cervical curvature, instability trauma,
(140 screws)
spinal kyphosis, degenerative cervical
G2: CT Nav n=29
disease(C2-C7)
(159 screws)
Forty adult spine cadavers(C1–T1)
G1:3D FluoroNav n=8
2013
controlled
M AN U
Random
SC
Liu et al.
Tian et al.
Pedicle screws misplacement
RI PT
Table 1
2D FluNAV
assessment
Nav 4/159
CT
10/80
0/80
CT
13/98
6/104
CT
0/140
0/80
CT
Accuracy
(80 screws)
cadaveric study(I)
G2: CT Nav n=8 (80 screws)
G3:2D FluoroNav n=8
TE D
(80 screws)
Retrospective
60 patients: sex unstated; age: 21-91 years
G1: CT Nav n=9
2007
cohort study(III)
Cervical spondylotic myelopathy, trauma,
(45 screws)
deformity, infection, OPLL. RA, osteoporosis
G2: 2D FluoroNav n=19
(C7, T1, T2)
(63 screws)
et
al. 2007
Retrospective
37 patients:20 male;27 female; age:35-81 years
G1:3D FluoroNav n=12
study cohort(III)
trauma, degenerative disease, tumor.
(94 screws)
AC C
Lekovic
EP
Lee et al.
5/94
CT
16/183
G2:2D Nav n=25 (183 screws)
Nottmeier et
Retrospective
220 patients: sex and age unstated, spinal
G1:3DFluoroNav n=140
al. 2009
cohort study(III)
disease unstated(T1-S1).
(637 screws) G2: CT Nav n=80 (314 screws)
42/637
29/314
CT
ACCEPTED MANUSCRIPT
Continued
Study (year)
Study design (class
Participants of
RI PT
Table 1
Interventions
(spinal level)
Pedicle screws misplacement
3D
evidence)
Accuracy assessment
2D FluNAV
CT Nav
3/108
5/112
CT
11/222
CT
FluNAV
Retrospective
79 patients: 52male;27 female; age:16-76 years
G1:3D FluoroNav n=24
al. 2004
cohort study(III)
Injuries and degeneration changes to the spine
(114 screws)
(cervical, thoracic, lumbar)
G2: CT Nav n=27
1/114
SC
Gruetzner et
M AN U
(112 screws)
G3:2D FluoroNav n=28 (108 screws)
Kotani et al.
Retrospective
61 patients:11 male;50 female; age:12-31 years
G1:3D FluoroNav n=32
2014
cohort study(III)
scoliosis with a thoracic curve
(416 screws)
13/416
TE D
G2: CT Nav n=29 (222 screws)
Prospective
67 patients:30 male;37 female; age:22-78 years
G1:3D FluoroNav n=43
2011
cohort study(II)
lumbar radiculopathy, instability with
(186 screws)
spondylolisthesis, trauma, tumor.
G2: 3D FluoroNav n=24
EP
Wood et al.
CT
7/110
(110 screws)
Fu et al.
Retrospective
24 patients:9 male;15 female; age:19-79 years
2008
cohort study(III)
Fracture. Spondylolisthesis, tuberculous
(76 screws)
spondylitis. ankylosing spondylitis. revision
G2: 2D FluoroNav n=13
(below T8 Level)
(74 screws)
AC C
3/186
G1: CT Nav n=11
5/74
3/76
CT
ACCEPTED MANUSCRIPT
Retrospective
42 patients: 29 male;13 female; age: 24-64 years
G1: CT Nav n=21
2009
cohort study(III)
Fracture. spondylolisthesis, lumbar disc
(104 screws)
herniation (thoracic, lumbar)
G2: 2D FluoroNav n=21
AC C
EP
TE D
M AN U
SC
(98 screws)
RI PT
Huang et al.
13/98
6/104
CT
ACCEPTED MANUSCRIPT
Influence analyses were conducted by consecutively removing one or all subgroup reviews CI) M-H, Fixed 0.01
0.1
1
10
0.31(0.16-0.61)
3D Fluoro Nav vs. 2D fluoro Nav Consecutively removing following one
Heterogeneity
RI PT
Risk Ratios (95%
Influence analysis
P=0.0008, I²=17%
SC
Table 2
Tian 2013
0.43(0.21-0.89)
Gruetzner 2004
0.31(0.15-0.63)
Lekovic 2007
0.17(0.06-0.47)
wood 2011
0.33(0.17-0.73)
P=0.006, I²=40%
CT Nav vs. 2D Fluoro Nav
0.49(0.29-0.81)
P=0.006, I²=31%
Fu 2008
0.47(0.27-0.82)
P=0.008, I²=49%
Gruetzner 2004
0.4(0.22-0.7)
P=0.002, I²=11%
Huang 2009
0.51(0.27-0.95)
P=0.03, I²=45%
0.44(0.24-0.81)
P=0.009, I²=48%
0.64(0.37-1.1)
P=0.10, I²=0%
P=0.001, I²=45% P=0.0008, I²=0%
M AN U
EP AC C
Tian 2013
TE D
Consecutively removing following one
Lee 2007
P=0.02, I²=0%
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
ACCEPTED MANUSCRIPT Abbreviations list
3D Fluoro Nav: Three-dimension fluoroscopy-based navigation
AC C
EP
TE D
M AN U
SC
CT Nav: Computer tomography-based navigation
RI PT
2D Fluoro Nav: Two-dimension fluoroscopy-based navigation