Secondary Logo

Journal Logo


Retrospective Review of Revision Surgery After Image-guided Instrumented Spinal Surgery Compared With Traditional Instrumented Spinal Surgery

Towner, James E. MD*; Li, Yan Icy PhD*; Singla, Amit MD; Moquin, Ross MD†,‡; Li, Yan Michael MD, PhD*

Author Information
doi: 10.1097/BSD.0000000000000949
  • Free


A diverse range of pathologies involves either acute or chronic compromise of the stability of the spine, requiring internal fixation. Optimally placed instrumentation should provide appropriate biomechanical support and fixation without causing neurovascular compromise. Although some malpositioned screws might not necessitate future revision surgery, they can result in adverse outcomes, either acutely with neural element impingement or vascular injury, or chronically in the form of decreased construct integrity.1 Depending on the level of vertebral bodies being instrumented, injuries to neural elements, including nerve roots and spinal cord, to vascular structures, such as the aorta or vertebral and iliac vessels, and to the lungs and pleural cavity could result from malpositioned screws (Fig. 1). Such complications from malpositioned screws can have devastating outcomes, add to the morbidity of the operation, and contribute to the economic burden associated with their management and treatment, including the need for a subsequent screw revision surgery.

Computed tomography scans showing malpositioned screws. The tip of a right lumbar pedicle screw is abutting the right common iliac artery in (A). A left thoracic pedicle screw is breaching the lateral wall of the pedicle abutting the aorta (B). A right thoracic screw is breaching into the pleural cavity abutting the right lung (C). A left thoracic screw is breaching the medial wall entering the spinal canal (D).

There are many adjuncts to the anatomic knowledge required for instrumentation, and there has been increasing adoption of a variety of technological tools to assist in the placement of spinal instrumentation. One commonly employed intraoperative modality in instrumented spine surgery is fluoroscopy. However, the use of intraoperative fluoroscopy is still associated with the risk of screw malpositioning.2,3 Several systems exist which offer image-guided navigation for instrumented spinal surgery. The O-Arm (Medtronic Inc., Louisville, CO) and Brainlab Spine Navigation (BrainLAB AG, Munich, Germany) are 2 such platforms. The O-Arm provides 3-dimensional (3D) fluoroscopic images that can be reconstructed in a multiplanar mode. Moreover, the images obtained by the O-Arm can be easily integrated with the StealthStation TREON (Medtronic Inc.) system for intraoperative navigation. The Brainlab platform allows for computed tomography (CT) navigation or 2-dimensional (2D) fluoroscopic navigation.

It is essential to understand the benefits and limitations of image-guided navigation as it becomes further integrated into daily practice and its clinical domain expands to incorporate adjuncts, such as image-guided tools or robotics.4,5 Although there are many previous studies comparing the accuracy of screws placed in different manners, there is a relative dearth of information regarding important clinical outcomes from misplaced screws, that is, what is the change in the screw revision rate between different methods.6 We sought to create a practical study, evaluating revision rates for all posteriorly placed instrumentation at our institutions for all pathologies to illustrate the collective experience of using this technology in a variety of spinal regions for a variety of pathologies. We evaluated the role of image guidance technology in reducing or potentially eliminating the need for subsequent screw revision surgeries, by comparing the rates of revision surgeries performed at The University of Rochester Medical Center (URMC) and State University of New York Upstate Medical University (SUNY Upstate) for malpositioned screws placed via a fluoroscopy-assisted or freehand technique compared with the use of image-guided navigation.


After obtaining approval for an exemption from the Institutional Review Board (IRB) for the Protection of Human Subjects, we performed a retrospective review of charts and radiographic data. The clinical data from SUNY Upstate was collected from September 2006 to March 2009 and from URMC was collected from July 2014 to January 2019. Only patients who were operated on by the senior authors and underwent spinal instrumentation procedures via a posterior approach, including lateral mass, pedicle, and iliac screws, were included in this review. Screws placed in an open and percutaneous manner were included. Patients who had hybrid procedures with both non-navigated and image-guided screw placement were included, with individual screws considered in their respective categories. Screws that had suboptimal positioning causing no symptoms or complications postoperatively were not included in estimating the risk of screw malposition.

A 2-tailed Fisher exact test was utilized to compare the rates of revised screws. Statistical analyses were performed using GraphPad QuickCalcs (; accessed January 2019). Statistical significance was set a priori at an α of 0.05.


From SUNY Upstate, 259 patients (133 males and 126 females) were analyzed, for a total of 3313 screws. From URMC, 404 patients (217 males and 187 females) were analyzed, for a total of 2492 screws. The absolute number of operations investigated was 663, however, the number of instances of instrumentation placement analyzed was 690. This discrepancy arises from the fact that 27 operations were performed as a hybrid operation; 3 instances involved use of O-Arm navigation and 24 instances involved Brainlab navigation. The method of screw placement was determined by the operative notes. Navigation was used in 271 procedures (40.87% of patients, 36.45% of screws), with 110 utilizing O-Arm navigation and 161 utilizing Brainlab navigation (Table 1). The age range of all groups is shown in Table 1. Among the 663 patients, there were 4 pediatric patients (<18 y old), 2 males (14 and 16 y old) and 2 females (13 and 16 y old), who had a total of 32 posteriorly placed screws.

Descriptive Analysis of Study Population, Screw Placement Modality, and Indications for Surgery

The underlying pathology encountered in our series was variable and included traumatic and nontraumatic causes. Degenerative spine disease was the most common underlying pathology for both males and females. Although there were more traumatic injuries in males than females, the distribution of degenerative and deformity-related pathology was comparable between males and females. Most of the screws (67.02%) were placed in the thoracolumbar region. Of the 5805 total screws implanted, 2116 (36.45%) were placed using navigation, with 1115 utilizing O-arm assistance and 1001 placed with Brainlab assistance (Table 2).

Screw Placement by Modality and Location

Of the 419 patients with 3689 screws placed without navigation assistance, 10 patients (5 male and 5 female) underwent subsequent revision, with a total of 15 malpositioned screws. Of the patients who had image-guided navigation, 3 required reoperations for screw repositioning (Table 3). Amongst all malpositioned screws requiring revision surgery, all but 1 occurred in the thoracic and lumbar spine (Table 3). The percent of revised screws in the non-navigation group was 0.12%, 0.64%, and 0.48% in the cervical, thoracic, and lumbar region, respectively. The percent of revised screws in the navigation group was 0%, 0.15%, 0.28%, in the cervical, thoracic, and lumbar region, respectively. There were no revised occipital, sacral, or iliac screws in either group.

The Number of Malpositioned Screws Requiring Revision Surgery, Broken Down by Spinal Level

Our statistical analysis did not reveal any significant difference in the revision rates between non-navigated and navigation-guided screw placement for all regions (P=0.09), in the cervical region (P=1), in the thoracic region (P=0.18), in the lumbar region (P=0.72), or between navigation platforms (P=0.61). There was no significant difference in total revisions by the institution when comparing both the number of screws and patients (P=0.48 and 0.25, respectively). In addition, there was no significant difference in revisions when comparing methods of placement (non-navigated and navigated) by the institution (P=0.58 and 1.00, respectively).


Malpositioned screws in instrumented spine surgery are an ever-present risk, with reported rates of pedicle breach ranging greatly in the literature, from 2% to 50%, depending on the definition of breach.7 Although the majority of malpositioned screws with suboptimal placement will not result in new postoperative symptoms or complications and thus not require revision surgery, malpositioned screws can have catastrophic effects. In the absence of postoperative symptoms or complications related to malpositioned screws, the decision of whether to proceed with a revision surgery or not has to be discussed with the patient.

In the 13 patients who underwent screw revision procedures, 6 were asymptomatic, 6 had radiculopathy, and 1 had a screw breaching the pleural space. Three of the 5 asymptomatic patients, including 1 female pediatric patient, had a medial breach of the pedicle resulting in the screw entering the spinal canal; 1 patient had a superior breach of the pedicle, 1 patient had a thoracic screw breaching laterally and abutting against the aorta, and the last patient had a lateral breach which was redirected during a subsequent revision surgery for adjacent level stenosis. Even though 5 patients had no symptoms related to or caused by malpositioned screws, they all opted to proceed with a revision surgery to prevent any possible future complications.

From this series, our overall rate of subsequent revision surgery for a malpositioned screw, placed posteriorly via a fluoroscopy-assisted or freehand technique, was found to be 2.39% (10/419 patients) compared with 1.11% (3/271 patients) in the navigation group. Overall, 0.41% of screws placed fluoroscopically or freehand (15/3689 screws) required revision. A prior study examining a single institution’s revision rates compared fluoroscopy-assisted to 3D fluoroscopy-navigated pedicle screw insertion in the thoracic and lumbar spine and found a statistically significant decrease in revision rates (1.35% in the 3D fluoroscopy group compared with 4.38% in the freehand group) [odds ratio (OR): 3.35, P<0.01].8 Our study is distinguished in the inclusion of multiple navigation techniques and including instrumentation in all regions of the spine (cervical, thoracic, lumbar, sacral, and iliac). Our figures are comparable to what was reported by the above authors, as well as others. Di Silvestre et al9 reported a rate of 4.3% for revision surgery with a screw malpositioning risk of 1.7% in their series of thoracic screws. A malposition rate of 6.1% for lateral mass cervical screws was reported by Graham et al,10 although the revision rate was not reported. Fluoroscopically assisted thoracic pedicle screw placement carries a risk of screw malpositioning of 0.26% with a rate of revision surgery of 2.1%, as reported by Bransford et al.2

Having grouped all navigation formats together (3D fluoroscopy, CT navigation, and 2D fluoroscopy), we found the rate of a revision surgery to be lower with image-guided navigation compared with fluoroscopic-assisted and freehand screw placement, although statistical significance was not reached. This is likely because the rate of clinically meaningful screw malpositioning is low. We calculate that based on a 1.28% difference in revision rates, a sample size of 3292 patients (1646 patients in each cohort) would be required to detect a significant difference between non-navigated and image-guided navigation. The decreased rate of revision surgery associated with navigation echoes findings of improved accuracy with navigation from prior meta-analyses.6,11 Tian et al,11 found CT-based navigation [OR 95% confidence interval (CI), in vivo: 0.32–0.60, P<0.01], 2D fluoroscopy navigation (OR 95% CI, in vivo: 0.27–0.48, P<0.01), and 3D fluoroscopy navigation (OR 95% CI, in vivo: 0.09–0.38, P<0.01) to all have a lower OR of pedicle violation compared with fluoroscopy-guided placement. Mason et al,6 similarly reported accuracy rates of 68.1%, 84.3%, and 95.5% with fluoroscopy-guided, 2D fluoroscopy navigation, and 3D fluoroscopy navigation, respectively. Multiple meta-analyses, including the 2 discussed above, have found 3D fluoroscopy to have improved accuracy relative to the other navigation strategies.6,11,12 However, there was no difference in the revision rate between navigation platforms in our cohort. Our results do indicate that navigation does not offer a significant advantage or added value in the placement of screws in the occiput, sacrum, or pelvis, where we had no revisions in any group. The potential for increased operative time and increased use of resources, and by extension cost, have led us to take a prudent approach to the use of navigation while placing instrumentation in these regions.

Despite having only 7 pediatric patients in our series, 22.22%, or 4 of 18 revised screws, occurred in 1 female pediatric patient, who had screws placed with fluoroscopic assistance alone. Accuracy rates of pediatric pedicle screws range from 77% to 99%, potentially related to differences in the anatomy of the spine in the pediatric population, which has shorter and smaller pedicles.13 The ability to visualize such anatomy intraoperatively could potentially reduce the risk of screw malpositioning in this group, as has been suggested in a retrospective review, which reported a 96.4% pediatric pedicle screw accuracy rate with image guidance.13

The use of intraoperative navigation can provide several advantages over conventional fluoroscopy or freehand techniques. First, there was a nonsignificant, lower incidence of revision surgery for malpositioned screws. Screws can be accurately directed and placed to the appropriate length using the virtual tip modality, thus, increasing the accuracy of screw placement (Fig. 2). Second, the high-quality images obtained and combined with navigation can serve as a great intraoperative teaching tool in helping trainees understand the anatomy of the spine. Finally, there is the potential advantage of reducing the amount of radiation exposure to the surgical staff.14

Intraoperative images obtained by the O-Arm showing the virtual tip modality being used for navigation and placement of a lumbar pedicle screw (A, B) and of an iliac screw (C, D).


No technology can supplant the spine surgeon’s foundational knowledge of anatomy, but the use of image-guided navigation in instrumented spine surgery may be associated with a decrease in the rate of malpositioned screws requiring revision surgery, although statistical significance was not reached. To further support and clarify the utility of this technology, future studies are warranted, including cost and quality of life analyses.


1. Maeda T, Higashino K, Manabe H, et al. Pullout strength of pedicle screws following redirection after lateral or medial wall breach. Spine (Phila Pa 1976). 2018;43:E983–E989.
2. Bransford R, Bellabarba C, Thompson JH, et al. The safety of fluoroscopically-assisted thoracic pedicle screw instrumentation for spine trauma. J Trauma. 2006;60:1047–1052.
3. Villavicencio AT, Burneikiene S, Bulsara KR, et al. Utility of computerized isocentric fluoroscopy for minimally invasive spinal surgical techniques. J Spinal Disord Tech. 2005;18:369–375.
4. Towner JE, Piper KF, Schoeniger LO, et al. Use of image-guided bone scalpel for resection of spine tumors: technical note. AME Case Rep. 2018;2:48.
5. Kochanski RB, Lombardi JM, Laratta JL, et al. Image-guided navigation and robotics in spine surgery. Neurosurgery. 2019;89:1179–1189.
6. Mason A, Paulsen R, Babuska JM, et al. The accuracy of pedicle screw placement using intraoperative image guidance systems. J Neurosurg Spine. 2014;20:196–203.
7. Woo EJ, DiCuccio MN. Clinically significant pedicle screw malposition is an underestimated cause of radiculopathy. Spine J. 2018;18:1166–1171.
8. Fichtner J, Hofmann N, Rienmuller A, et al. Revision rate of misplaced pedicle screws of the thoracolumbar spine-comparison of three-dimensional fluoroscopy navigation with freehand placement: a systematic analysis and review of the literature. World Neurosurg. 2018;109:e24–e32.
9. Di Silvestre M, Parisini P, Lolli F, et al. Complications of thoracic pedicle screws in scoliosis treatment. Spine (Phila Pa 1976). 2007;32:1655–1661.
10. Graham AW, Swank ML, Kinard RE, et al. Posterior cervical arthrodesis and stabilization with a lateral mass plate. Clinical and computed tomographic evaluation of lateral mass screw placement and associated complications. Spine (Phila Pa 1976). 1996;21:323–328; discussion 329.
11. Tian NF, Huang QS, Zhou P, et al. Pedicle screw insertion accuracy with different assisted methods: a systematic review and meta-analysis of comparative studies. Eur Spine J. 2011;20:846–859.
12. Du JP, Fan Y, Wu QN, et al. Accuracy of pedicle screw insertion among 3 image-guided navigation systems: systematic review and meta-analysis. World Neurosurg. 2018;109:24–30.
13. Larson AN, Santos ER, Polly DW Jr, et al. Pediatric pedicle screw placement using intraoperative computed tomography and 3-dimensional image-guided navigation. Spine (Phila Pa 1976). 2012;37:E188–E194.
14. Costa F, Tosi G, Attuati L, et al. Radiation exposure in spine surgery using an image-guided system based on intraoperative cone-beam computed tomography: analysis of 107 consecutive cases. J Neurosurg Spine. 2016;25:654–659.

fluoroscopy-assisted; malpositioned; navigation; O-Arm; Brainlab; revision; pedicle screw; lateral mass

Copyright © 2020 Wolters Kluwer Health, Inc. All rights reserved.