Effect of posterior ligamentous complex integrity on O-arm navigation accuracy in thoracolumbar trauma: a prospective comparative study

Article information

Asian Spine J. 2026;.asj.2025.0335

Publication date (electronic) : 2026 March 31

doi : https://doi.org/10.31616/asj.2025.0335

Siddharth Aiyer, Sameer Nagpal, Ajay Kothari, Shailesh Hadgaonkar, Pramod Bhilare, Parag Sancheti

Sancheti Institute of Orthopaedics and Research, Pune, India

Corresponding author: Sameer Kumar Nagpal, Spine Department, Sancheti Institute of Orthopaedics and Research, Pune, Maharashtra, 411005, India, Tel: +91-8427025262, Fax: +91-02028999108, E-mail: sameernagpal.97@gmail.com

Received 2025 June 14; Revised 2025 November 27; Accepted 2026 January 6.

Abstract

Introduction

The thoracolumbar spine lies between the biomechanically rigid thoracic region and highly mobile lumbar region and is involved in nearly 50% of all traumatic spinal fractures [1]. With a better understanding of the biomechanics of the spine, posterior ligamentous complex (PLC) injury may influence the choice of treatment [2].

AO has classified the thoracolumbar fractures based on their stability and fracture morphology. Subtypes A0, A1, and A2 are considered stable fractures, whereas subtypes A3, A4, B, and C are considered unstable fracture variants. However, fractures with B2 and C subtypes are accompanied by injury to the PLC, leading to macro- and micromotions at the fracture site [3].

Posterior pedicle screw fixation is often used for stabilizing thoracolumbar fractures and offers excellent strength and three-columbar purchase. The evolution of pedicle screw placement techniques from free-hand techniques to advanced navigation-based imaging modalities underscores the persistent challenge of achieving precise and accurate screw placement. The breach rate, as high as 10%–40%, was reported with the free-hand technique, which can contribute to complications such as vascular, visceral, and neurological injuries [4]. Various navigation techniques can facilitate accurate placement of pedicle screws, including two-dimensional fluoroscopic navigation and more recent three-dimensional (3D) computed tomography (CT)-based navigation systems, such as the O-arm navigation.

O-arm is based on cone-beam CT, which rely upon geometric calibration via a specific arrangement of fiducial markers relative to the patient’s bony anatomy, and it displays virtual image guidance on the O-arm navigation screen [5].

Drift phenomenon in O-arm navigation is a term used to define the loss of accuracy of the O-arm as a result of soft tissue and facet release in deformity surgeries. It is caused by a change in the relative geometric alignment of the anatomy of the bony structures with respect to the reference frame of the O-arm [6]. In contrast, in unstable fractures, the inaccuracy occurs early because of PLC-related micromotions rather than the time-dependent drift seen in deformity surgery.

The author postulates that unstable fracture variants with PLC injury, such as A3/A4+B2 and C subtypes, raise a significant concern about accuracy because of possible micro- and macromotions at fracture sites that can lead to a loss of accuracy of the displayed image in the O-Arm console because of interference of geometric calibration. This may be similar to the drift phenomenon seen with deformity surgery.

A literature gap exists regarding the accuracy of screw placement in thoracolumbar fractures with PLC injury. Thus, this study aimed to evaluate the effect of PLC injury and potential drift phenomenon on O-arm instrumentation, which may cause misplacement of pedicle screws and more cranial facet joint violations.

Materials and Methods

This prospective study enrolled consecutive patients who underwent thoracolumbar fracture fixation using an intraoperative 3D navigation system (O-arm Surgical Imaging System; Medtronic, Minneapolis, MN, USA) in conjunction with a navigation platform (StealthStation S8; Medtronic) over 24 months (June 2022 to May 2024).

The study was approved by the Institutional Review Board of Sancheti Institute of Ortopedics and Research, Pune (IEC-SIOR Agenda 082), and all patients provided written informed consent. Ethical standards of the Declaration of Helsinki and its further amendments were followed.

All patients were treated at a single tertiary spine center (Sancheti Institute of Ortopedics and Research). Surgeries were performed by four fellowship-trained spine surgeons with at least 5 years of post-fellowship experience following a standardized navigation protocol.

Patients aged >18 years presenting with high-energy thoracolumbar fractures classified as AO subtypes A3, A4, A3+B2, A4+B2, or C were included. Conversely, patients with low-energy fractures, pathological fractures, and morphologically stable fractures (AO subtypes A1 and A2) were excluded [7].

For analysis, patients were stratified according to PLC integrity: Group A comprised isolated A3 and A4 fracture patterns without PLC disruption. Group B included fractures with PLC injury (subtypes A3+B2, A4+B2, and C).

PLC injury was defined on preoperative magnetic resonance imaging as the presence of discontinuity, edema, or high signal intensity on T2-weighted or short-tau inversion recovery sequences involving one or more of the following structures: (1) supraspinous ligament, (2) interspinous ligament, (3) ligamentum flavum, or (4) facet joint capsule. Complementary CT findings, such as facet subluxation, widening of the interspinous distance, or avulsion fractures, were also recorded. These findings were further confirmed intraoperatively by direct observation of the discontinuity in the PLC. Assessments were conducted independently by two fellowship-trained spine surgeons, and disagreements were resolved by consensus. All patients underwent surgery as per the institutional trauma protocol, and no systematic difference was set in surgical timing between groups A and B.

Radiological outcomes were evaluated using O-arm data. Pedicle screw accuracy was graded using the Gertzbein–Robbins criteria [8]: breaches <2 mm were considered nonsignificant, whereas those >2 mm were significant. Significant breaches were further classified as noncritical (lateral, anterior, or superior) or critical (inferior or medial) owing to neurovascular risk. Cranial facet violation was assessed using the modified Parks criteria [9]. Local kyphosis was measured with the Cobb angle from the adjacent intact vertebrae. Loss of O-arm accuracy was verified against fixed bony landmarks before each screw insertion. Additional parameters included radiation dose and placement of intermediate (fracture-level) screws.

Other variables that were assessed perioperatively included total blood loss, operative time, length of hospital stay, improvement of the scores in the Visual Analog Scale (VAS), and American Spinal Injury Association Impairment Scale (ASIA). These patients were followed up at months 1, 6, 12, and 24 to assess for any implant failure or nonunion.

Operative procedure

All procedures were performed under general anesthesia, with the patient placed in the prone position on a radiolucent table with standard chest and pelvic padding. No postural reduction was attempted. Any reduction achieved by prone positioning was accepted, and no additional reduction was made before screw insertion. This was done to avoid inadvertent changes in vertebral alignment that could affect the accuracy of O-arm navigation.

The reference frame was attached to the cranial-most instrumented spinous process. Navigation accuracy was verified against a fixed bony landmark before screw insertion. Pedicle entry points were marked using the navigated awl and confirmed using a probe. For intermediate screws, the trajectory was directed toward the intact bone and endplate, with adjustments made in the mediolateral and craniocaudal orientations (Fig. 1). To avoid calibration errors, all screws were placed before any reduction maneuvers. Loss of accuracy, if detected, was corrected with a repeat O-arm spin. A final scan was performed to confirm screw placement, and any mispositioned screws were revised intraoperatively. Reduction and decompression, when required, were performed after screw placement.

Fig. 1

Placement of an intermediate pedicle screw in the fractured vertebra under O-arm navigation in a group B fracture pattern.

Results

A total of 81 patients were included in the study, and 93 fracture levels were fixed. Seventy fracture levels were included in group A and 23 in group B. Groups A and B had mean ages of 53 and 41.7 years, respectively. The L1 vertebra was the most commonly fractured vertebra (30%), followed by D12 (28%) and L2 (16%). Fracture subtypes were also segregated based on AO classification according to fracture morphology (Table 1).

Table 1

Demography of the patient included in the study

A total of 762 pedicle screws were inserted under O-arm navigation guidance. Among these screws, 550 were placed in group A and 212 screws in group B. A total of 25 pedicle screw breaches (4.5%) occurred in group A, of which eight were significant, including one critical breach. Group B recorded 22 breaches (10.4%), with 17 classified as significant, including four critical breaches.

The incidence of significant pedicle breaches was 1.5% (eight out of 550 screws) in group A and 8% (17 out of 212 screws) in group B, demonstrating a significant difference (p<0.01). Similarly, the rates of critical pedicle breaches were 0.2% in group A and 1.8% in group B, which were also significant (p=0.023). In contrast, the rate of nonsignificant breaches was comparable between the groups—3% in group A and 2.5% in group B, with no significant difference (p=0.80).

Of the total breaches, 21 (45%) were medial, 20 (43%) were lateral, 4 (8.5%) were anterior, and 2 (4.2%) were inferior (Table 2). Of all critical pedicle screw breaches, 4 (80%) occurred in group B. All pedicles with critical breaches were revised and confirmed intraoperatively, and no postoperative neurological worsening was noted (Fig. 2). The cranial facet joint violation rates were 2.5% (3/120) in group A and 9.5% (4/42) in group B (p=0.07), which was not significant but suggested a higher trend of violation in group B.

Table 2

Pedicle screw breach analysis

Fig. 2

Lateral pedicle breach in the left pedicle of a dorsal vertebra in a group B fracture variant (A). Significant medial pedicle breach on the right side in an upper lumbar vertebra in a group B fracture variant (B).

At the fracture level, a total of 58 screws were inserted across 44 fracture segments (group A, n=44; group B, n=14). The breach rate at the fracture level was 9.1% in group A (4/44 screws, of which one was significant and three were nonsignificant) and 7.1% in group B (1/14 screws, nonsignificant). For screws placed at adjacent nonfracture levels, 506 screws in group A had 21 breaches (4.1%), whereas 198 screws in group B had 21 breaches (10.6%) (Table 3).

Table 3

Level-specific pedicle screw breach analysis

The mean preoperative kyphosis values were 14.7° in group A and 18° in group B (p=0.112), which was not significant. The mean kyphosis corrections upon stabilization in groups A and B were 4.5° and 5.9°, respectively (p=0.198, not significant). In total, six instances of loss of accuracy were observed, where a repeat O-arm scan was taken before continuing screw insertion: 3 (5%) in group A and 3 (14.3%) in group B (p=0.112, not significant) (Fig. 3).

Fig. 3

Illustration of loss of accuracy in a group B fracture pattern. The O-arm navigation initially demonstrated an optimal pedicle screw trajectory (A); however, the final O-arm scan revealed a 3 mm medial breach of the right-sided pedicle screw at a lower thoracic level (B).

Of the 60 patients in group A, 9 (15%) presented with neurological deficits, compared with 10 of 21 patients (47.6%) in group B (p=0.005), indicating a significantly higher neurological burden in group B. Recovery of at least one ASIA grade was achieved in 77.8% of patients in group A versus 20.0% in group B (p=0.012), again favoring group A (Table 4).

Table 4

Comparison of clinical and radiological outcomes in group A and group B

For the total study population, the mean duration of surgery was 215 minutes in both groups, with an average blood loss of 230 mL and a mean radiation exposure of 611 mGy·cm². The mean reduction in VAS score was 4.7 points, with mean VAS scores of 8.01 and 3.23 preoperatively and postoperatively, respectively. These values were identical between the groups and thus were not subjected to statistical comparison.

One case of implant failure in group B (4.8%) was due to rod breakage 18 months postoperatively. The fractured vertebra showed good consolidation. Rod exchange with implant shortening was performed. The difference in implant failure between the groups was not significant (p=0.259). No wound complications or infections occurred during the 2-year follow-up period in either group.

Discussion

The concept of screw fixation through the pedicle was first described by Roy-Camille [10] in 1970s. In the last few decades, pedicle screws underwent significant advancement in design and composition, as they have a promising ability to achieve three-column fixation and superior biomechanics. Pedicle screws provide better construct stability and better correction of a deformity over wires and hybrid hook screw constructs [11].

Over the last few decades, various modalities have emerged for the accurate placement of pedicle screws, with CT-based navigation being the most accurate system. In a study by Lu et al. [12], the overall anatomical pedicle screw breach rate in O-arm–navigated thoracolumbar fracture was 8.3%, which is consistent with the 6.2% anatomic breach rate in the present study. Given that reduction maneuvers can affect the accuracy of the O-arm, we did not attempt fracture reduction before screw insertion. However, we still had more pedicle breaches in group B fractures. This is comparable to the drift phenomenon described by Zhao et al. [6], who reported an incidence of 16.7% in deformity cases. In our series, accuracy loss was observed in six cases, with an incidence of 14% in group B fractures and 5% in group A. In each instance, a repeat O-arm scan was performed to restore accuracy. However, the classical “drift phenomenon” in deformity surgery typically refers to progressive misregistration that develops over time with extensive soft tissue or facet releases. In contrast, the inaccuracy seen in our unstable fracture cohort occurred early, likely due to micro- and macromotions set the fracture site, leading to transient misalignment with the reference frame. Thus, although conceptually related, the mechanism in trauma differs from the time-dependent drift described in deformity correction. Therefore, checking the accuracy before every screw insertion is recommended.

In this study, group A had a critical pedicle breach rate of 0.2% and a significant pedicle breach rate of 1.45%. In contrast, group B had higher breach rates, at 1.8% for critical and 8% for other significant breaches. Of all critical breaches observed, 80% occurred in group B, indicating a greater technical challenge in these fracture patterns, as they had micro- and macromotions, which led to a higher rate of accuracy loss and a higher number of pedicle breaches than group A fractures. The relatively low post hoc power for critical breaches reflects the limited number of such events; however, the observed effect size indicates a clinically meaningful difference that warrants confirmation in larger cohorts.

Cranial facet joint preservation has been given much importance with better understanding of the spinal motion segment, which can result in adjacent segment disease [8]. Ohba et al. [13] demonstrated that the cranial facet joint violation rates were as high as 30% in the free-hand technique of pedicle screws compared with 3.8% in 3D CT computer navigation-guided screws. In the present study, we could observe seven facet joint violations, of which 4 (57%) were observed in group B, indicating the direct association of violation rate with PLC injury and motion at the fracture site, leading to a loss of geometric calibration [13].

In the present study, PLC injury was found to be an independent factor for accuracy loss, pedicle screw breaches, and cranial facet violations. No correlation with the distance of screw level from the reference frame was noted, which further confirmed the relative loss of accuracy in group B fractures. Multivariate analysis did not identify age as an independent predictor of pedicle screw accuracy or breach rates, confirming that PLC injury remained the primary factor associated with navigational inaccuracy.

Kyphosis >20° is a known predictor of posttraumatic progression [14]. In our cohort, the mean postoperative kyphosis was 9.7°, with corrections of 4.5° in group A and 5.9° in group B (Fig. 4). Group B demonstrated greater preoperative kyphosis (18° vs. 13°), a difference that persisted postoperatively (12° vs. 8.6°). These findings highlight the role of the PLC in maintaining sagittal alignment and limiting postoperative kyphosis (Fig. 4).

Fig. 4

(A, B) Illustration of kyphosis correction in a group B fracture variant, with preoperative O-arm scan showing 20.9° kyphosis and postoperative scan showing correction to 12.8°.

In this level-specific analysis, breaches were uniformly distributed across all instrumented levels rather than concentrated distally. At the fracture (intermediate) level, breach rates were similar (group A, 9.1%; group B, 7.1%), whereas at adjacent levels, group B showed a markedly higher rate than group A (10.6% vs. 4.1%). These findings suggest that PLC injury causes a global reduction in navigational accuracy, likely due to micromotions that disrupt the patient reference geometry and produce system-wide drift rather than level-specific inaccuracy.

In this study, the placement of intermediate or fracture vertebrae screws is the distinct advantage of using an O-arm, which significantly reduces the incidence of posttraumatic kyphosis. Dick et al. [4] suggested that short-segment constructs with two fractured vertebrae screws were 84% stiffer than four-screw constructs. Under O-arm guidance, we have placed at least one fracture vertebrae screw in 44 levels (47%). In a study on intermediate screws, Mahar et al. [15] observed a correction of 5° in segmental kyphosis, which is comparable to our finding, as we could achieve a correction of 4.9°.

Nguyen and Phan [16] demonstrated a hardware failure rate of 16.7% at a 53-month follow-up for short-segment fixation, and screw breakage above the fractured vertebrae was the most common form of hardware failure.

In the present study, we observed one case of implant breakage 18 months postoperatively, in which the patient underwent surgery for an L4 burst fracture of AO subtype A4+B2. In this case, the index screw was not placed in either of the pedicles, demonstrating the role of the intermediate screw in terms of construct stability in fractures at junctional levels (Fig. 5).

Fig. 5

Postoperative case of implant failure at 18-month follow-up showing bilateral rod breakage (A). Revision surgery was performed with rod exchange and implant shortening (B).

O-arm can essentially help minimize the radiation exposure to zero for the operating staff. In this study, the mean radiation exposure to the patient was 611 mGy·cm, which is 5–6 times less than the threshold dose [17].

This study has several limitations. Although all surgeons followed a standardized protocol, differences in technique may have influenced the accuracy of screw placement. Trauma mechanism (high vs. low energy) could also affect intraoperative stability and navigation precision. Group B included a relatively small sample, which may limit statistical power. Finally, as a single-center series with a modest cohort size, the findings may not be broadly generalizable.

Conclusions

This study highlights that PLC injury significantly reduces the accuracy of O-arm navigation-guided pedicle screw placement in thoracolumbar fractures. Fracture patterns with PLC injury are associated with higher rates of navigation errors, critical pedicle breaches, and intraoperative accuracy loss despite strict adherence to navigation protocols, likely due to increased micro- and macromotions at the fracture site. Surgeons should remain vigilant in such cases and consider repeat scans when necessary to minimize pedicle screw malposition.

Article information Continued