INTRODUCTION
Cervical instrumented fixation is aimed at achieving cervical stabilization and alignment correction by choosing a procedure that affords the requisite fixation strength [1,2]. However, possible surgical complications, such as nerve and vertebral artery (VA) injuries, necessitate careful attention [3,4]. Pedicle screws (PSs) and lateral mass screws (LMSs) are predominantly utilized for middle to lower cervical spinal fixation [5,6]. Although PSs provide more solid fixation than LMSs do, they are associated with a higher risk of nerve and VA injuries than LMSs are [7].
Aramomi and Maki [8] reported the effectiveness and safety of paravertebral foramen screws (PVFSs) as an alternative technique for posterior cervical fixation [2]. VA injuries are less likely with PVFSs than with PSs because PVFSs are shorter screws that do not to reach the transverse foramen. PVFSs are also characterized by a larger screw diameter and better cortical purchase than LMSs because the tips of the screws are directed toward the cortical bone of the vertebral foramen. This unique screw selection and insertion procedure for PVFSs significantly reduces the risk of nerve and VA injury compared with PSs and provides stronger fixation than LMS does. Therefore, increased cortical purchase of screw tips is crucial for maximizing the fixation strength of the PVFS [2]. The PVFS trajectory serves as an effective alternative in posterior cervical spine surgery, particularly when conventional LMSs provide insufficient stability or when PSs pose a high risk of VA injury. Additionally, it serves as a salvage option in cases where LMS insertion results in lateral mass fractures. Compared to PSs, PVFS reduces the risk of vascular injury while maintaining adequate fixation strength. Its ability to engage denser cancellous bone also makes it a suitable choice for patients with osteoporosis.
Although PVFSs were originally inserted under lateral fluoroscopic guidance [8], fluoroscopic guidance may not be sufficiently reliable. Shimizu et al. [9] reported the clinical outcomes of 94 PVFSs inserted using fluoroscopic images. Among the 94 screws, 14 (14.9%) were placed in inappropriate positions with 9 screws (9.8%) penetrating the vertebral foramen and 5 screws (5.3%) penetrating the transverse foramen while no nerve or VA injuries occurred. Furthermore, cortical bone contact was achieved with only 39 screws (41.5%) with 41 screws (43.6%) positioned in the lateral mass.
An intraoperative image-based navigation system, called the O-arm navigation system, is highly reliable and beneficial for enhancing the accuracy and safety of screw placement [3]. It is characterized by automatic registration and is based on the acquisition of postpositioning images with a significant reduction in the surgeon’s degree of radiation exposure [3,10-12]. However, there have been no previous reports of PVFSs inserted under O-arm navigation. Thus, there is little evidence on the accuracy of the positioning of the screw tips, which is critical for the optimal fixation strength of PVFSs. The navigation system has an inherent margin of error within a few millimeters, which should be considered in screw insertion [3,11,13-17]. However, it is important to recognize that O-arm navigation has been primarily studied in the context of cervical PS placement, and no prior reports exist for PVFS placement. Therefore, while direct comparison is not possible, previous PSs studies suggest that navigation-assisted screw placement typically results in errors of less than 4 mm, which should be taken into account when using O-arm navigation for PVFS placement. Thus, we devised a novel technique for PVFS placement under O-arm navigation to place the screw tips in ideal locations.
In this study, we reported our novel surgical technique of PVFS placement to achieve good cortical purchase of screw tips.
MATERIALS AND METHODS
1. Surgical Techniques
1) Overview of stepwise PVFS
All screws were placed using the novel surgical technique described here, called stepwise PVFS. The entry point and trajectory itself were the same as those used in the original PVFS technique. Although fluoroscopy was used for screw insertion in the original PVFS method, all procedures were performed with navigation guidance without the use of fluoroscopy in our stepwise PVFS approach. The O-arm was harnessed for intraoperative image acquisition, and StealthStation S8 served as the navigation system. The reference frame was usually attached to the C2 spinous process using a single spine clamp tool or to the two most rostral spinous processes within the fixation levels using a double spine clamp tool. Screws with a diameter of 4.5 mm (InfinityTM; Medtronic) were exclusively utilized for all procedures.
2) Five key steps of stepwise PVFS
(1) Determining the insertion point and trajectory
The optimal insertion point and trajectory were determined using O-arm navigation, following the original PVFS placement technique [2]. The starting points are the same as those in the original PVFS method, 0–1 mm medial to the intersection of the midline of the lateral mass and the lateral notch of the lateral mass. The original method describes an insertion angle of 20° to 25° (Fig. 1A). A navigated high-speed drill (StealthMidas, Medtronic) was used to create starting holes. The screw trajectory was also drilled and tapped aiming to put the screw tip in the cortical bone of the lateral edge of the vertebral foramen at the entrance of the pedicle.
(2) Measuring screw length
The trajectory length was measured on navigation images. A screw two mm longer than the measured length was selected for the purpose of stepwise depth adjustment (Fig. 1B).
(3) Initial screw insertion
The screw was inserted to reach the optimal position on the navigation screen. Due to the extra length of the screw, approximately 2 mm of the screw head and thread were extracted to project beyond the prepared screw hole and the lamina surface (Fig. 1C).
(4) Verifying the screw position
A second O-arm scan image was obtained to ascertain whether the screws had reached or stopped short of the ideal position (Fig. 1D). The ideal position (ideal screw) was defined as the screw tip touching the lateral cortex of the vertebral foramen at the entrance of the pedicle, without any penetration into the transverse foramen or vertebral foramen (i.e., the spinal canal).
(5) Postimaging depth adjustment
After verification, if the screw tips did not reach the ideal position, screws were further inserted to achieve the ideal placement. For screws already located in the ideal position but with extra space at the tip, minimum screw head adjustments were made to match up the screw head positions (Fig. 1E).
2. Patient Characteristics
A retrospective analysis was conducted. Between October 2022 and February 2024, stepwise PVFS was performed in 12 consecutive patients (78 screws), including 5 men and 7 women patients, with an average age of 75 years (range, 52–85 years). The mean follow-up period was 471 days (range, 47–834 days). The indications for surgery included cervical spondylosis (5 patients), cervical spine injury (3 patients), ossification of the posterior longitudinal ligament (2 patients), and dropped head syndrome (1 patient). The levels of fusion were as follows: C1–4 (1 patient), C2–7 (1 patient), C2–T1 (1 patient), C2–T2 (2 patients), C3–4 (1 patient), C3–6 (1 patient), C4–6 (1 patient), C4–7 (2 patients) C4–T3 (1 patient), and C5–7 (1 patient) (Table 1). A total of 78 screws were inserted via a stepwise PVFS in 12 patients as follows: 13 in C3, 22 in C4, 18 in C5, 19 in C6, and 6 in C7. The surgical indications for stepwise PVFS were similar to those for LMSs. Each screw trajectory was planned preoperatively, and no screws were inserted as a salvage technique for any other screws (Fig. 2). This study received approval from the ethics committee of the affiliated hospital (approval No. Reiwa5-16), and informed consent was obtained from all patients.
3. Radiological Assessment
All 12 patients underwent plain cervical x-ray imaging and CT multiplanar reconstruction the day after surgery. We assessed the accuracy of screw placement with great attention to observing the locations of the screw tips. The ideal position (ideal screw) was defined as the screw tip touching the lateral cortex of the vertebral foramen at the entrance of the pedicle. Short-length screws were defined as screws whose tips stopped short of the cortical bone of the vertebral canal. Complications associated with stepwise PVFS, such as neurovascular injuries, screw malposition and screw loosening, were documented from medical records and radiological evaluations.
RESULTS
There were no adverse events related to screw insertion. Among the total 78 screws, 12 (15%) were classified as ‘ideal position (ideal screws)’ according to the intraoperative second O-arm CT, while the remaining 59 (76%) were categorized as ‘short-length screws,’ 4 (5%) showed lateral deviation, and 3 screws could not be evaluated due to unclear imaging. There were no excessively long screws penetrating the cortex of the vertebral canal or transverse foramen. A total of 67 screws (86%) were further inserted: including 8 ideal screws (10%) adjusted to match up the screw heads if extra space existed and 59 short-length screws (76%) inserted to achieve the ideal position. Of the 11 screws that were not further inserted, 4 (5%) were left because they were already in the ideal position without extra space, 4 (5%) were not further inserted due to lateral deviation, and 3 (4%) were not considered for further insertion due to unclear imaging. Postoperative CT scans revealed that 70 screws (90%) were ultimately placed in the ideal position. The lengths of the screws used were 10 mm for 2 screws, 12 mm for 18 screws, and 14 mm for 58 screws. There were no cases of screw loosening or pseudoarthrosis at the final follow-up. The fusion rate was 10 out of 12 cases (83%). In the remaining 2 cases, bony fusion could not be confirmed due to the short follow-up period; however, radiographic evaluation demonstrated stability. Radiographic evaluation confirmed stable screw positioning in all patients, and no revision surgery was required due to mechanical failure (Table 2).
DISCUSSION
The results of our novel surgical technique were much better than those of the conventional technique of PVFS placement in previous reports [9]. In previous studies, fluoroscopy-guided PVFS placement achieved an appropriate position rate of 80%–100%, while the ideal position was achieved in only 41.5% of cases. Additionally, malposition rate of 15%–20% have been reported, including vertebral foramen and transverse foramen perforation (Table 3). In contrast, our study using stepwise PVFS achieved an appropriate position rate of 100% and an ideal position rate of 90%, with no cases of malposition or complications. The 95% confidence interval for the ideal position accuracy of O-arm guided stepwise PVFS placement was 81.0%–94.7%, further supporting its reliability. The evolution of intraoperative imaging and navigation techniques has fundamentally changed surgical procedures involving the cervical spine [18]. These advancements aim to increase the safety and accuracy of the procedure. Nevertheless, errors cannot be completely eliminated even with the use of the O-arm system because of the high mobility of the cervical spine [3,11,14,16,19].
We should always be aware that the images displayed on the navigation screen are not real-time data. Rotation or an anterior shift of the vertebral bodies can always occur throughout the procedure. Although repeating the registration for each vertebral body can minimize navigation errors, this process is timeconsuming [14,19]. Even if we try to reduce various sources of errors as much as possible, the system technically has possible errors of a few millimeters [13,17,20]. We have referenced studies on cervical PSs placement using O-arm navigation, which provide insight into the expected error margins. According to previous studies, the majority of PSs placements using O-arm navigation resulted in less than 4-mm deviation, with accuracy rates ranging from 88.9% to 99.3% (Table 4) [11,17,21-23]. Therefore, an error margin of 1 mm or a few millimeters should be accounted for, particularly in cervical fixation, where such a discrepancy can have significant consequences [16]. The stepwise PVFS method allows for precise surgical adjustments by correcting the screw position based on the intraoperative CT in a stepwise manner. In our study, stepwise PVFS method achieved an ideal screw position rate of 90%, with no cases of malposition exceeding acceptable deviation. While a direct comparison with cervical PSs is challenging due to anatomical and technical differences, our findings align with previous reports using O-arm navigation in terms of precision and safety (Table 4).
Although PVFSs can provide more robust fixation than LMSs do, the fixation strength might be compromised if the screw tips are not properly inserted into the cortex around the vertebral foramen at the entry zone of the pedicle [2,24]. Several biomechanical studies on PVFSs emphasize the importance of screw tip cortical purchase around the vertebral foramen [2,24]. In a cadaveric study by Maki et al. [2], PVFS demonstrated a mean pullout strength of 234±114 N, which was superior to LMS (158± 91 N, p=0.06) and comparable to salvage PVFS (195±125 N). This suggests that PVFS can provide stronger initial fixation than LMS and serve as a viable salvage technique in cases of LMS failure. Furthermore, Tsuda et al. [24] conducted a CT-based study evaluating screw trajectories for PVFS, PSs, and LMS, demonstrating that PVFS trajectories exhibited higher CT attenuation values than LMS trajectories at every cervical level and higher than PSs trajectories at C5 and C6. Given that CT attenuation values correlate with bone mineral density and pullout strength, these findings further support the biomechanical advantage of PVFS placement. While fluoroscopy is not reliable enough for the ideal insertion of PVFSs based on previous reports, relying solely on the conventional use of a navigation system might fall short of expectations. In previous reports, only 40% of PVFSs were ideally inserted under fluoroscopic guidance. Furthermore, penetration into the spinal canal or transverse foramen has been detected in 5% to 20% of cases, although neurological deficits or VA injuries have not been observed. The present study revealed a high success rate of 90% in achieving ideal positions without any case of screw penetration into the vertebral canal or transverse foramen [9,25] (Table 3). Our proposed stepwise PVFS method, as described in this study, enhances the safety and accuracy of O-arm navigation by considering its inherent error margin. This highlights not only the accuracy improvements but also the biomechanical benefits of optimal screw trajectory and cortical engagement. While our study did not perform direct pull-out strength testing, the combination of previous biomechanical findings and our accuracy results strongly supports the clinical feasibility and mechanical stability of stepwise PVFS method.
Navigation errors can lead to screw tip misplacement, causing it to fall short of the ideal position. This difference is thought to stem from the mobility of the cervical spine. All procedures of screw insertion may lead to rotation and a ventral shift of the vertebral bodies [11,26]. Because the reference arm was typically fixed on the C2 spinous or on the 2 most cranial spinous processes, a ventral shift or rotation of the vertebral body could have oc-curred at other levels. These intraoperative bone shifts could result in shallower screw tip placement or more lateral screw translation than initially planned. As a result, the rate of ideal positioning of PVFS tips predictably falls short of expectations under conventional O-arm navigation guidance.
In this study, 67 out of 78 screws (86%) required additional advancement based on intraoperative CT image obtained after initial screw insertion under navigation guidance. Ultimately, 70 screws (90%) were confirmed to be properly positioned postoperatively. The high rate of additional insertions does not necessarily reflect inaccuracies in navigation. The inherent errors in navigation systems, which may result in shallow screw placement, should be considered throughout the procedure. The insertion torque of the PVFS remarkably increases as the tip of the PVFS approaches the cortical bone [2,27], which serves as an important clue to reach the ideal positions. For initial screw insertion, when an increase in torque is detected, it is prudent to avoid excessive advancement to avoid penetration of the transverse foramen or vertebral foramen.
Even if the screw head is projected 2 mm compared with other screws with no additional screw advancement, bending techniques and polyaxial screws facilitate easy rod fixation. However, there could also be a concern of unintentional alignment correction in rod fixation of unmatched screw heads, because forcible rod fixation could lead to vertebral rotation or anterior displacement. In this context, further screw insertion of ‘ideal screws’ with enough space can be justified to avoid excessive rod bending and unintentional alignment corrections can be avoided.
This study has several limitations. First, the sample size was relatively small, with only 12 patients; however, the total number of screws (78) was comparable to or larger than those in previous studies on PVFS accuracy (Table 3). PVFS using O-arm navigation is a relatively novel technique, making it difficult to accumulate a larger patient cohort. Despite the limited number of patients, our analysis based on screw-level data provides meaningful insights into the accuracy of this technique. While the follow-up period was relatively short, this study was not an investigation of the outcomes of PVFSs but rather a report of the technical tips to achieve ideal PVFS insertion. Long-term follow-up of a larger number of cases is necessary for further verification of the effectiveness of our novel surgical techniques.
A shallow insertion angle with reduced triangulation could compromise the pull-out strength of the PVFS. In our current study, some screws were inserted at a shallower angle, particularly those in the lateral mass. The original method describes an insertion angle of 20 to 25 degrees; however, some screws were reported to deviate into the spinal canal [9]. It is possible that navigation-assisted PVFS may result in shallower angles as a trade-off to avoid excessive deviation into the spinal canal. This trade-off between avoiding excessive medial deviation and ensuring sufficient cortical purchase should be carefully considered, and further studies are needed to determine the optimal balance. Lateralizing the entry point compared to the original technique could compensate for this tendency. Furthermore, even when tapping is performed, we hypothesize that in case with exceptionally hard lateral cortex of the vertebral foramen, the screw trajectory may tend to deviate laterally as the screw tip begins to contact the hard cortex. However, this has not been fully verified. This triangulation loss factor, which could impact the strength of the PVFS, is acknowledged as a limitation of this study.
CONCLUSSION
In this study, we investigated a novel method for achieving the ideal positioning of PVFS using O-arm navigation, considering the inherent error margin of the navigation technique. The stepwise PVFS method enhances safety and accuracy, improving fixation strength by placing the screw tip at the ideal position, compared with fluoroscopy-guided PVFS placement.
PDF Links
PubReader
ePub Link
XML Download
Full text via DOI
Full text via PMC
Download Citation
Print
