Modified paravertebral foramen screw trajectory for posterior cervical spine fixation: feasibility of computed tomographic evaluation

Article information

Asian Spine J. 2026;.asj.2025.0487

Publication date (electronic) : 2026 January 21

doi : https://doi.org/10.31616/asj.2025.0487

Sadaki Mitsuzawa^1,2, Eijiro Onishi¹, Satoshi Ota¹, Shinnosuke Yamashita¹, Yoshihiro Tsukamoto¹, Hisataka Takeuchi¹, Tadashi Yasuda¹, Shuichi Matsuda²

¹Department of Orthopaedic Surgery, Kobe City Medical Center General Hospital, Kobe, Japan

²Department of Orthopaedic Surgery, Kyoto University Graduate School of Medicine, Kyoto, Japan

Corresponding author: Sadaki Mitsuzawa, Department of Orthopaedic Surgery, Kobe City Medical Center General Hospital, 2-1-1 Minatojima-minamimachi, Chuo-ku, Kobe 650-0047, Japan, Tel: +81-78-302-4321, Fax: +81-78-302-7537, E-mail: sadakimitsuzawa@gmail.com

Received 2025 August 21; Revised 2025 September 3; Accepted 2025 October 14.

Abstract

Introduction

The use of cervical posterior fusion surgery is increasing in aging populations, particularly among elderly individuals with a healthy life expectancy. Cervical screws provide better fixation than wiring or hook systems. Pedicle screws (PSs) and lateral mass screws (LMSs) are two major types of posterior cervical spine fixation. The paravertebral foramen screw (PVFS) is an emerging technique designed to enhance stability and biomechanical support [1]. By contrast to the traditional PS or LMS, the PVFS is intended to achieve robust fixation with minimal disruption to surrounding structures, making it a promising alternative for treating cervical degenerative diseases, trauma, tumors, and deformities [2]. Advances in surgical techniques and instrumentation have contributed to the increasing interest in PVFS in spinal surgery.

PVFS has certain limitations despite the advantages of a safer trajectory, minimal invasiveness, and improved biomechanical stability. A major concern is the steep learning curve associated with this technique during the initial learning phase, requiring precise anatomical knowledge. Although studies have demonstrated the pull-out strength of PVFS, the effectiveness of very short screws in elderly patients with poor bone quality remains uncertain [3]. Additionally, long-term clinical studies that compare the durability and efficacy of PVFS with other well-established fixation methods are lacking. Surgeons should select the longest possible screw to avoid postoperative complications, such as screw loosening.

To address these concerns, we developed a modified PVFS (mPVFS) with a longer trajectory. This adaptation involves inserting the screw upward, similar to LMS, and aligning it with the axis of the lateral mass facet. We aimed to investigate the trajectories and bone qualities of the mPVFS on computed tomography (CT) images and compare them with those of the PVFS and LMS. We also evaluated age and sex subgroups in future clinical situations.

Materials and Methods

The Ethics Committee of Kobe City Medical Center General Hospital (approval no., 24242) approved this retrospective study performed in accordance with the principles of the Declaration of Helsinki (as revised in 2013). All patients provided informed consent before study inclusion and after the publication of the anonymized results.

Patients who visited the emergency department of Kobe City Medical Center General Hospital between 2023 and 2024 and underwent cervical spine CT were examined. We included patients who underwent cervical CT due to head trauma or neck pain. We excluded patients with compression myelopathy, ossification of the posterior longitudinal ligament, cervical spine injury, or metabolic bone disease. Moreover, we compared extreme age groups to provide novel insights. Groups 1 and 2 included female and male patients, respectively, aged 20–35 years. Groups 3 and 4 included female and male patients, respectively, aged 85–100 years. The study included 40 patients, with each group comprising 10 patients.

A cervical spine CT scan was performed using a 320-row multidetector CT system (Aquilion One PRISM; Canon Medical Systems, Tokyo, Japan). The imaging conditions were tube voltage, 120 kV; tube current, 300 mA; collimator, 0.5 mm; reconstruction slice thickness, 1 mm; slice interval, 1 mm; field of view, 250 mm; matrix, 512×512; and kernel FC30. Multiplanar reconstruction images were generated from the acquired data (Synapse Vincent; Fujifilm, Tokyo, Japan).

The C5 left lateral mass and C5 vertebral body were measured to assess bone mineral density (BMD). A cubic region of interest (ROI) of 5.0×5.0×5.0 mm was defined at the center of the lateral mass and vertebral body in a three-dimensional model. CT attenuation values were measured, as previously reported (Fig. 1) [2,4].

Fig. 1

A 5.0-mm×5.0-mm×5.0-mm cubic region of interest (yellow square) was defined at the center of the lateral mass and vertebral body in the three-dimensional model. Computed tomography attenuation values were measured.

Fig. 2 shows the insertion methods for the three screw types. In the PVFS method, oblique cross-sectional images aligned with the vertebral pedicle axis are created. The insertion point is 1 mm medial to the center of the lateral mass in the transverse section, with the screw tilted inward at 20° [5]. The ROI is defined as an ellipse with a short diameter of 4.5 mm.

Fig. 2

Insertion method for the three screw types (yellow ellipse). The long diameter, area, and computed tomography attenuation values of the region of interest of each screw were measured. In the paravertebral foramen screw (PVFS) method, oblique cross-sectional images aligned with the vertebral pedicle axis are created. An ellipse with a short diameter of 4.5 mm and tilted inward at an angle of 20° is applied. In the modified PVFS method, oblique cross-sectional images aligned with the axis of the lateral mass facet are created. An ellipse with a short diameter of 4.5 mm and tilted inward at an angle of 20° is applied. In the lateral mass screw (LMS) method, on the same plane as modified paravertebral foramen screw (mPVFS), an ellipse with a short diameter of 3.5 mm and tilted outward at an angle of 25° is applied.

In the mPVFS method, oblique cross-sectional images aligned with the axis of the lateral mass facet are created. The insertion point is positioned 1 mm lateral to the center of the lateral mass in the transverse section, and the screw is tilted inward at 20°. The ROI is designated as an ellipse with a short diameter of 4.5 mm.

The trajectory is measured using the Magerl method in the LMS method. Oblique cross-sectional images aligned with the axis of the lateral mass facet are created. The insertion point is 2 mm medial to the center of the lateral mass in the transverse section, and the screw is tilted outward at 25°. The ROI is defined as an ellipse with a short diameter of 3.5 mm [6].

Measurements of the long diameter, area, and CT attenuation values of the ROI for each screw were taken on the left side of the C3–C6 vertebrae in each patient. CT attenuation values represent the X-ray absorption capacity of a substance within a small unit volume in a CT examination, with values expressed in Hounsfield units (HU).

The data were described as means and standard deviations. The Tukey-Kramer method was used to compare differences in CT attenuation values between different groups and screw trajectories. All data analyses were performed using IBM SPSS ver. 25.0 (IBM Corp., Armonk, NY, USA), with a significance level of p<0.01. Additionally, η²p was used to calculate effect size with the following cut-off values: 0.01–0.06, 0.06–0.14, and >0.14 interpreted as small, medium, and large effects, respectively. G*Power ver. 3.1.9.6 (Heinrich-Heine-Universität Düsseldorf, Düsseldorf, Germany; http://www.gpower.hhu.de/) was applied to determine the suitable sample size. An a-priori power analysis with α=0.05 and an effect size of 0.7 showed that at least 10 patients in each group were adequate to detect significant differences with a power of 90%.

Results

The mean age of patients was 25.6±4.0 and 24.7±4.1 years in groups 1 and 2, respectively, and 90.9±4.7 and 90.4±2.4 years in groups 3 and 4, respectively. CT attenuation values of the C5 lateral mass in the four groups were 523±108, 364±93, 221±85, and 338±194 HU, respectively, with significantly higher values in group 1 than in group 3 (p<0.001, η²p=0.441). CT attenuation values of the C5 body in the four groups were 403±68, 342±54, 157±33, and 229±64 HU, respectively. Groups 1 and 2 had significantly higher values than groups 3 and 4 (p<0.001, η²p=0.761) (Fig. 3).

Fig. 3

The computed tomography attenuation values of the C5 lateral mass and vertebral body in the four groups. The horizontal black lines above each bar graph indicate significant differences between pairs of groups. HU, Hounsfield unit.

The long diameters of PVFS, mPVFS, and LMS were 10.5±1.1, 12.9±1.2, and 12.2±1.3 mm in group 1, respectively (η²p=0.421); 12.1±1.4, 14.6±1.3, and 13.5±1.4 mm in group 2, respectively (η²p=0.357); 11.8±1.3, 15.1±1.4, and 14.0±1.3 mm in group 3, respectively (η²p=0.523); and 12.8±1.3, 16.0±1.7, and 14.2±1.6 mm in group 4, respectively (η²p=0.418). The mPVFS and LMS in groups 1, 2, and 3 were not significantly different. However, all other comparisons were significantly different (p<0.001) (Fig. 4).

Fig. 4

Length of each screw in the four groups. The horizontal black lines above each bar graph indicate significant differences between pairs of screw types. PVFS, paravertebral foramen screw; mPVFS, modified paravertebral foramen screw; LMS, lateral mass screw.

The areas of PVFS, mPVFS, and LMS were 37.6±4.0, 45.8±4.5, and 33.5±3.4 mm² in group 1, respectively (η²p=0.628); 43.1±4.6, 51.3±4.9, and 36.8±3.8 in group 2, respectively (η²p=0.646); 41.7±5.2, 53.0±5.1, and 38.8±3.6 mm² in group 3, respectively (η²p=0.636); and 45.5±5.0, 56.3±6.2, and 39.0±4.3 mm² in group 4, respectively (η²p=0.656). The PVFS and LMS in group 3 were not significantly different. However, all other comparisons were significantly different (p<0.001) (Fig. 5).

Fig. 5

Screw insertion area of each screw in the four groups. The horizontal black lines above each bar graph indicate significant differences between pairs of screw types. PVFS, paravertebral foramen screw; mPVFS, modified paravertebral foramen screw; LMS, lateral mass screw.

CT attenuation values of PVFS, mPVFS, and LMS were 500±142, 611±148, and 538±145 HU in group 1, respectively (η²p=0.094); 380±84, 459±119, and 414±120 in group 2, respectively (η²p=0.082); 216±80, 255±87, and 233±77 HU in group 3, respectively (η²p=0.038); and 423±177, 475±170, and 428±150 HU in group 4, respectively (η²p=0.020). The CT attenuation value of the mPVFS was significantly greater than that of the PVFS in groups 1 and 2 (p=0.002, 0.005, respectively) (Fig. 6).

Fig. 6

The computed tomography attenuation value of each screw in the four groups. The horizontal black lines above each bar graph indicate significant differences between pairs of screw types. PVFS, paravertebral foramen screw; mPVFS, modified paravertebral foramen screw; LMS, lateral mass screw.

Discussion

This study revealed that the mPVFS has a longer trajectory and larger occupancy area than do the PVFS and LMS, regardless of age and sex. CT attenuation values varied widely, depending on age and sex, with mPVFS values tending to be higher than those of the PVFS and LMS, but the difference was not statistically significant.

Two factors determine cervical screw strength: bone quality and screw–bone contact area. Poor bone quality increases the risk of postoperative complications, such as screw loosening. Aging affects all organ systems, including the musculoskeletal system. As BMD decreases, the HU values of the cancellous bone in the cervical vertebrae also decline progressively with age [7]. This trend is consistent with our findings because the HU values in groups 3 and 4 were lower than those in groups 1 and 2, with decline particularly pronounced in women (i.e., groups 1 and 3). Lovecchio et al. [4] and Liang et al. [7] reported significant differences in HU values from C3 to C7, with the highest and lowest HU values observed in C4, and C6 and C7, respectively. The interobserver reliability was higher in the vertebral body than in the lateral mass [4]. Our study showed that the CT attenuation values of the vertebral body seemed lower than those of the lateral mass. However, vertebral body values had smaller variations than those in the lateral mass, which is largely consistent with the findings of a previous study [4].

With respect to factors influencing the screw–bone contact area, both diameter (thread size) and length (number of threads) are directly related to the available surface. Matsukawa et al. [8], in their finite element analysis of osteoporotic vertebrae, showed that larger diameters and longer screws significantly increase pullout strength. They also demonstrated that screw diameter had a greater impact on resistance to pullout and flexion–extension loads, whereas screw length more strongly influenced stability against lateral bending and axial rotation. These results emphasize that screw size should be tailored based on anatomical constraints, and the use of longer and larger screws, whenever feasible, provides optimal fixation strength and stability. Compared with the LMS and PVFS, the PS provides the highest fixation strength, such as pull-out strength and rotational resistance, making it a promising option for long-term stability. However, PS insertion is technically demanding, with a potential risk of vertebral artery and spinal cord injuries. Its application at C3–C6 is challenging in patients with a narrow pedicle. The LMS can be inserted with relative safety and minimal risk to the vertebral artery and spinal cord. However, its weak pull-out strength can cause postoperative screw loosening in patients with osteoporosis. Compared with PS and PVFS insertion, LMS insertion poses the risk of lateral mass fracture because the LMS is inserted outward. Recently, the PVFS has been reported as a new screw type that has gained increasing attention. One of its primary advantages is its superior biomechanical stability. Studies have demonstrated that the PVFS provides comparable or even greater pull-out strength than that of the LMS, thereby reducing the risk of hardware failure [2]. Additionally, the PVFS provides a safer trajectory by avoiding critical neurovascular structures, thereby decreasing the likelihood of complications, such as vertebral artery injury and nerve root impingement. Chen et al. [9] reported larger safety angle of PVFS insertion reduces vertebral artery injury risk than PS insertion, suggesting a relative advantage in terms of surgical safety. Maki et al. [2] and Shimizu et al. [3] have reported that intraoperative use of the posterior vertebral wall line on lateral fluoroscopy combined with a drill stopper adjusted to a length shorter than the preoperatively measured distance to the transverse foramen allowed safe placement of PVFS without a navigation system. In their clinical series, penetration into the transverse foramen was exceedingly rare, with no encountered vertebral artery injuries, supporting the reliability of this method as a practical surgical indicator [3]. Moreover, the PVFS can be used as a salvage technique for failed LMS insertions, proving effective in clinical practice. Another significant benefit is its minimally invasive nature, which may contribute to reduced postoperative pain, shorter hospital stays, and faster recovery times.

The PVFS has several limitations despite these advantages. The insertion technique poses a challenge, particularly during the initial learning phase because short monocortical screws are inserted into the cancellous bone. Another limitation is the lack of long-term clinical studies assessing PVFS durability and efficacy compared with well-established fixation methods. Shimizu et al. [3] reported a screw-loosening rate of 3.2% (3/94 screws), suggesting a potential risk of failure when using the PVFS to anchor screws at the upper or lower ends; however, the number of clinical studies remains limited.

By selecting the longest possible screws, surgeons can enhance fixation strength and avoid postoperative complications, such as loosening. Therefore, we introduced the mPVFS, a novel trajectory featuring increased length and surface area compared with the traditional LMS and PVFS. Specifically, the LMS trajectory is determined under lateral fluoroscopy parallel to the facet in the sagittal plane, with an outward tilt in the axial plane. By contrast, the PVFS trajectory is aligned parallel to the pedicle in the sagittal plane under fluoroscopy, with an inward tilt in the axial plane. Thus, the mPVFS combines these two approaches: it is parallel to the facet in the sagittal plane and tilted inward in the axial plane. This hybrid trajectory facilitates safe and reproducible insertion even without navigation systems, thereby offering a practical alternative. In clinical practice, the maximum screw length that avoids breaching the transverse foramen is preoperatively measured and selected, which is similar to the approach used in the standard PVFS technique. Although intraoperative CT navigation is not essential if screw length is determined in this manner, the incorporation of navigation systems can further enhance the safety and accuracy of mPVFS insertion. Additionally, mPVFS offers several advantages for cervical spine fixation. First, the increased screw length enhances pull-out strength by improving bone–screw engagement, indicating that the mPVFS is 2.5–3.0 mm longer than the PVFS and approximately 1 mm longer than the LMS. Its longer trajectory may contribute to its superior mechanical strength. Second, a larger surface area allows greater bone contact, thereby reducing the risk of loosening. The mPVFS area was 1.2 times larger than that of PVFS and 1.4 times larger than that of LMS in our study. Third, the BMD at the screw insertion site is also important. The PVFS has demonstrated better anchoring capabilities than LMS because it engaged with denser cortical bone. The mPVFS similarly targets a high-density region [5]. The HU values of mPVFS tended to be higher than those of the PVFS and LMS, although with not statistically significant differences. In summary, mPVFS can be inserted without requiring special instruments, navigation systems, or highly advanced surgical skills. Despite its simplicity, it offers the advantage of reducing the potential risk of neurovascular injury and is particularly suitable for surgeons who prefer using larger and longer screws whenever possible.

This study has several limitations. First, the mechanical strength of each screw type was not directly evaluated in cadaveric specimens, and PS was compared. Previous biomechanical studies showed that PVFS tend to provide greater pull-out strength than LMS, whereas Chen et al. [10] reported that PS (636 N) achieved nearly twice the pull-out resistance of PVFS (327 N) [2]. Furthermore, Kowalski et al. [11] demonstrated that the pull-out strength of PS is independent of lateral mass structure, with no difference between the standard and Abumi techniques. Because the mPVFS was designed with a 4.5-mm diameter, extended length (2.5–3.0 mm longer), and 1.2-fold larger contact area, it may achieve equal or slightly greater strength than PVFS, although unlikely to match PS. However, these assumptions remain speculative and represent the main limitation of this study, and further biomechanical studies are required. Second, this study was restricted to CT-based analysis without clinical application, thus no accumulation of clinical evidence is available. Because this is the first conceptual feasibility study of the mPVFS trajectory, our observations should be regarded as preliminary findings and exploratory in nature. Nevertheless, the possibility of inserting screws 2.5–3 mm longer than conventional PVFS may provide surgeons with a potentially reassuring alternative, although future case series and comparative studies are needed to confirm its actual clinical benefit. Third, data are lacking for the middle-aged population (36–84 years). Previous literature suggests a continuous decline in HU values with age across all age groups [7,12]. In clinical practice, safer screw trajectories are typically required in younger patients with good bone quality, whereas stronger fixation is often prioritized in elderly patients with poor bone quality. We deliberately compared these two extreme age groups to gain novel insights. The trend that the mPVFS achieved the greatest screw length and contact area was consistent in both young and elderly patients.

Conclusions

The mPVFS, which balances safety and strength, may have great clinical utility. This new screw trajectory can be easily applied in clinical settings without a navigation system. Although the mPVFS could not demonstrate fixation strength equivalent to that of the PS, its ability to safely accommodate screws that are 2.5–3.0 mm longer compared with the conventional PVFS, regardless of age and sex, represents a potential clinical advantage. Further biomechanical and clinical studies are required to validate the superiority of mPVFS over existing techniques.

Article information Continued