Oblique-View Method Improves the Detection of cement leakage in Cement-Augmented Pedicle Screw Fixation

Abstract

Introduction

Cement-augmented pedicle screws (CAPS) are increasingly used in surgeries for osteoporotic vertebral fractures (OVFs); however, cement leakage (CL) remains a concern. Lateral-view (LV) fluoroscopy, the conventional method for monitoring cement injection, has limited sensitivity in detecting early leakage. The oblique-view (OV) method may provide improved detection, although its role in CAPS fixation remains unclear.

Methods

Data from 116 patients and 535 screws who underwent CAPS fixation for OVFs using the LV or OV method were retrospectively compared. Patient demographics, surgical factors, and CL incidence were analyzed. A novel computed tomography-based grading system for segmental vein CL was developed as follows: grade 0 (no CL), grade 1 (≤10 mm), grade 2 (>10 mm without exceeding the anterior border or ≤10 mm beyond the anterior border), and grade 3 (>10 mm beyond the anterior border).

Results

Baseline demographics and surgical factors were comparable between the groups. Overall, CL was observed in 56% of the patients and 20.9% of the screws. The screw-based incidence of CL was significantly lower in the OV group than in the LV group (15.0% vs. 22.9%, p=0.047), although the CL severity did not differ. Intraoperative CL detection was significantly more effective in the OV group than in the LV group (37.2% vs. 91.7%, p<0.01), including grade 1 and 2 cases, and 67% of leakages were identified within 0.4 mL of injection. No major complications, such as pulmonary embolism, occurred.

Conclusions

The OV method enhanced the intraoperative detection of CL during CAPS fixation for OVFs, facilitating the early recognition of minor leakage and reducing the overall leakage frequency without increasing the incidence of severe complications. These findings support the clinical utility of the OV method and the proposed grading system, which may guide safe cement injection in patients with osteoporosis.

Introduction

In Japan, an aging population has led to an increasing number of spinal surgeries^1-3). Additionally, the number of patients with osteoporosis has increased. Osteoporotic vertebral fractures (OVFs) are among the most common fragility fractures in older people and often lead to persistent pain, progressive deformity, and impaired quality of life. Surgical stabilization with pedicle screw fixation is indicated in selected cases, particularly in those with severe collapse unsuitable for vertebroplasty. In spinal fusion surgery for patients with osteoporosis, screw loosening and backout are major complications, and some cases require reoperation, including extension of fixation^4,5). The use of cement-augmented fenestrated pedicle screws (CAPS) as a new material for spinal fusion surgery reduces the incidence of loosening and increases fixation strength⁶⁾. In particular, its advantages in patients with osteoporosis include a reduced risk of reoperation compared to conventional pedicle screws and a lower risk of screw loosening^7-9). However, cement leakage (CL) remains a major concern, as it can lead to serious complications such as neural compression or pulmonary embolism. CL is common, ranging from 8.9% to >80%, depending on the detection method and patient population^10-13). Although most CLs are asymptomatic, cases of cement PE¹⁴⁾ and cement extending into the right atrium have also been reported¹⁵⁾. Intraoperative fluoroscopy is commonly employed to monitor cement injection; however, the lateral view (LV), which is the standard method, has limitations. In our previous study, the intraoperative incidence of CL with the LV was 25.6%¹⁶⁾. Accordingly, we developed an oblique-view (OV) method aimed at detecting CL at an early intraoperative stage and preventing serious complications. We hypothesized that the OV method would allow earlier detection of CL than the LV method, thereby reducing the incidence of severe leakage. This study aimed to compare the efficacy of LV and OV fluoroscopic methods in detecting the CL during CAPS fixation of OVF.

Materials and Methods

Patient selection

This multicenter retrospective study was approved by the institutional review board (Approval number: 5147). Data from 150 patients who underwent spinal surgery using CAPS at our hospital and five affiliated institutions between 2022 and 2025 were retrospectively reviewed. Among these patients, 116 who received OVF-related surgery were included in the analysis (Figure 1). The patients were divided into two groups: the LV group (83 patients, 375 screws), treated between April 2022 and September 2024, and the OV group (33 patients, 160 screws), treated between October 2024 and March 2025. The inclusion criteria were as follows: age >60 years, diagnosis of OVF, and spinal surgery using CAPS. The exclusion criteria were vertebroplasty alone, degenerative disease, spinal infection, spinal tumor, prior spinal surgery at the index level, and spinal deformity. The mean age was 81.0±6.4 years, and 69 patients (59.5%) were female.

Figure 1.

Patent flow chart.

CAPS: cement-augmented pedicle screw; LIF: lateral lumbar interbody fusion; LV: lateral view; OV: oblique view; OVF: osteoporotic vertebral fracture; PES: penetrating endplate screw; TLIF: transforaminal interbody fusion; VA: vertebral augmentation

Fracture level and surgery type for OVF

The distribution of the fractured vertebrae was as follows: T9, three patients; T10, one patient; T11, five patients; T12, 35 patients; L1, 34 patients; L2, 12 patients; L3, 13 patients; L4, 13 patients; and L5, three patients. The types of surgery included balloon kyphoplasty or vertebral body stenting using CAPS in 76 patients, corpectomy using CAPS in 15 patients, lateral lumbar interbody fusion using CAPS in three patients, transforaminal lumbar interbody fusion using CAPS in 12 patients, and CAPS only or CAPS with an endplate penetrating screw in 10 patients.

Surgical procedure

CAPS was used as reinforcement for OVFs in combination with posterior spinal fixation. They were generally placed at the upper and lower instrumented vertebrae or adjacent to the vertebroplasty site or cage. Pedicle screws from the Viper^Ⓡ or Verse^Ⓡ spinal system (DePuy Synthes, Raynham, MA, USA) were used, and screw length was determined preoperatively. Intraoperative fluoroscopy and postoperative computed tomography (CT) were used to evaluate CL. Postoperative CT scans were performed for all patients as part of the standard protocol for postoperative assessment, not specifically for research purposes.

Intraoperative fluoroscopy method for CL detection

1) LV method

In the LV group, 1.0 mL of cement was injected in two stages. In the first stage, 0.3 mL of cement was slowly injected into the screw of the contralateral vertebra (Figure 2a), followed by the injection of the same volume (0.3 mL) into the screw on the opposite side (Figure 2b). A second injection was performed in the same sequence until the target volume (0.7 mL) was achieved (Figure 2c). Cement was injected under lateral fluoroscopy with careful intermittent monitoring to detect CL. The procedure was immediately discontinued if leakage was observed. Finally, the distribution of the cement was confirmed using anteroposterior fluoroscopic imaging. Grade 3 CL was identified on intraoperative radiographs (Figure 2d) and confirmed on postoperative CT scans (Figure 2e).

Figure 2.

Lateral-view method.

Lateral view method for cement injection with a total of 1.0 mL: (a) 0.3 mL contralateral, (b) 0.3 mL opposite side, and (c) the remaining 0.7 mL was injected following the same order. (d) Intraoperative lateral fluoroscopy showing cement leakage (white arrow). (e) Postoperative CT scan demonstrating grade 3 cement leakage.

CT: computed tomography

2) OV method

In the OV method, 1.0 mL of cement was injected in three stages. First, 0.4 mL was slowly injected into the ipsilateral vertebra (Figure 3a). The OV was then confirmed (Figure 3b). Next, an additional 0.4 mL was injected into the contralateral vertebrae (Figure 3c). After reconfirming the OV (Figure 3d), a second 0.4 mL was injected into the contralateral vertebra under lateral fluoroscopy following the same sequence. In the third stage, the remaining 0.2 mL was injected into each screw in the same order. If grade 3 CL was detected on lateral fluoroscopy before the injected volume reached 0.4 mL, cement injection was immediately discontinued. Even when CL was not observed under the LV (Figure 4a), if CL was observed under the OV (Figure 4b), the injection sequence was modified so that the affected screw was injected last, and the cement volume was reduced accordingly. Postoperative CT revealed a grade 1 CL (Figure 4c).

Figure 3.

Oblique-view method.

Oblique view method for cement injection with a total of 1.0 mL: (a) 0.4 mL injected into the ipsilateral screw under lateral fluoroscopy, (b) oblique view to confirm absence of leakage, (c) contralateral injection checked with lateral fluoroscopy, and (d) contralateral oblique view to confirm absence of leakage.

Figure 4.

Representative case of grade 1 CL.

(a) CL was observed under the oblique view (black arrow), (b) Postoperative CT scan showing grade 1 CL.

CL, cement leakage; CT, computed tomography

Postoperative CT evaluation of CL

The degree of segmental vein CL on CT was graded from 0 to 3 according to leakage size and anterior vertebral border involvement. Grade 0 indicated no CL; grade 1 was CL ≤10 mm; grade 2 was CL >10 mm without exceeding the anterior border of the vertebral body or CL ≤10 mm with extension beyond the anterior border; and grade 3 suggested CL >10 mm with extension beyond the anterior border of the vertebral body (Figure 5). In addition, the cement volume and time to cement injection were assessed in relation to the CL.

Figure 5.

Grade of cement leakage into the segmental veins.

Statistical analysis

Baseline patient characteristics, including age, sex, body mass index, young adult mean, fracture level, time to surgery, and surgery-related factors, including the presence of diffuse idiopathic skeletal hyperostosis, osteoporosis fractur (OF) classification¹⁷⁾, and pre-existing vertebral fracture, were analyzed and compared between the LV and OV groups. Continuous variables were analyzed using Student's t-test or Mann-Whitney U test, as appropriate. Categorical variables were compared using the chi-squared test or Fisher's exact test. Statistical significance was set at p<0.05. All statistical analyses were performed using EZR (Saitama Medical Center, Jichi Medical University, Saitama, Japan), which is a graphical user interface for R (R Foundation for Statistical Computing, Vienna, Austria). It is a modified version of R Commander designed to add statistical functions frequently used in biostatistics¹⁸⁾.

Results

No significant differences in baseline patient characteristics, including age, sex, and body mass index, were observed between the LV and OV groups. In addition, no significant differences were observed between the two groups in the presence of diffuse idiopathic skeletal hyperostosis or OF classification (Table 1). However, fracture levels differed significantly between the groups, with lumbar fractures being more frequent in the OV group than in the LV group. Regarding surgery-related factors, no significant differences in the type of surgery or time to first cement injection were observed between the LV and OV groups; however, cement volumes were significantly greater in the OV group than in the LV group (Table 2). Overall, CL was observed in 56.0% of patients and 20.9% of screws. The screw-based incidence of CL was significantly lower in the OV group than in the LV group (15.0% vs. 22.9%, p=0.047). By contrast, no significant difference in CL grade was observed between the two groups (Table 3). Intraoperative fluoroscopy identified CL more effectively in the OV group than in the LV group (37.2% vs. 91.7%, p<0.01). Moreover, in the OV group, intraoperative CL was detected at a high rate, even in cases classified as grade 1 or grade 2 (p<0.01). In the OV group, CL was detected in 66.7% (16/24) of the cases within the first 0.4 mL of injection. No major complications, such as pulmonary embolism due to CL, were observed in any patient.

Table 1.

Baseline Patient Characteristics.

Variable	LV Group (n=83)	OV Group (n=33)	p-value
Age (years)	81.1±5.9	81.0±7.6	0.48
Sex, female [n, (%)]	49 (58)	20 (62.5)	0.61
BMI (kg/m2)	21.2±3.7	21.0±2.5	0.43
Lumbar YAM (%)	83.6±18.7	86.6±19.2	0.24
Total hip YAM (%)	72±13.5	69.1±11.2	0.06
Time to surgery (days)	56.4±51.3	47.4±49	0.20
Fracture level
Thoracic (T5-9) [n, (%)]	1 (1.2)	2 (5.9)	0.03*
Thoracolumbar (T10-L2) [n, (%)]	67 (78.8)	20 (58.8)
Lumbar (L3-5) [n, (%)]	17 (20.0)	12 (35.3)
DISH			1
Yes [n, (%)]	32 (38.6)	13 (39.4)
No [n, (%)]	51 (61.4)	20 (60.6)
OF classification
OF 2 [n, (%)]	6 (72.3)	1 (3.0)	0.87
OF 3 [n, (%)]	16 (19.3)	6 (18.2)
OF 4 [n, (%)]	51 (61.4)	23 (69.7)
OF 5[n, (%)]	10 (12.0)	3 (9.1)
Pre-existing VF, n (%)			0.30
Yes [n, (%)]	43 (51.8)	13 (41.9)
No [n, (%)]	40 (48.2)	20 (58.1)

Values are presented as mean±standard error.

BMI: body mass index; DISH: dish diffuse idiopathic skeletal hyperostosis; LV: lateral-view; OF: osteoporotic fracture; OV: oblique-view; VF: vertebral fracture; YAM: young adult mean

*p<0.05.

Table 2.

Surgery-Related Factors.

Variable	LV Group (n=83)	OV Group (n=33)	p-Value
Type of surgery with CAPS
VA (BKP or VBS) [n, (%)]	59 (71.1)	17 (51.5)	0.053
Corpectomy [n, (%)]	6 (7.2)	9 (27.3)
LIF [n, (%)]	2 (2.4)	1 (3.0)
TLIF [n, (%)]	8 (9.6)	4 (12.1)
CAPS only or with EPS [n, (%)]	8 (9.6)	2 (6.1)
Cement volume per screw (cc)	1.1±0.6	1.5±0.2	<0.01*
Time to cement injection (min)	12.7±1.5	12.5±0.7	0.541

Values are presented as mean±standard error.

BKP: balloon kyphoplasty; CAPS: cement augmented pedicle screw; EPS: endplate penetrating screw; LV: lateral-view; OV: oblique-view; VA: vertebral augmentation; VBS: vertebral body stenting; LIF: lateral lumbar interbody fusion; TLIF: transforaminal lumbar interbody

*p<0.05.

Table 3.

Comparison of CL between the LV Group and the OV Group.

Variable	LV Group (n=83, 375 screws)	OV Group (n=33, 160 screws)	p-Value
Total CL [n, (%)]	86 (22.9)	24 (15.0)	0.047*
Grade 1 [n, (%)]	43 (11.5)	12 (7.5)	0.126
Grade 2 [n, (%)]	27 (7.2)	5 (3.1)
Grade 3 [n, (%)]	16 (4.3)	7 (4.4)
Intra-op CL detection (per CL screw) [n / CL screws, (%)]	32/86 (37.2)	22/24 (91.7)	<0.01*
Grade 1 [n/CL screws, (%)]	9/43 (20.9)	10/12 (83.3)	<0.01*
Grade 2 [n/CL screws, (%)]	9/27 (33.3)	5/5 (100)
Grade 3 [n/CL screws, (%)]	14/16 (87.5)	7/7 (100)

Intraoperative detection rates are calculated per-CL screw.

CL: cement leakage; LV: lateral-view; OV: oblique-view

*p<0.05.

Discussion

Comparing the LV and OV methods using CAPS for OVF surgeries revealed that CL was observed in more than half of the patients, with a screw-based incidence of 20.9%. Importantly, the incidence of CL was significantly lower in the OV group than in the LV group, although CL severity did not differ significantly between the groups. Moreover, intraoperative fluoroscopy was more effective in detecting CL in the OV group, allowing the identification of leakage even in grade 1 and grade 2 cases, and frequently within 0.4 mL of injection. No major complications, such as pulmonary embolism, occurred in either group, supporting the overall safety of the procedures.

The superior performance of the OV method in detecting the CL can be explained by its ability to visualize the vertebral region more comprehensively than the LV alone. In the LV, small amounts of cement, especially those extending slightly beyond the anterior vertebral border, may be obscured by overlapping filled cement in the vertebral body. By contrast, the OV provides a clearer angle for identifying early leakage around the vertebra. This allows earlier recognition of grade 1 and grade 2 leakages, which might otherwise remain undetected in the LV. Moreover, the OV method enabled the detection of CL within the 0.4 mL of injection in >90% of cases, indicating that it can alert the surgeon at a very early stage of cement injection. Early detection is particularly valuable for preventing progression to higher-grade leakage and minimizing the risk of clinically significant complications. Overall, the OV method was not only more sensitive for detecting minor leakage but also contributed to the overall safety profile of CAPS fixation in OVFs.

In the OV group, over 90% of CL cases were detected within 0.4 mL of injection, emphasizing the clinical importance of early detection. Although intermittent fluoroscopy has been described as a strategy to confirm cement distribution¹⁹⁾, radiation exposure remains a concern. For example, cadaveric studies have demonstrated that lateral fluoroscopy can result in a nearly 30-fold greater radiation exposure than undertable fluoroscopy, with exposure increasing in proportion to fluoroscopy time²⁰⁾. Our study provides novel findings by demonstrating the utility of oblique fluoroscopy as a sensitive yet practical technique for intraoperative CL detection.

Traditionally, CL has been classified using Yeom's system, which categorizes leakage by location into type S (segmental veins), type B (basivertebral veins), and type C (paravertebral tissues via cortical defects)²¹⁾. However, this classification does not adequately reflect the degree of leakage, nor does it reliably differentiate between types S and C. Hence, we introduced a new CT-based grading system (grades 0-3), specifically for segmental vein CL, incorporating both leakage size and anterior border involvement. To prevent severe complications, the relationship with major vessels is particularly important. Using the anterior border as an indicator, we created a severity classification in which the severity increases even if the CL is small but exceeds the anterior border. Although we initially hypothesized that the OV method would reduce the incidence of grade 3 CL by enabling earlier detection and intervention, the overall frequency of CL decreased. However, the proportion of grade 3 cases remained comparable to that observed with the LV method. A plausible explanation is that even small volumes of cement may occasionally result in grade 3 leakage, making it difficult to prevent severe leakage entirely despite the enhanced detection capability of the OV method. Other contributing factors, such as bone quality, venous anatomy, and technical variations in screw placement, may also play a role and warrant further investigation^11,22).

Interestingly, 67% of CL cases occurred at injection volumes of <0.4 mL, indicating that leakage risk may be highest in the very early stages of injection. Previous studies have emphasized that the total or maximal cement volume is a major risk factor, with higher volumes being strongly associated with greater leakage rates. For example, Guo et al.¹⁰⁾ have reported an 81.7% leakage rate in 202 patients (950 screws) and demonstrated a significant positive association between injected volume and leakage risk. Similarly, Janssen et al.²³⁾ observed a leakage rate of approximately 73.3% and suggested that a greater cement volume increased the incidence. Our findings extend knowledge by showing that leakage can occur at significantly lower injection volumes than previously appreciated, highlighting the critical importance of meticulous fluoroscopic observation from the beginning of cement injection. In the present study, cement volume was greater in the OV group than in the LV group. Possible explanations include the higher proportion of middle-to-lower lumbar levels in the OV group, which generally require larger volumes for adequate augmentation²⁴⁾, as well as the tendency of the LV method to involve smaller volumes in the early phase of CAPS.

This study had some limitations. First, as this was a retrospective study, selection and information biases were possible. Second, the OV method was introduced later in the study period, and therefore, increased operator experience may have influenced the reduction in CL. A learning-curve effect in cement-handling procedures has been reported previously; for example, Yuan et al.²⁵⁾ demonstrated that complication rates in percutaneous kyphoplasty decreased as surgical experience increased. Thus, we cannot fully exclude the possibility that improved operator proficiency contributed to the lower leakage rate observed in the OV group. However, the primary clinical significance of the OV method lies in enhancing early intraoperative detection of leakage rather than merely reducing its incidence. Third, the CL grade classification was established independently in this study, and further consideration is needed regarding its reproducibility and clinical usefulness in other institutions. Finally, the number of OV cases was smaller than that of LV cases, which potentially limited the statistical power of our study.

Conclusions

The OV method proved more effective than the LV method for intraoperative detection of the CL during CAPS fixation for OVFs. Although the overall frequency of CL was lower in the OV group, the incidence of severe leakage (grade 3) was comparable between the two groups. The OV method facilitated the earlier detection of leakage, frequently within the first 0.4 mL of cement injection, and enabled the identification of even low-grade leakage. Thus, the OV method could enhance the sensitivity of intraoperative monitoring and may contribute to improving the safety of CAPS fixation without increasing the risk of major complications. Further prospective studies are warranted to validate these results, refine the proposed grading system, and establish standardized protocols for the clinical application of the oblique method.

Author Contributions: Conception and design, K.M., K.K.; provision of clinical materials or patients, K.M., K.K., F.A., M.H., N.Y.; data collection and assembly, K.K., T.Y., H.O., N.K., Y.I., K.T.; data analysis and interpretation, K.K., T.Y.; manuscript writing, K.M., K.K.; final approval of manuscript, All authors.

Conflicts of interest: The authors declare that there are no relevant conflicts of interest.

Ethical Approval and Informed Consent: This study was approved by the Institutional Review Board of Hyogo Medical University (Approval number: 5147). All patients provided comprehensive written informed consent at the time of admission for the use of their clinical data in future research. The study was conducted in accordance with the Declaration of Helsinki.