The cervical ligamentum flavum area: A new sensitive morphological parameter for identifying the cervical spinal stenosis

Abstract

Thickening of the cervical ligamentum flavum (CLF) has been considered as a main cause of cervical spinal stenosis (CSS). A previous study reported that cervical ligamentum flavum thickness (CLFT) is correlated with CSS. However, the whole hypertrophy is different from focal thickness. Therefore, to analyze hypertrophy of the CLF, we created a new morphological parameter, called the cervical ligamentum flavum area (CLFA). We hypothesized that the CLFA is an important morphological parameter in the diagnosis of CSS. CLF samples were acquired from 83 patients with CSS, and from 84 controls who underwent cervical magnetic resonance imaging (C-MRI). T2-weighted axial C-MRI images were acquired. We measured the CLFA and CLFT at the C6-C7 intervertebral level on C-MRI using appropriate image analysis software. The CLFA was measured as the cross-sectional area of the entire CLF at the level of C6-C7 stenosis. The CLFT was measured by drawing a straight line along the ligament side towards the spinal canal at the C6-C7 level. Mean CLFA was 25.24 ± 6.43 mm² in the control group and 45.34 ± 9.09 mm² in the CSS group. The average CLFT was 1.48 ± 0.28 mm in the control group and 2.09 ± 0.35 mm in the CSS group. CSS patients had significantly higher CLFA (P < .01) and CLFT (P < .01). For the validity of both CLFA and CLFT as predictors of CSS, a receiver operating characteristic curve analysis revealed an optimal cutoff point for the CLFA was 31.66 mm², a sensitivity of 92.8%, specificity of 88.4%, and an area under the curve of 0.97 (95% CI, 0.94–0.99). The optimal cut off-point of the CLFT was 1.79 mm, with a sensitivity of 83.5%, specificity of 84.5%, and an area under the curve of 0.92 (95% CI, 0.87–0.96). Both CLFT and CLFA were significantly related to CSS, but CLFA was the more sensitive measurement parameter. Therefore, to evaluate patients with CSS, treating physicians should test for CLFA.

1. Introduction

Cervical spinal stenosis (CSS) occurs when the vertebrae located in the cervical spine narrows significantly enough to compress the cervical nerve roots of the spinal cord. Although CSS is found in 26% of asymptomatic older adults, it frequently causes upper extremity pain, neck, and shoulder pain.^[1–3] Patients with CSS-induced pain may have reduced functional ability and quality of life.^[4,5] Cervical magnetic resonance imaging (C-MRI) is the most sensitive technique for showing structures of soft tissue, such as intervertebral disc and ligamentum flavum. Accurate C-MRI diagnosis of CSS is important to determine appropriate management and rule out the possibility of peripheral neuropathy.^[6,7] CSS is usually caused by age-related cervical degenerative changes, including swelling or herniation of intervertebral disc, ossification and osteophyte of the posterior longitudinal ligament.^[8,9] The thickness of the cervical ligamentum flavum (CLF) has also been considered to be a major cause of CSS.^[10] Thickening of the CLF can compress the nerve roots and dural sac, resulting in the above symptoms.^[11] However, the overall hypertophy of the CLF is somewhat different from the focal thickness. Measurement errors can occur whenever the treating physician only checks cervical ligamentum flavum thickness (CLFT). Therefore, we made a new morphological parameter called CLF area (CLFA) to evaluate the hypertrophy of the entire CLF. Unlike the CLFT, the CLFA measures the cross-sectional area of the entire CLF. The CLFA has not yet been analyzed for its association with CCS. We hypothesized that CLFA is a key morphological parameter for CSS diagnosis. Therefore, we compared CLFA and CLFT between CSS patients and normal controls via C-MRI.

2. Materials and methods

2.1. Patients

This study protocol was approved and reviewed by the Institutional Review Board of International St. Mary’s Hospital, College of Medicine, Catholic Kwandong University. (IRB protocol number: IS18RISI0016). Due to the retrospective nature of this protocol, we skipped the requirement for informed consent. Patients who underwent C-MRI from June 2017 to August 2018 and were diagnosed with CSS were included. Patients over 50 years of age were recruited only if they had clinical symptoms suitable for CSS. The inclusion criteria were defined as: (1) neck and shoulder pain with or without symptom spreading to the upper extremities resulting from CSS, (2) weakness of the fingers or hands and tingling or loss of sensation in the upper extremities, (3) the most stenosis at C6–C7, and (4) C-MRI performed within 6 months of the diagnosis that was available for review.

Exclusion criteria were as follows: (1) pathological changes in the cervical spinal cord, (2) spinal infection, (3) past surgical history of cervical spine, (4) injury of cervical spinal cord, (5) congenital cervical spine defects, (6) space occupying lesions such as cysts, tumors, and syringomyelia.

We enrolled a total of 97 patients after the CSS diagnosis was confirmed by 2 experienced, board-certified neuroradiologists. In the CSS group, there were 49 (50.05%) males and 48 (49.95%) females with a mean age of 58.96 ± 7.06 years (range: 50–81 years; Table 1). To compare the CLFA and CLFT between patients with and without CSS, we enrolled a group of control patients who had undergone C-MRI as part of routine medical examinations and who had no CSS-related symptoms. The control group consisted of 97 participants (49 males [50.05%] and 48 females [49.95%]) with a mean age of 56.97 ± 6.29 years (range: 50–79 years; Table 1). We examined the CLFA and CLFT in the control group at the C6–C7 facet joint level.

Table 1.

Comparison of the characteristics of control and CSS group.

Variable	Control group n = 97	CSS group n = 97	Statistical significance
Gender (male/female)	49/48	49/48	NS
Age (yrs) CLFT (mm)	56.97 ± 6.29 1.48 ± 0.28	58.96 ± 7.06 2.09 ± 0.35	NS P < .001
CLFA (mm2)	25.24 ± 6.43	45.34 ± 9.09	P < .001

Data represent the mean ± standard deviation (SD) or the numbers of patients.

CLFA = cervical ligament flavum area, CLFT = cervical ligament flavum thickness, CSS = cervical spinal stenosis, NS = not statistically significant (P > .05).

2.2. Imaging parameters

The C-MRI examinations had been performed with 3T Avanto (Siemens Medical Systems, Inc) with 3T Philips Achieva scanner. C spine axial T2-weighted images with 4 millimeter thick slices had been obtained. The following other parameters were used as well: Zoom: 301.25%, 0.4-mm intersection gap, 1180-ms/15-ms repetition time/echo time, 160 × 160 cm field of view, 320 × 224 matrix, and 15 echo train length.

2.3. Image analysis

Axial T2-weighted MR images of the cervical spine were obtained at the facet joint level in individual patients. Both CLFA and CLFT of the cervical facet joints were measured by MRI using image analysis system (INFINITT PACS Healthcare, Seoul, Republic of Korea).

The CLFA was measured as the cross-sectional area of the entire CLF at the level of C6–C7 stenosis (Fig. 1A).

Figure 1.

The cervical ligamentum flavum thickness (CLFT) was measured by drawing a straight line along the ligament side towards the spinal canal and towards the laminar side of the cervical ligament curvature and measuring the thickest point at the C6–C7 level (A). The cervical ligamentum flavum area (CLFA) was measured as the cross-sectional area of the entire cervical ligamentum flavum at the level of C6–C7 stenosis (B).

The CLFT was measured by drawing a line along the ligament side towards the spinal canal and towards the laminar side of the cervical ligament curvature and measuring the thickest point at the C6–C7 level (Fig. 1B).

2.4. Statistical analysis

Independent t-tests were used to compare CLFA and CLFT between control and CSS groups. Chi-square analysis was used to examine associations between gender. Receiver operating characteristic (ROC) analysis was done with SPSS version 22. Each cutoff values acquired by the ROC curve analysis were evaluated in a logistic regression model as independent variables. P-values <.05 is considered as sufficient evidence that the result is statistically significant.

3. Results

Chi-square statistical test indicated no significant difference in the gender distribution.

Mean CLFA was 25.24 ± 6.43 mm² in the control group and 45.34 ± 9.09 mm² in the CSS group. The average CLFT was 1.48 ± 0.28 mm in the control group and 2.09 ± 0.35 mm in the CSS group. CSS patients had significantly higher CLFA (P < .01) and CLFT (P < .01) than controls (Table 1). For the validity of both CLFA and CLFT as predictors of CSS, ROC curve analysis showed that the most suitable cut off point for CLFA was 31.66 mm², with a sensitivity of 92.8%, specificity of 88.4% (Table 2), and an area under the ROC curve (AUC) of 0.97 (95% CI, 0.94–0.99) (Fig. 2). The optimal cut off point for the CLFT was 1.72 mm, with a sensitivity of 83.5%, specificity of 84.5% (Table 3), and an AUC of 0.92 (95% CI, 0.87–0.96) (Fig. 2).

Table 2.

Sensitivity and specificity of each cutoff point of the CLFT.

CLFT (mm)	Sensitivity (%)	Specificity (%)
0.89	100	1.2
1.49	95.9	46.4
1.66	94.8	77.3
1.79*	83.5	84.5
1.89	72.2	91.8
2.00	55.7	94.8

CLFT = cervical ligament flavum thickness.

^*The best cutoff point on the receiver operating characteristic (ROC) curve.

Figure 2.

Receiver operating characteristic curve (ROC) of the cervical ligamentum flavum thickness (CLFT) and cervical ligamentum flavum area (CLFA) for prediction of cervical spinal stenosis. The optimal cut off-point of the CLFT was 1.79 mm, with a sensitivity of 83.5%, specificity of 84.5%, and AUC of .92. The ROC curve analysis revealed an optimal cutoff point for the CLFA was 31.66 mm², a sensitivity of 92.8%, specificity of 88.4%, and AUC of .97. AUC = area under the curve.

Table 3.

Sensitivity and specificity of each cutoff point of the CLFA.

CLFA (mm2)	Sensitivity (%)	Specificity (%)
14.29	100	1.2
27.48	97.9	67.0
30.84	95.9	79.4
31.66*	92.8	88.4
36.57	85.6	96.4
41.11	63.9	97.6

^*The best cutoff point on the receiver operating characteristic (ROC) curve; CLFA = cervical ligament flavum area.

4. Discussion

CSS can lead to difficulties with manual dexterity, imbalance and occasionally in the neck pain.^[12] It results in reactive hypertrophy of endplate osteophytes and uncal, ligamentous structures, with swelling of the cervical disc.^[2,13–15] CSS also describes osteoarthritic changes in the spine, including osteophytes, spondylosis and facet joint disease, in addition to ligament abnormalities such as ossification or hypertrophy of the CLF.^[16–19]

Previous research has analyzed that cervical dural sac-thickness, the cervical canal’s diameter, cervical lateral masses and cervical pedicles.^[20] Kwon et al reported anatomical differences in cervical dura mater thickness according to spine height and age.^[21] Prasad et al measured the diameter of the cervical spinal canal as a threshold indicator or anteroposterior straight-line distance for CSS.^[7]

Matsuura et al investigated medial and lateral diameters in the middle pedicle level of the cervical spinal canal and determined that the ratio of these diameters was predictive of spinal cord injuries.^[22] Chaput et al have insisted that the morphology of facet joint can predict cervical degenerative spondylolisthesis.^[23]

Degenerative changes of the CLF are a common finding associated with the aging process of the intervertebral disc. The main categories of CLF degeneration are ossification, calcification, and hypertrophy.^[17] To analyze the hypertrophy of CLF, a new morphological parameter called CLFA was investigated. We hypothesized that the CLFA is an important morphological parameter in CSS diagnosis.

The CLF significantly contributes to moment resistance, especially in flexion.^[24,25] There are many previous reports of ligamentum flavum hypertrophy in the thoracolumbar spine.^[26,27] We hypothesized that the same histopathological and biomechanical changes also occur in the cervical spine. Of course, previous studies have shown that CSS and CLFT are closely related.^[10,11] However, the overall enlargement of the CLF is somewhat different from the focal thickness. Measurement errors can occur at any time if the treating physician only tests the CLFT. Therefore, we devised a new morphological parameter called CLFA to evaluate the hypertrophy of the entire CLF.

In this study, CLFA showed a sensitivity of 92.8%, specificity of 88.4%, and AUC of 0.97 (95% CI, 0.94–0.99) in predicting CSS. On the other hand, the CLFT showed a sensitivity of 83.5%, specificity of 84.5%, and AUC of 0.92 (95% CI 0.87–0.96). These results suggest that CLFA predicts CSS better than CLFT. In this study, we analyzed both the CLFA and CLFT from C-MRI images. Because, the C-MRI studies are the most important in the diagnosis of degenerative disease of the CSS and in the highly detectable hypertrophy of CLF.^[7]

The current study has several limitations. Anatomically, degenerative CSS can divide the central, lateral recess, and foraminal.^[28] However, we focused on central canal only. Because, in contrast to lumbar spinal anatomy, CLF did not enter the neural foramen in the cervical spine.^[29] Second, several different methods to evaluate CSS, such as the cord-canal-area ratio, the space available for the cord, the transverse spinal cord area, the sagittal canal diameter, and the cord compression ratio (ratio transverse: sagittal cord diameter) have been shown to be effective at discriminating CSS.^[30–40] However, since we only investigated measurements of CLFA and CLFT, there may be some limitations regarding the measurement of other morphological changes. Third, there may be technical errors associated with measuring CLFA and CLFT in C-MRI. Although these morphological parameters were measured on T2-weighted axial images that best represent CLF at the facet joint level, the T2-weighted axial images analyzed for parameter measurements may be inconsistent due to differences in cut angle or cut plane. Due to technical causes, involuntary movements or individual anatomical deformations in C-MRI. Also, the 4.0 mm axial T2-weighted C-MR image slices are thicker than ideal slices. Fourth, this study has a retrospective nature. A prospective study is needed to validate the results. Fifth, we measured only the C6–C7 level in this study. Despite these limitations, this is the first study documenting the association between CLFA and CSS. We hope that these results provide a new reliable evaluation tool for CSS.

5. Conclusions

Both CLFT and CLFA were significantly related to CSS, but CLFA was a more sensitive measurement parameter for CSS than CLFT. We identified the best cutoff value for CLFA as 31.66 mm² with 92.8% sensitivity and 88.4% specificity. When evaluating patients with CSS, physicians should evaluate CLFA carefully.

Acknowledgments

All authors thank the International St. Mary’s Hospital.

Author contributions

Conceptualization: Jae Ni Jang, Young Uk Kim.

Data curation: Jae Ni Jang, Young Uk Kim.

Formal analysis: Young Uk Kim.

Funding acquisition: So Yeon Kim.

Investigation: Jae Ni Jang, Young-Soon Choi, Keum Nae Kang, Young Uk Kim.

Methodology: Jae Ni Jang, Young-Soon Choi, Sukhee Park, Young Uk Kim.

Project administration: Sukhee Park.

Resources: Sukhee Park, Jungmin Yi, Young Uk Kim.

Software: Young Uk Kim.

Supervision: Young Uk Kim.

Validation: Yumin Song, Jae Won Kim, Young Uk Kim.

Visualization: Yumin Song, Young Uk Kim.

Writing – review & editing: So Yeon Kim.

Writing – original draft: Jungmin Yi, Young Uk Kim.