Normal range of clivoaxial angle in adults using flexion and extension cervical magnetic resonance imaging scans

Abstract

Objective

An abnormally decreased clivoaxial angle (CXA) is used during the clinical evaluation for corrective skull base surgery. Published normal ranges of CXA using x-ray, computed tomography, or magnetic resonance imaging (MRI) vary dramatically, especially with neck flexion or extension. The aim of this study was to use high-resolution MRI to determine the normal range of CXA in various neck positions using a reproducible measurement technique.

Methods

The CXA was measured in 10 healthy volunteers on sagittal T2 SPACE c-spine MRI in supine and prone positions and with the neck both neck and extended. CXA is strictly defined as the angle between a line along the inferior third of the dorsal clival cortex and a line from the superior/posterior cortex of the dens to the posterior/inferior corner of the C2 body. Statistical analysis was performed in all positions and included mean CXA, range, standard deviation (SD), inter-reader agreement, and group comparisons.

Results

The mean CXA overall was 156.92° (SD=4.23°; range 134–179°). The mean value for extension CXA was 169.20° (SD=5.81°), and the mean value for flexion CXA was 144.73° (SD=5.71°), the difference being statistically significant (p<0.0001) regardless of supine or prone position. Concordant correlations of reader measurements showed substantial agreement in the supine position at 0.96, with lower agreement in the prone position at 0.87.

Conclusions

We report normal ranges for CXA in various neck positions based on 3D T2-weighted MRI, using a reproducible measurement method. There was a significant difference in the CXA values between neck extended and neck flexed positions but not between supine and prone positions.

Introduction

The clivoaxial angle (CXA) is one of many measurements used to study the craniocervical relationship, and it is increasingly being used to evaluate brain-stem deformity and craniocervical instability, as well as craniocervical surgical realignment.¹ Abnormally low CXA values are associated with brain-stem compression or medullary kinking and are linked to neurological symptoms, including headache, neck pain, bulbar symptoms, and myelopathy.² It has been shown that angulation <130° leads to kinking of the brain stem and that correction to angles >145° improves the clinical symptoms associated with clivoaxial deformity.²,³ Recently, a clinical algorithm for the correction of kyphotic CXA has been proposed, where a normal lower limit of 135° acts as a threshold between surgical treatment and conservative management.⁴ In recent neurosurgical literature, surgical normalization of the CXA via a combination of decompression and stabilization techniques has been associated with improved clinical outcomes.³,^5–9

CXA values were traditionally measured via radiographs but are increasingly being determined from computed tomography (CT) and even magnetic resonance imaging (MRI) scans. Historically, normal measurements were reported as a range from 150° to 180°.^10–12 More recently, normal CXA values have been recorded on a clinical basis as a range from 145° to 160°.³,⁴ In addition, CXA values can change dramatically by 11–30° between flexion and extension of the neck.¹⁰,¹¹,¹³,¹⁴ Current normal values of CXA vary throughout the literature and may not be clearly defined.¹³,^15–17 Part of this wide variation may in part be due to CXA measurement methods and descriptions that are irreproducible or insufficient.

Given the progressive use of craniocervical fixation, the increasing reliance upon CXA in the evaluation of candidates for craniocervical fixation, and the change in normal CXA values upon neck position, establishing the normal range for CXA using relevant imaging modalities and various positions is essential. The purpose of this study was twofold: (a) to determine the normal CXA in various neck positions on high-resolution MRI, and (b) to provide a reproducible approach for CXA measurement. To our knowledge, there has been no similar report of MRI-based CXA measurements in the literature to date.

Methods

Method design

This was an Institutional Review Board (IRB)-approved, Health Insurance Portability and Accountability Act (HIPAA)-compliant, retrospective review of MRI cervical spine imaging performed on 10 healthy volunteers (images obtained for a prior IRB-approved, HIPAA-compliant study). The selection criterion included age >18 years. Exclusion criteria were: known cervical anomaly, history of cervical spine injury, known cervical disc herniations, cervical canal or neuroforaminal stenosis, or cervical spine surgery; acute or chronic neck pain; and any contraindication to MRI. The mean age of the subjects was 30.3 years (range 26–34 years). Six subjects were male, and four subjects were female. Of note, no incidental findings precluded inclusion into this study.

Each study was performed on a 3T MRI scanner (Magnetom Verio; Siemens, Erlangen, Germany) using a six-element body surface coil and a 16-element table spine coil and included only T2 sampling perfection with application-optimized contrasts using different flip angle evolution (SPACE) sagittal imaging. Sequences were obtained in four neck positions: prone flexion, prone extension, supine flexion, and supine extension. SPACE imaging parameters included: TE 123 ms; TR 1200 ms; flip angle 125°; partition thickness 0.9 mm; FOV 280 mm×280 mm; matrix size 320 × 317; receiver bandwidth 744 Hz; parallel imaging factor 2; and two excitations. Neck positions were maintained at the maximum amount of flexion or extension that the patient could tolerate and hold within the space limitations of the coil. In the supine position, the head was raised with pillows to allow the patient to hold a flexed position comfortably, while in the extension position, the pillows were instead placed under the lower neck/shoulder region. In the prone position, the patients maintained flexed and extended positions without extra support.

Image analysis

All measurements were made on a true midline sagittal MRI scan of the cervical spine. The midline image was determined via 3D manipulation of the isotropic data on a separate workstation (AquariusNET; TeraRecon, San Mateo, CA). In each volunteer, the CXA was strictly defined as the angle between a line drawn along the inferior third of the dorsal clival cortex and a second line drawn from the most superior/posterior-most cortex of the dens tip to the posterior/inferior-most corner of the C2 vertebral body (Figures 1 and 2). Measurements were performed independently by two individual fellowship-trained neuroradiologists (A.S. with seven years of experience and F.B. with 27 years of experience).

Figure 1.

Schematic demonstrating the method for clivoaxial measurement. Solid yellow lines indicate landmarks used in the technique including the inferior third dorsal clival cortex and the line from the most superior/posterior-most cortex of the dens tip to the posterior/inferior-most corner of the C2 vertebral body. Dotted lines are extensions of the solid lines to measure the subsequent angle that forms at their junction, the clivoaxial angle (CXA), in blue.

Figure 2.

A 32-year-old male normal volunteer. (a) Measurements taken from the midline sagittal cervical spine 3D T2 image in the neck flexed, supine position with clivoaxial angle measuring 148°. (b) Measurements taken from midline sagittal cervical spine 3D T2 image in the neck extended, supine position with the CXA measuring 168°.

Statistical analysis

Statistical analysis was performed on the CXA data in all neck positions. The descriptive statistics of continuous variables were summarized by their mean and standard deviation (SD). The agreement between two raters regarding continuous variable CXA was assessed by Lin’s concordance correlation coefficient. The reference for agreement included: <0.90 poor, 0.90–0.95 moderate, 0.95–0.99 substantial, and >0.99 almost perfect. The mean CXA differences were tested by paired t-test or Wilcoxon’s signed rank test, depending on the distribution of the data. When CXA was compared at extended and flexed positions, the differences of CXA were tested at those positions. A similar analysis was performed when the CXA was compared between supine and prone positions. Box plots were used to illustrate the distribution of raw data, and Bland–Altman plots were used to examine concordance. Analysis was performed using SAS v9.4 (SAS Institute, Cary, NC).

Results

CXA was measured on midsagittal views of the cervical spine 3D T2-weighted MRI of all 10 volunteers. The mean CXA in all positions was 156.92° (SD=4.23°; range 134.00–179.00°). The mean CXA in extension was 169.20° (SD=5.81°; range 152.00–179.00°). The mean CXA in flexion is 144.73° (SD=5.71°; range 134.00–163.00°). The mean CXA for each position is presented in Table 1 and Figure 3.

Table 1.

Measurements by raters and agreement between two raters by concordant correlation coefficient.

Variable	Rater I CXA°	Rater II CXA°	Overall CXA°	Difference	Concordant correlation (95% CI)
Supine extension	165.60 (7.44)	165.80 (8.75)	165.70 (8.08)	–0.20 (1.75)	0.98 (0.94–0.99)
Supine flexion	144.50 (5.50)	143.70 (6.15)	144.10 (5.77)	0.80 (1.75)	0.95 (0.82–0.98)
Prone extension	172.60 (5.85)	172.80 (5.07)	172.70 (5.31)	–0.20 (2.70)	0.89 (0.61–0.97)
Prone flexion	144.40 (5.83)	146.30 (7.39)	145.35 (6.47)	–1.90 (3.14)	0.85 (0.58–0.95)

Values are means (standard deviations) unless indicated otherwise.

CXA: clivoaxial angle; CI: confidence interval.

Figure 3.

Distribution of CXA measurements as a box plot at the various neck positions by rater.

There was a statistical difference (p<0.0001) in the mean CXA between extension and flexion regardless of whether the subject was prone or supine. The mean supine CXA was 165.70° (SD=8.08°) in extension and 144.10° (SD=5.77°) in flexion, while the mean prone CXA was 172.70° (SD=5.31°) in extension and 145.35° (SD=6.47°) in flexion. These data are presented in Tables 2 and 3 and Figure 4(a)–(d).

Table 2.

Test of difference between extension and flexion position.

Outcome variables	Position		p-Value
	Extension	Flexion
Supine	165.7° (8.08)	144.1° (5.77)	<0.0001
Prone	172.7° (5.31)	145.4° (6.47)	<0.0001

Table 3.

Test of difference between supine and extended position.

Outcome variables	Position		p-Value
	Supine	Prone
Extension CXA	165.7° (8.08)	172.7° (5.31)	0.0343
Flexion CXA	144.1° (5.77)	145.4° (6.47)	0.6539

Figure 4.

Bland–Altman plots for CXA measurements at the various neck positions: (a) supine flexion, (b) supine extension, (c) prone flexion, and (d) prone extension.

There was overall moderate agreement of CXA measurement between the two readers (Table 1 and Figure 5), with an overall concordant correlation (95% confidence interval (CI)) for all the CXA measurements of 0.91 (0.06), 95% CI 0.82–1.00. The correlation coefficients in the four various positions ranged from 0.85 to 0.98. In the supine position the agreements were stronger at 0.98 in extension and 0.95 in flexion. The agreements with prone measurements were less strong at 0.89 in extension and 0.85 in flexion.

Figure 5.

Distribution of the difference of CXA measurements as a box plot at the various neck positions by rater.

Discussion

Patients with craniocervical junctional abnormalities often have neurological symptoms, including headache, neck pain, bulbar symptoms, and myelopathy. Abnormally decreased CXA values in these patients are thought to result in stretched neurons within the brain stem and upper cord, which are then more vulnerable to further insult.¹⁸ The CXA is an important metric in the initial assessment for potential craniocervical junctional reduction/stabilization. Such surgical corrections are becoming more commonplace and include a variety of techniques: suboccipital decompression, reduction, fusion, and/or stabilization; open traction-reduction; posterior stabilization/fusion; and ventral decompression. CXA is used preoperatively to evaluate a surgical candidate clinically, to plan surgically, and to assess surgical outcomes postoperatively.^3–9,¹⁶,¹⁹,²⁰

Considering this growing dependency on the CXA in the neurosurgical realm, it is imperative to study this metric using currently relevant imaging modalities and standardized measurement techniques. Bollo et al. urged surgeons to evaluate craniocervical junction malformations using normal ranges based on CT instead of plain radiography.¹⁹ Hussain et al. concluded that the evaluation of the pre- and postoperative measurements should be based upon the same imaging modality, whether CT or MRI.¹⁷ In current practice, most patients acquire MRI cervical spine imaging as part of their routine workup. As such, we present normal ranges of CXA based upon high-resolution cervical spine MRI. We did not find any similar reports of MRI-based measurement of normal CXA values in healthy volunteers using a strictly defined technique for angle measurement.

Initial, widely used reports cite the normal CXA range as 150–180°, with values taken from x-ray imaging based on work by Smoker et al. and Van Guilder et al.¹⁰,¹¹ In 2015, however, an evaluation based upon 100 normal CT studies reported a lower CXA normal range of 138–168°, with 24% of patients demonstrating a CXA of <150°.¹⁵ Furthermore, Botelho and Ferreira demonstrated the normal range in 33 normal volunteers as 129–179° based upon MRI.¹⁶ These and other similar studies question the validity of the long-held normal value range of 150–180°. In 2016, a new parameter, termed the “clivodens angle,” was proposed as a diagnostic method for basilar invagination using CT imaging, with reports that it may be more accurate than CXA.²¹ This is further evidence that objective methods for evaluating basilar invagination and craniocervical alignment need better defining and updating—on more current modalities as opposed to early radiographic studies—especially since these values are affecting surgical candidacy. Our current study further validates the idea of a lower CXA normal range based on high-resolution MRI, with CXA in all positions at approximately 135–179°, being lower in neck flexion (135–163°) than in extension (150–179°). Our values are in line with the recent clinical statement of using 135° as the lower limit of normal when using CXA to decide between surgical correction and conservative treatment, which is different than the 150–180° range.⁴

The CXA is generally formed between the line connecting the dorsum sellae to the basion and the line drawn along the dorsum of the C2 vertebral body. Over the years, many have attempted to elucidate the normal range for the CXA via various modalities with varying results. Part of the variability likely originates from the inherent difficulty in pinpointing the exact anatomic landmarks from which to create the angle measurement. In reviewing the literature, the precise way in which the CXA measurement is carried out is often unapparent or inconsistent between studies. Most recently, a multidisciplinary consensus statement determined the CXA as the angle between the lower third of the dorsal clivus line and the dorsal C2 line,²² but this description still leaves room for interpretation. The detailed updated method we describe here for measuring the CXA was highly reproducible in our study, with substantial agreement based upon correlation coefficients. We urge future studies of the CXA to use a similar standardized approach in order to develop more reproducible results. We hope this description of a reproducible measurement technique can be an impactful contribution to assist in the assessment of interventions and individual care.

The correlation for supine measurements in our study on average demonstrate substantial value for strength of agreement at 0.96 (0.95–0.99 reflecting substantial value), which is most helpful, since spine imaging is practically always performed supine. We note, however, that the correlation values in our prone measurements are overall lower, with an average of 0.87 (poor value). Upon further subjective investigation of the prone images, the lower correlation coefficients are most likely due to imaging factors unique to prone positioning rather than the measurement method itself. These factors include less spatial and contrast resolution thought to be due to the decreased ability for the patient to remain still, increased breathing motion artifacts, and a further distance from the surface coil to the craniocervical region in the prone position.

The literature also demonstrates that CXA values change with neck flexion (decreases CXA by 9–11°) and extension (increases CXA by the same).¹³,¹⁴ Older literature reports variations up to 30° between flexion and extension.¹⁰,¹¹ The current study validates these findings in that there is statistical significance between CXA values obtained in neck extension versus neck flexion, even at low power. We also found that these CXA differences are not affected by whether the patient is supine or prone, with no statistical difference between supine or prone positioning. To our knowledge, no one has reported normal CXA values using 3D MRI during neck flexion and extension within the scanner, although a similar approach has been used to measure the posterior cervical subarachnoid space accurately.²³

Our study has several limitations. The sample of normal volunteers was small. However, we did find that the variation of CXA values between neck flexion and extension was statistically significant (p<0.0001). We also did not undertake a sex or gender sub-analysis, given the small sample size. Our patient population was young and may mimic that of the population who are evaluated for craniocervical instability or basilar invagination. However, future research based on a wider age range or older age range would be beneficial. Additionally, we have described above the inherent decrease in image quality with prone positioning due to motion and breathing artifacts, as well as target distance to the surface coil. Flexion/extension angles were likely limited by MRI coils and apparatus, and exact measurements of neck angle positions were not assessed. Further studies could include evaluation of abnormal subjects with basilar invagination in addition to healthy volunteers.

Conclusions

The use of CXA for evaluation of basilar invagination and craniocervical instability is becoming more commonplace, especially for patient selection in corrective skull base surgeries. We present the normal range of CXA in various neck positions on high-resolution MRI via a reproducible method for angle measurement using healthy volunteers. We urge the use of a reproducible measurement technique in future studies of the CXA, as well as in clinical work. We report high variability in CXA values between neck flexion and extension, but no significant difference in prone versus supine positioning. Our normal values are more in agreement with recently published surgical algorithms and could be used for more reliable evaluation of craniocervical alignment and evaluation of potential skull base surgical patients. Furthermore, our data are based on MRI, a modality that most of these patients obtain as part of their clinical evaluation.