The Spine Journal

-

(2014)

-

Clinical Study

Increased spinal cord movements in cervical spondylotic myelopathy Irene M. Vavasour, PhDa,*, Sandra M. Meyers, BScb, Erin L. MacMillan, PhDc, Burkhard M€adler, PhDd, David K.B. Li, MDa, Alexander Rauscher, PhDa,e, Talia Vertinsky, MDf, Vic Venu, MDf, Alex L. MacKay, D Phila,b, Armin Curt, MDg,h a Department of Radiology, 2211 Wesbrook Mall, University of British Columbia, Vancouver, BC, Canada, V6T 2B5 Department of Physics and Astronomy, 6224 Agricultural Rd, University of British Columbia, Vancouver, BC, Canada, V6T 1Z1 c Department of Medicine, 2775 Laurel St, 10th Floor, Vancouver, BC, Canada, V5Z 1M9 d Department of Neurosurgery, Sigmund-Freud-Str. 25, Univerity of Bonn, Germany, 53105 e UBC MRI Research Centre, M10 Purdy Pavilion, 2111 Wesbrook Mall, University of British Columbia, Vancouver, BC, Canada, V6T 2B5 f Department of Radiology, 855 W 12th Ave, Vancouver General Hospital, Vancouver, BC, Canada, V5Z 4E3 g Spinal Cord Injury Center, Forchstrasse 340, University of Zurich, CH-8008 Zurich, Switzerland h International Collaboration on Repair Discoveries (ICORD), 818 West 10th Ave, University of British Columbia, Vancouver, BC, Canada, V5Z 1M9 b

Received 22 July 2013; revised 10 December 2013; accepted 17 January 2014

Abstract

BACKGROUND CONTEXT: Magnetic resonance imaging (MRI) is a very useful diagnostic test for cervical spondylotic myelopathy (CSM) because it can identify degenerative changes within the spinal cord (SC), disclose the extent, localization, and the kind of SC compression, and help rule out other SC disorders. However, the relationships between changes in cerebrospinal fluid (CSF) flow, cord motion, the extent and severity of spinal canal stenosis, and the development of CSM symptoms are not well understood. PURPOSE: To evaluate if changes in the velocity of CSF and SC movements provide additional insight into the pathophysiological mechanisms underlying CSM beyond MRI observations of cord compression. STUDY DESIGN: Prospective radiologic study of recruited patients. PATIENT SAMPLE: Thirteen CSM subjects and 15 age and gender matched controls. OUTCOME MEASURES: Magnetic resonance imaging measures included CSF and SC movement. Cervical cord condition was assessed by the Japanese Orthopaedic Association (JOA) score, compression ratio (CR), and somatosensory evoked potentials (SSEPs) of the tibial and ulnar nerves. METHODS: Phase-contrast imaging at the level of stenosis for patients and at C5 for controls and T2-weighted images were compared with clinical findings. RESULTS: Cerebrospinal fluid velocity was significantly reduced in CSM subjects as compared with controls and was related to cord CR. Changes in CSF velocity and cord compression were not correlated with clinical measures (JOA scores, SSEP) or the presence of T2 hyperintensities. Spinal cord movements, that is, cord displacement and velocity in the craniocaudal axis, were increased in CSM patients. Increased SC movements (ie, total cord displacement) both in the controls and CSM subjects were associated with altered spinal conduction as assessed by SSEP. CONCLUSIONS: This study revealed rather unexpected increased cord movements in the craniocaudal axis in CSM patients that may contribute to myelopathic deteriorations in combination with spinal canal compression. Understanding the relevance of cord movements with

FDA device/drug status: Not applicable. Author disclosures: IMV: Grant: Cervical Spine Research Society (B, CAD); Research Support Investigator Salary: MS Society of Canada (D/year); Research Support Staff/Materials: MS Society of Canada (D/year). SMM: Fellowship Support: NSERC (C). ELM: Grant: Cervical Spine Research Society (E, USD, Paid directly to institution), Michael Smith Foundation for Health Research (C, CAD); Fellowship Support: Multiple Sclerosis Society of Canada (D). BM: Nothing to disclose. DKBL:Nothing to disclose. AR: Nothing to disclose. TV: Nothing to disclose. VV: Nothing to disclose. ALM: Nothing to disclose. AC: Nothing to disclose. 1529-9430/$ - see front matter Ó 2014 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.spinee.2014.01.036

The disclosure key can be found on the Table of Contents and at www. TheSpineJournalOnline.com. There are no conflicts of interest with any of the authors. This study was funded by the Cervical Spine Research Society and ELM was supported by the Michael Smith Foundation for Health Research. * Corresponding author. UBC MRI Research Centre, Room M10, Purdy Pavilion, 2221 Wesbrook Mall, Vancouver BC V6T 2B5, Canada. Tel.: 1-604-822-0357; fax: 1-604-827-3339. E-mail address: [email protected] (I.M. Vavasour)

2

I.M. Vavasour et al. / The Spine Journal

-

(2014)

-

respect to supporting the clinical CSM diagnosis or disease monitoring requires further long-term follow-up studies. Ó 2014 Elsevier Inc. All rights reserved. Keywords:

Magnetic resonance imaging; Phase contrast; Somatosensory evoked potentials; Cervical spondylotic myelopathy; Spinal cord movement; Stenosis

Introduction Cervical spondylotic myelopathy (CSM) is the leading cause of spinal cord (SC) dysfunction in people older than 55 years in North America, with the most common level of abnormality at C5–C6 [1,2]. Diagnosis of CSM is primarily based on clinical symptoms that can include clumsy hands and numbness of feet, increased reflexes, and gait disturbance. Magnetic resonance imaging (MRI) is the most useful diagnostic imaging test for CSM because it can confirm degenerative changes within the SC, disclose the extent, localization, and kind of SC compression, and help rule out other SC disorders (eg, tumors, syringomyelia, or cord malformation). In CSM, cord damage is typically associated with a narrowing of the spinal canal, called stenosis, which can also alter the flow of cerebrospinal fluid (CSF) through the canal [3,4]. The area of stenosis is often characterized by T2 hyperintensities on conventional MRI [5]. However, the relationships between changes in CSF flow, cord motion, the extent and severity of spinal canal stenosis, and the development of CSM symptoms are not well understood. Previous studies of CSF flow using cine phase-contrast MRI with cardiac gating show a pulsatile motion in the cranial to caudal direction synchronized to the blood pulse flow [6]. The amplitude of the pulsatile CSF flow has been found to be correlated with the severity of the myelopathy [4,7], with flow amplitude decreasing because of stenosis. Spinal cord motion has been first reported in the 1980s using intraoperative ultrasonography, where in some cases SC motion was rated as increased in subjects with SC compression [8], while studies of cord motion in controls could not be performed. The development of phase imaging for measuring motion with MRI allowed SC motion to be characterized in healthy controls [9,10]. Shortly after cardiac systole, the SC is reported to initially move in a caudal direction and then recover in the cranial direction [9,10]. The maximum velocity of the SC was found to range from 7 to 13 mm/s, but differed between individuals. Previous measurements of maximal cord displacement in healthy controls showed values ranging from 0.22 [11] to 0.5 mm [10]. Complementary to clinical testing, recordings of somatosensory evoked potentials (SSEPs) are highly sensitive in the diagnosis of SC disorders and provide measures of cord function, that is, conductivity within dorsal columns that cannot be determined by clinical means [12]. Specifically in disorders that frequently present changes in sensory function (such as numbness or paraesthesia), objective measures of sensory function independent of patient rating can be of value in the appreciation of neurologic impairment [13].

The purpose of this study was to compare velocity patterns of CSF and cord movements between CSM subjects and controls. It was hypothesized that degenerative changes of the cervical spine with increased spinal canal stenosis will not only compromise CSF flow but will also affect cord movements, whereas the interaction between these changes and how they relate to the development of CSM might be less predictable. Materials and methods Subjects Thirteen subjects (mean age, 62 years, range 50–77 years) with diagnosed CSM and 15 age and gender matched controls (mean age, 58 years, range 50–73 years) were recruited. Both groups underwent MRI and electrophysiological and clinical evaluations. The examinations of Japanese Orthopaedic Association (JOA), stenosis grading, and SSEP were performed by independent examiners that were blinded to the results in other domains. Written informed consent was obtained with approval from the Clinical Research Ethics Board of our institution. Magnetic resonance imaging and analysis Magnetic resonance imaging scans were performed with a phased array spine coil, using only the first four coil elements in the vicinity of the cervical spine, on a Philips 3.0T Achieva system (Philips Healthcare, Best, The Netherlands). All subjects were scanned with a localizer and sagittal T2-weighted imaging (T2WI) sequence (repetition time (TR)/echo time (TE)53,314/120 ms, 11 3 mm thick slices) to orient axial slices perpendicular to the SC, as illustrated in Fig. 1; an additional axial higher resolution T2WI sequence (TR/TE51,500/12 ms, field of view (FOV) 150
112
67 mm3, reconstructed matrix 336
250
12) was acquired for improved grey and white matter contrast and to identify any T2 hyperintense regions. Velocity imaging was performed using a threedimensional phase contrast sequence, retrospectively gated to the peripheral pulse (TR/TE513/8.6 ms, flip angle510 , two averages, FOV 140
140
25 mm3, partial parallel imaging acceleration factor52, reconstructed matrix 256
256
5, velocity encoding parameter55 cm/s in the craniocaudal direction). The single five slice stack was oriented perpendicular to the SC at the level of the stenosis for CSM subjects and at the C5 level for controls. If a subject had more than one stenosis within the cervical spine then the stack was centered on the stenosis closest to C5. Phase images are defined in the interval between 0 and 2p. Quantities that influence the phase of the MRI signal, such as flow or magnetic susceptibility, are mapped into this

I.M. Vavasour et al. / The Spine Journal

-

(2014)

-

3

Fig. 1. T2-weighted images from a control (left column) and CSM subjects (right column). The axial images are taken around the C5 level and shown by the lines in A and E. (G) The lack of CSF space and compression of the cord can be seen in the CSM subject at the stenosis, (F) but the slice above the stenosis shows more normal characteristics. (D and H) The CSF and cord velocities represent an average over five slices centered at the C5 level. CSM, cervical spondylotic myelopathy; CSF, cerebrospinal fluid.

interval, even when the true value lies outside this range. That is, when the phase exceeds 2p, it is wrapped back to 0. Various algorithms have been developed to remove the resulting phase wraps. We used a combination of FUn [14] and an in-house program to unwrap the phase images. In short, the FUn algorithm starts in areas of high signal to noise and flat-phase topography. New pixels are unwrapped by computing a prediction from already unwrapped adjacent pixels. If the prediction disagrees with the phase of that

pixel by more than a noise-threshold, a multiple of 2p is added to that pixel. Therefore, the algorithm attempts to remove large changes in phase between adjacent pixels. Maps of mean and standard deviation (SD) of absolute value of the velocity over all cardiac time points were calculated for each pixel from the unwrapped phase images. Velocity noise level, taken to be the mean absolute velocity within a region of interest (ROI) in stationary background tissue, was determined to be 0.08 cm/s. Regions of interest

4

I.M. Vavasour et al. / The Spine Journal

-

(2014)

-

Table 1 Age, SSEP classification, JOA score, compression ratio and status of T2 hyperintensities for CSM subjects and controls

CSM subjects

Controls

S no. Age (y) SSEP* JOAy CRz

Hyperintensitiesx

1 2 3 4 5 6 7 8 9 10 11 12 13 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

0/0/0 1/0/0 0/1/0 1/1/1 1/1/1 0/1/0 0/1/1 1/2/1 1/1/1 0/1/0 0/2/0 0/1/1 1/1/0 0/0/0 0/0/0 0/0/0 0/0/0 0/0/0 0/0/0 0/0/0 0/0/0 0/0/0 0/0/0 0/0/0 0/0/0 0/0/0 0/0/0 0/0/0

58 76 61 61 50 62 56 52 77 69 65 68 46 60 58 58 50 50 57 59 53 58 65 51 53 53 58 50

N A A A N N A N A A A N N A N A A A N N N N N A A N A N

16.5 11 15.5 13.5 17 9.5 17 16 14 17 15 17 17 17 17 17 17 17 17 17 17 17 17 17 15.5 17 17 17

0.39 0.32 0.28 0.21 0.35 0.33 0.33 0.18 0.37 0.30 0.28 0.37 0.23 0.43 0.40 0.34 0.50 0.37 0.37 0.50 0.52 0.41 0.55 0.44 0.53 0.38 0.46 0.69

CSM, cervical spondylotic myelopathy; SSEP, somatosensory evoked potential; JOA, Japanese Orthopedic Association; CR, compression ratio; N, normal; A, abnormal. * Somatosensory evoked potentials divided into N and A. y Japanese Orthopedic Association score. z Compression ratio taken at the level of stenosis for CSM subjects and at C5 for controls. x Hyperintensities5hyperintensities on T2-weighted images with the location labeled as above/at/below the level of stenosis and with the number indicating the intensity (05no signal abnormality, 15ill-defined T2 hyperintensity, 25well-defined T2 hyperintensity).

were drawn around the canal and SC for each slice on magnitude images, with reference to SD maps and higher resolution T2WI to help distinguish structures. The CSF ROI was obtained by removing the SC ROI from the canal ROI. Mean absolute velocity throughout all cardiac time points was calculated within each ROI for both CSF and SC and then averaged over all slices. Mean velocity curves were piecewise integrated to obtain displacement curves, and cord displacement was plotted over time for each subject. Because the cord should begin and end each cardiac cycle with no net displacement, and phase is a relative measure, cord velocity curves were adjusted by adding or subtracting a uniform velocity from each cardiac time point to make the net displacement over a cardiac cycle equal to zero; the same velocity was added or subtracted from CSF and SC velocity curves. Maximum displacement and total cord displacement (absolute area under the displacement curve) were determined.

Fig. 2. Mean absolute CSF velocity averaged over the entire heart cycle and compression ratio for control subjects and subject with CSM. The subjects are subdivided into those with normal (N) and with abnormal (A) somatosensory evoked potential values. Significant differences are shown (*p!.05, **p!.001, ***p!.0001). Error bars indicate standard error. The compression ratio was measured at the level of stenosis for subjects with CSM and at C5 for control subjects. CSF, cerebrospinal fluid; CSM, cervical spondylotic myelopathy.

Electrophysiological study and analysis Tibial and ulnar nerve SSEPs were measured in the right and left sides for all participants [12]. Changes in amplitude, latency (corrected for body height), and configuration were calculated to score SSEP recordings and classify the recordings as normal or abnormal by an independent neurologist. If any of the measured SSEPs were considered abnormal, then the subject (either control or CSM subject) was classified as having an abnormal SSEP condition regardless of CSM diagnosis. JOA scores The severity of CSM was ranked using the JOA scoring system [15], which is a validated and internationally recognized score for the clinical appreciation of CSM symptoms [16]. Stenosis grading The compression ratio (CR: maximum anteroposterior diameter of the SC over maximum right/left diameter [17]) of the cord was measured at the stenosis for CSM subjects and at C5 for controls. T2 hyperintensities on the T2WI were visually evaluated by two radiologists below, above and at the level of stenosis, where 05no signal abnormality, 15ill-defined T2 hyperintensity, and 25well-defined T2 hyperintensity. The radiologists were blinded to the diagnosis and SSEP recordings. Statistics Velocities and displacements from CSM subjects and controls with and without abnormal SSEP were compared

I.M. Vavasour et al. / The Spine Journal

-

(2014)

-

5

Fig. 3. Plots of CSF velocity over the heart cycle for (Top) control subjects and (Middle) subjects with CSM. The velocity noise level is delineated by the shaded area. (Bottom) A correlation plot of the compression ratio versus the mean CSF velocity is also shown. The regression line has an R50.7 (p!.0001). The compression ratio was measured at the level of stenosis for subjects with CSM and at C5 for control subjects. CSF, cerebrospinal fluid; CSM, cervical spondylotic myelopathy.

using a two-tailed Mann-Whitney U test. Spearman correlation coefficients for correlations between mean velocity, displacement measures, and clinical scores were calculated. Significance was set at p!.05.

Results Demographics, electrophysiological, clinical, and MRI scores for compression and signal changes in all participants

6

I.M. Vavasour et al. / The Spine Journal

Table 2 Mean velocity (cm/s) and cord displacement (cm) results for CSM subjects and controls

CSM subjects

Controls

S no.

Mean CSF

Mean cord

Total disp

Maximum disp

1 2 3 4 5 6 7 8 9 10 11 12 13 Avg 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 Avg

0.286 0.375 0.278 0.549 0.261 0.393 0.245 0.202 0.244 0.226 0.130 0.281 0.378 0.295 0.558 0.685 0.897 0.997 0.680 0.861 0.699 0.993 1.061 1.066 0.804 1.084 0.915 0.650 1.001 0.881

0.262 0.170 0.388 0.161 0.116 0.201 0.078 0.108 0.135 0.154 0.365 0.274 0.091 0.192 0.095 0.067 0.302 0.208 0.105 0.055 0.083 0.098 0.091 0.114 0.135 0.185 0.137 0.106 0.122 0.144

0.176 0.126 0.571 0.315 0.075 0.279 0.163 0.023 0.299 0.169 1.086 0.976 0.133 0.338 0.123 0.028 0.329 0.410 0.039 0.026 0.059 0.151 0.052 0.082 0.168 0.259 0.124 0.091 0.057 0.133

0.071 0.032 0.189 0.059 0.016 0.074 0.040 0.002 0.059 0.025 0.264 0.095 0.015 0.072 0.007 0.005 0.054 0.050 0.003 0.013 0.012 0.010 0.008 0.009 0.023 0.027 0.015 0.018 0.015 0.018

CSM, cervical spondylotic myelopathy; CSF, cerebrospinal fluid; disp, displacement; avg, average.

are shown in Table 1. There was no statistical difference in the mean age of the two groups. The clinical scoring revealed significant reduction of JOA scores in patients (15.162.5). Somatosensory evoked potentials scoring showed abnormal findings in 54% of CSM subjects and 47% of controls. MRI The CR was significantly decreased in CSM subjects compared with controls independent of normal or abnormal SSEP (Fig. 2). None of the controls had a T2 hyperintensity in the SC, whereas all but one of the CSM subjects had a hyperintense area in the cord (Table 1). CSF velocity Fig. 1 shows MRIs from two exemplary participants: a control with axial slices aligned at C5 (A, B, and C) and a CSM subject with images selected above and at the level of the most prominent SC compression (E, F, and G). In the CSM subject, intramedullar signal changes can be identified in sagittal and axial slices, in addition to defining the

-

(2014)

-

level of stenosis. The corresponding velocity plots in d and h highlight the large differences in CSF and cord velocity patterns in the control subject compared with the CSM patient. A movie of the velocity over a heart cycle from the control and CSM subjects in Fig. 1 is shown in Movie 1 (Supplementary material). Mean absolute CSF velocity averaged over the entire heart cycle is compared between subject groups in Fig. 2. No significant differences in CSF velocity were found between controls with and without normal SSEP or between CSM subjects with and without normal SSEP. Significant differences were found when comparing each control group to each CSM group. Cerebrospinal fluid velocity correlated with CR (Fig. 3; R50.7, p!.0001). Plots in Fig. 3 show the CSF velocity in the craniocaudal direction across the heart cycle for controls and CSM subjects (the mean CSF velocities [cm/second] are presented in Table 2 for each CSM subject and control). Controls showed the characteristic velocity pattern with a peak and a trough. Although the shape of the CSF velocity pattern for CSM subjects was similar to that of the controls, the velocity amplitude was smaller for all CSM subjects. Spinal cord motion No correlations were found between SC velocity or displacement and the CR (Fig. 4). Plots in Fig. 4 show the SC velocity across the heart cycle for controls and CSM subjects (the mean cord velocities [centimeter/second] are presented in Table 2 for each CSM subject and control). Controls showed a fairly narrow range of SC velocity centered at zero, mostly within the velocity noise level. Spinal cord velocity in some CSM subjects had higher peaks than controls and exhibited a comparable velocity pattern across the heart cycle as normal CSF velocity. No systematic increases or decreases were observed between the cord velocity at, above, or below the stenosis or in subjects with multiple or diffuse areas of stenosis. Mean absolute SC velocities averaged over the entire heart cycle are compared between subject groups in Fig. 5. A significant difference was found between the mean SC velocity in controls with normal SSEP and CSM subjects with abnormal SSEP. Cord displacement plots over the heart cycle are shown in Fig. 6 for controls and CSM subjects (the maximum and total cord displacements [centimeter] are presented in Table 2 for each CSM subject and control). The SC of CSM subjects exhibited significantly greater mean maximum and total displacement than control SC. When subdivided into those subjects with and without abnormal SSEP, maximum displacement was able to separate CSM subjects with abnormal SSEP from controls with normal and abnormal SSEP. In addition, the total cord displacement was able to distinguish between controls with normal SSEP from CSM subjects with abnormal SSEP, as well as

I.M. Vavasour et al. / The Spine Journal

-

(2014)

-

7

Fig. 4. Plots of spinal cord velocity over the heart cycle for (Top) control subjects and (Middle) subjects with CSM. The velocity noise level is delineated by the shaded area. (Bottom) A correlation plot of the compression ratio versus the mean cord velocity is also shown. No correlation was found. The compression ratio was measured at the level of stenosis for subjects with CSM and at C5 for control subjects. CSM, cervical spondylotic myelopathy.

controls with normal SSEP from controls with abnormal SSEP (Fig. 5). Correlations between parameters No significant correlations or trends were found for mean CSF velocity versus mean cord velocity for all

subjects or between mean CSF velocity and cord displacement measures. No correlation was found between the subject height and the mean velocities. Furthermore, no significant correlations were found between the mean velocities and the JOA score in CSM subjects.

8

I.M. Vavasour et al. / The Spine Journal

Fig. 5. Mean cord vel, max disp, and total disp for control subjects and subjects with CSM. The subjects are subdivided into those with normal (N) and with abnormal (A) SSEP values. Significant differences are shown (p!.05). Error bars indicate standard error. cord vel, cord velocity; max disp, maximum cord displacement; total disp, total cord displacement; CSM, cervical spondylotic myelopathy; SSEP, somatosensory evoked potential.

Discussion The CSF velocity measurements presented herein confirm previous studies that applied phase-contrast MRI to show a reduction in CSF velocity in the spinal canal in subjects suffering from CSM because of spinal canal stenosis [3,4,7]. However, this study revealed a previously unreported observation that patients with CSM show increased SC movements that were not obviously related to changes in CSF flow or cord compression. Interestingly, CSM subjects and controls with signs of altered SC conduction by means of abnormal SSEP showed increased SC movement. These findings might provide further insights into underlying pathophysiological mechanisms in CSM that are complementary to the effects of cord compression. CSF velocity The CSF velocity results are in accordance to previous studies that show a heart cycle associated cranial to caudal flow of CSF in the spinal canal [4,18,19]. Although the relationship between CR and CSF velocity was weak within each subject group, both parameters significantly distinguished control from CSM subjects. As expected, a reduction in the width of the spinal canal (quantified by the CR at the stenotic area) was related to a reduced CSF velocity that is likely because of obstruction of the CSF space [4,7]. The assessment of CSF flow (ie, deterioration in CSF flow parameters) has been considered to be of value to indirectly estimate the impact of the spinal stenosis on the cord and might be applicable to assess the efficacy of decompression surgery [20,21]. Accordingly, in one study, subjects with reduced CSF velocity compared with controls revealed a normalization of CSF flow amplitudes after surgery [3]. Although CSF flow is altered in degenerative spine disorders with and without associated SC pathology, the

-

(2014)

-

relationship between changes in CSF flow and the degree of compression are not well understood and do not show a reliable correlation with the clinical presentation and severity of SC disorder [22,23]. On average, the CR may decrease with symptom severity; however, there are subjects with mild stenosis but severe CSM symptoms and vice versa [24]. This indicates that factors such as SC ischemia and other mechanical factors, such as increased strain and shear forces, might also contribute to CSM symptoms [25–28], although there may be influences from other unknown factors. Correspondingly, spinal canal narrowing could be assumed to cause turbulent CSF flow that is not well characterized by unidirectional CSF velocity measurements and might have an unpredictable impact of inducing stress on the SC [22]. Spinal cord motion In healthy subjects, the noncompromised SC undergoes movement in all directions, both axial and perpendicular, with almost rhythmic oscillations [29,30]. Several studies focused on movements in the craniocaudal direction (here measured as cord displacement) that are closely related to the heart cycle and are considered to be induced by the CSF flow [8–11,30]. Previous studies on healthy controls have found maximum craniocaudal SC velocities ranging from 0.7 to 1.29 cm/s [9,31]. Our values were slightly larger; however, our cord displacement measurements in controls were in agreement with previous measurements [10,11]. Changes in SC movement in patients suffering from CSM have not been addressed in previous studies and the results of increased craniocaudal cord velocity in CSM compared with controls were unexpected. Previous imaging studies indicate that a narrowing within the spinal canal should prevent cord oscillation and the restriction of movement would actually improve the imaging of the cord [9,19]. However, in 6/13 of our CSM subjects, increased cord velocities (ie, more than 2 SD above the mean of controls with normal SSEP) could be observed. Furthermore, the CSM subjects showed a mean increase in cord velocity and cord displacement compared with controls. There is only one report where a similar finding of increased SC displacement has been shown intraoperatively by ultrasound measurements [8]. In patients with various pathologic SC conditions, ultrasound measures revealed increased SC oscillations that were described to be related to arterial pulsations [8]. It was even reported that in some instances, after release of the cord compression, the SC movements diminished [8]. However, there are so far no studies available in the literature that measure pathologic cord movements before and after surgical intervention. Only one study, again using ultrasound, which measured craniocaudal movement of the SC in CSM patients after surgical decompression claimed that movements ranged from just perceptible to about 2 mm, but no mention of the state of cord motion before surgery was made [32].

I.M. Vavasour et al. / The Spine Journal

-

(2014)

-

9

Fig. 6. Plots of spinal cord displacement over the heart cycle for (Top) control subjects and (Middle) subjects with CSM. (Bottom) A correlation plot of the compression ratio versus the total displacement is also shown. No correlation was found. The compression ratio was measured at the level of stenosis for subjects with CSM and at C5 for control subjects. CSM, cervical spondylotic myelopathy.

Correlation between imaging and clinical measures Although CSM subjects did have a significantly reduced CSF velocity and CR compared with controls, no correlation was found between these MRI parameters and standard clinical measures. The findings for CSF velocity are in

agreement with Watabe et al. [4] who found no correlation between CSF flow velocity at C1 and JOA scores in CSM subjects. Only Shibuya et al. [7] reported correlations between the amplitude of CSF flow at C3 and C7 and JOA scores for CSM subjects. Although CSM patients on

10

I.M. Vavasour et al. / The Spine Journal

average show the typical changes in CSF flow and cord compression, the extent of clinical symptoms varies considerably. This discrepancy holds true for our findings; patients might show rather mild stenotic changes but present with severe clinical symptoms and vice versa, whereas subjects with enormous canal encroachment are almost without symptoms [2,33,34]. Another approach to reveal changes in SC function that are independent of the clinical complaints but objectively assess cord function is the application of SSEPs in spine disorders. Somatosensory evoked potentials provide independent information complementary to clinical testing and have been proven to be sensitive to changes in spinal conductivity even during surgical intervention [35–38]. Perlik and Fisher [39] found that all their CSM subjects with rather severe CSM showed abnormal SSEP. In the present study, with less severely affected CSM subjects, not all patients revealed abnormal SSEPs. Interestingly, in the age-matched elderly controls, altered SSEP was also found, although the subjects did not complain of CSM symptoms. These findings confirm that SSEPs provide signs of cord impairment complementary to the clinical assessment. Most interesting is the finding that subjects with abnormal SSEP (both in CSM and controls) showed a trend of increased cord velocity and cord displacement irrespective of whether they complained of CSM symptoms. This finding might indicate that changes in cord movement (in terms of velocity and displacement) have some impact on cord function and may contribute to the development of CSM even before the onset of symptoms. Abnormal signal intensities on T1- and T2WIs of the SC have been described in many studies regarding CSM and can indicate a wide range of pathologies such as traumatic edema, inflammation, vascular ischemia, gliosis, or myelomalacia [40]. In patients with cervical compressive myelopathy, an increase in T2 signal intensity has been associated with more elderly subjects and those with longer disease duration [41]. Also, subjects with an increase in T2 signal intensity have a better prognosis after decompression surgery, with about 50% of SCs returning to normal T2 signal intensity [42]. Again, there is no clear relation between the extent and localization of these signal changes and the clinical presentation. In this study, only one subject with CSM showed no signal abnormalities on T2WIs, whereas none of the controls showed signal abnormalities. However, this one subject with normal appearing cord intensity still had a reduced CSF velocity and increased SC motion compared with controls. Conclusion In patients with CSM, measures of CSF velocity and cord motion provide additional insight beyond the quantification of cord compression (eg, CR). The finding of abnormal increased cord motion in CSM subjects and even in controls with subclinical cord impairment (by means of

-

(2014)

-

altered SSEP) is rather unexpected but could reveal pathophysiological mechanisms (ie, forces) that might induce strain on the SC and ultimately have an impact on the development of CSM before symptom onset. Beyond CSF and cord movements, dynamic interferences originating from the vascular system (intra- and extramedullar) need to be considered to be involved in the development of CSM, which requires further studies. Further research into the reason for the increase in cord displacement, whether turbulent flow plays a role in cord damage, and whether the increase in SC motion in some subjects with CSM could be used as a predictor of clinical course or response to treatment is warranted.

Appendix Supplementary data Supplementary data related to this article can be found at http://dx.doi.org/10.1016/j.spinee.2014.01.036 References [1] Boden SD, McCowin PR, Davis DO, et al. Abnormal magneticresonance scans of the cervical spine in asymptomatic subjects. A prospective investigation. J Bone Joint Surg Am 1990;72: 1178–84. [2] Matsumoto M, Fujimura Y, Suzuki N, et al. MRI of cervical intervertebral discs in asymptomatic subjects. J Bone Joint Surg Br 1998;80: 19–24. [3] Tominaga T, Watabe N, Takahashi T, et al. Quantitative assessment of surgical decompression of the cervical spine with cine phase contrast magnetic resonance imaging. Neurosurgery 2002;50: 791–5; discussion 796. [4] Watabe N, Tominaga T, Shimizu H, et al. Quantitative analysis of cerebrospinal fluid flow in patients with cervical spondylosis using cine phase-contrast magnetic resonance imaging. Neurosurgery 1999;44: 779–84. [5] Takahashi M, Sakamoto Y, Miyawaki M, Bussaka H. Increased MR signal intensity secondary to chronic cervical cord compression. Neuroradiology 1987;29:550–6. [6] Bhadelia RA, Bogdan AR, Kaplan RF, Wolpert SM. Cerebrospinal fluid pulsation amplitude and its quantitative relationship to cerebral blood flow pulsations: a phase-contrast MR flow imaging study. Neuroradiology 1997;39:258–64. [7] Shibuya R, Yonenobu K, Koizumi T, et al. Pulsatile cerebrospinal fluid flow measurement using phase-contrast magnetic resonance imaging in patients with cervical myelopathy. Spine 2002;27:1087–93. [8] Jokich PM, Rubin JM, Dohrmann GJ. Intraoperative ultrasonic evaluation of spinal cord motion. J Neurosurg 1984;60:707–11. [9] Levy LM, Di Chiro G, McCullough DC, et al. Fixed spinal cord: diagnosis with MR imaging. Radiology 1988;169:773–8. [10] Mikulis DJ, Wood ML, Zerdoner OA, Poncelet BP. Oscillatory motion of the normal cervical spinal cord. Radiology 1994;192: 117–21. [11] Enzmann DR, Pelc NJ. Brain motion: measurement with phasecontrast MR imaging. Radiology 1992;185:653–60. [12] Hausmann O, Min K, Boni T, et al. SSEP analysis in surgery of idiopathic scoliosis: the influence of spine deformity and surgical approach. Eur Spine J 2003;12:117–23. [13] Ellaway PH, Kuppuswamy A, Balasubramaniam AV, et al. Development of quantitative and sensitive assessments of physiological and

I.M. Vavasour et al. / The Spine Journal

[14]

[15]

[16]

[17]

[18]

[19]

[20]

[21]

[22]

[23]

[24] [25] [26] [27]

functional outcome during recovery from spinal cord injury: a clinical initiative. Brain Res Bull 2011;84:343–57. Witoszynskyj S, Rauscher A, Reichenbach JR, Barth M. Phase unwrapping of MR images using Phi UN–a fast and robust region growing algorithm. Med Image Anal 2009;13:257–68. Hirabayashi K, Miyakawa J, Satomi K, et al. Operative results and postoperative progression of ossification among patients with ossification of cervical posterior longitudinal ligament. Spine 1981;6: 354–64. Bartels RHMA, Verbeek ALM, Benzel EC, et al. Validation of a translated version of the modified Japanese orthopaedic association score to assess outcomes in cervical spondylotic myelopathy: an approach to globalize outcomes assessment tools. Neurosurgery 2010;66:1013–6. Shin JJ, Jin BH, Kim KS, et al. Intramedullary high signal intensity and neurological status as prognostic factors in cervical spondylotic myelopathy. Acta Neurochir (Wien) 2010;152:1687–94. Enzmann DR, Pelc NJ. Normal flow patterns of intracranial and spinal cerebrospinal fluid defined with phase-contrast cine MR imaging. Radiology 1991;178:467–74. Levy LM. MR imaging of cerebrospinal fluid flow and spinal cord motion in neurologic disorders of the spine. Magn Reson Imaging Clin N Am 1999;7:573–87. Guha A, Tator CH, Endrenyi L, Piper I. Decompression of the spinal cord improves recovery after acute experimental spinal cord compression injury. Paraplegia 1987;25:324–39. Furlan JC, Kalsi-Ryan S, Kailaya-Vasan A, et al. Functional and clinical outcomes following surgical treatment in patients with cervical spondylotic myelopathy: a prospective study of 81 cases. J Neurosurg Spine 2011;14:348–55. Parkkola RK, Rytokoski UM, Komu MES, Thomsen C. Cerebrospinal fluid flow in the cervical spinal canal in patients with chronic neck pain. Acta Radiol 2000;41:578–83. Golash A, Birchall D, Laitt RD, Jackson A. Significance of CSF area measurements in cervical spondylitic myelopathy. Br J Neurosurg 2001;15:17–21. Nurick S. The pathogenesis of the spinal cord disorder associated with cervical spondylosis. Brain 1972;95:87–100. Brain R. Cervical spondylosis. Ann Intern Med 1954;41:439–46. Doppman JL. The mechanism of ischemia in anteroposterior compression of the spinal cord 1975. Invest Radiol 1990;25:444–52. Shedid D, Benzel EC. Cervical spondylosis anatomy: pathophysiology and biomechanics. Neurosurgery 2007;60:S7–13.

-

(2014)

-

11

[28] Henderson FC, Geddes JF, Vaccaro AR, et al. Stretch-associated injury in cervical spondylotic myelopathy: new concept and review. Neurosurgery 2005;56:1101–13; discussion 1101–13. [29] Figley CR, Stroman PW. Investigation of human cervical and upper thoracic spinal cord motion: implications for imaging spinal cord structure and function. Magn Reson Med 2007;58:185–9. [30] Matsuzaki H, Wakabayashi K, Ishihara K, et al. The origin and significance of spinal cord pulsation. Spinal Cord 1996;34:422–6. [31] Kohgo H, Isoda H, Takeda H, et al. Visualization of spinal cord motion associated with the cardiac pulse by tagged magnetic resonance imaging with particle image velocimetry software. J Comput Assist Tomogr 2006;30:111–5. [32] Moses V, Daniel RT, Chacko AG. The value of intraoperative ultrasound in oblique corpectomy for cervical spondylotic myelopathy and ossified posterior longitudinal ligament. Br J Neurosurg 2010;24:518–25. [33] Bednarik J, Kadanka Z, Dusek L, et al. Presymptomatic spondylotic cervical cord compression. Spine 2004;29:2260–9. [34] Lee S-H, Kim K-T, Suk K-S, et al. Asymptomatic cervical cord compression in lumbar spinal stenosis patients: a whole spine magnetic resonance imaging study. Spine 2010;35:2057–63. [35] Tani T, Yamamoto H, Kimura J. Cervical spondylotic myelopathy in elderly people: a high incidence of conduction block at C3-4 or C4-5. J Neurol Neurosurg Psychiatry 1999;66:456–64. [36] Malhotra NR, Shaffrey CI. Intraoperative electrophysiological monitoring in spine surgery. Spine 2010;35:2167–79. [37] Garcia RM, Qureshi SA, Cassinelli EH, et al. Detection of postoperative neurologic deficits using somatosensory-evoked potentials alone during posterior cervical laminoplasty. Spine J 2010;10:890–5. [38] Petersen JA, Wilm BJ, von Meyenburg J, et al. Chronic cervical spinal cord injury: DTI correlates with clinical and electrophysiological measures. J Neurotrauma 2012;29:1556–66.http://dx.doi.org/10.1089/ neu.2011.2027. [39] Perlik SJ, Fisher MA. Somatosensory evoked response evaluation of cervical spondylytic myelopathy. Muscle Nerve 1987;10:481–9. [40] Kameyama T, Ando T, Yanagi T, et al. Cervical spondylotic amyotrophy: magnetic resonance imaging demonstration of intrinsic cord pathology. Spine 1998;23:448–52. [41] Yukawa Y, Kato F, Yoshihara H, et al. MR T2 image classification in cervical compression myelopathy: predictor of surgical outcomes. Spine 2007;32:1675–8; discussion 1679. [42] Alafifi T, Kern R, Fehlings M. Clinical and MRI predictors of outcome after surgical intervention for cervical spondylotic myelopathy. J Neuroimaging 2007;17:315–22.