|
|
|
|
|
INTRODUCTION
|
In patients with bulging cervical discs, MRI demonstrates the presence of spinal cord hyperintense signal opposite one or more of the bulging discs. The clinical significance and pathology of this spinal cord hyperintensity are not fully understood. Several studies have failed to show a significant relationship between intramedullary hyperintense signal and clinical symptoms.(11,15,17) From pathological point of view, the spinal cord hyperintensity has been related to degenerative changes in the white matter, cystic changes in grey matter, and spinal cord oedema or congestion.(10,11,18,20 )
The aim of our study was to investigate the clinical significance of this spinal cord hyperintensity from several aspects: Firstly, whether its presence would represent a permanent change in the spinal cord; Secondly, the usefulness of surgical decompression in patients with this MRI finding; Thirdly, the best approach to be used for surgical decompression of such patients: anterior, posterior or combined.
|
MATERIALS
AND METHODS
|
In addition, the neurological status was evaluated according to the Neurosurgical Cervical Spine Scale (NCSS) (Table 1). NCSS is a method of scoring motor function of the upper and lower extremities and sensory deficits.(7) The maximum score in this scale (patient’s free of neurological deficits) is 14 and the minimum (patient’s severely disabled with motor and sensory deficits) is 3.
All patients had preoperative MRI. MRI studies included T1-weighted images (spin-echo; TR 500-600 msec, TE 15-20 msec) and T2-weighted images (gradient-echo; TR 200-300 msec, TE 15-20 msec; flip angle 10-12 degrees) of the cervical spine obtained in the sagittal and axial planes. Analysis of the MRI especially focused on:
| 1. | The deformity of the spinal cord observed at the maximum compression level expressed by the ratio of the sagittal diameter divided by the transverse diameter of the spinal cord observed on T1-weighted images (cord deformity ratio = sagittal diameter/transverse diameter) (Fig. 1).(17,22) | ||
| 2. | Intramedullary signal changes on T2-weighted images. | ||
| 3. | Associated disc protrusion diagnosed as a soft-tissue mass protruding posteriorly from the intervertebral disc space beyond the posterior aspects of adjacent vertebral bodies in the T1- and T2-weighted images |
Patients were separated into two groups according to the intramedullary
signal changes observed in T2-weighted MR images:
 |
Group A (n=27): included patients having bulging cervical discs associated with cervical cord hyperintense signals in the T2-weighted MR images. | ||
|
Group B (n=25): included patients having bulging cervical discs but with normal spinal cord signals in the T2-weighted MR images. |
All patients underwent surgical decompression of the spinal cord.
Anterior cervical discectomy was performed in 24 patients (46%). In
this group, anterior cervical discectomy was performed at one level
in 13 patients, at two levels for 8 patients and at three levels in
3 patients. Interbody fusion using autogenous iliac bone graft was performed
when discectomy involved 2 or 3 cervical levels. On the other hand,
posterior decompression by means of wide laminectomy was performed in
25 patients (48%) where three-level laminectomy was done in 12 patients,
four-level laminectomy in 5 patients and more than four-level laminectomy
was performed in 8 patients. The remaining 3 patients (6%) had cervical
canal stenosis and were managed initially by anterior cervical discectomy
and interbody fusion and 6 months later by laminectomy.
Postoperative neurological status was determined according to NCSS for
all patients every three months. Follow-up periods ranged from 3 to
27 months (mean 16 months). The degree of neurological recovery, expressed
as the ratio between the postoperative gain in NCSS and the pre-operative
deficit in NCSS, was calculated for each patient using the equation:
NCSS
at last follow-up – Initial NCSS score
Degree of neurological recovery = -------------------------------------------------------
Maximum
NCSS score (ie. 14) – Initial NCSS score
In addition, MRI studies were repeated in 17 patients in Group A after
3 months or more from surgery to study the changes in the spinal cord
hyperintense signals after cord decompression.
The mean values of the spinal cord deformity ratio, NCSS scores and
degree of neurological recovery were expressed as means +/- standard
deviation. For comparison of mean values, the standard T-test was used.
P values less than 0.05 were considered significant.
Table
1 - Neurosurgical Cervical
Spine Scale*
|
 |
|
Figure
1 — Scatterplot graph (left) depicting the relationship
between pre-operative neurological deficits as scored by the NCSS
and spinal cord deformity shown in T1-weighted axial images at
the maximum compression level (cord deformity = sagittal diameter
(a) transverse diameter (b) of the spinal cord (right)).
|
|
RESULTS
|
Table 2 - Features of Clinical Presentation
|
Clinical features
|
Group A
|
Group B
|
| Age: - Range |
34-83 yrs
|
32-63 yrs
|
| - Mean |
57.5 +/- 9.5 yrs
|
51.8 +/- 8.8 yrs
|
| Sex: - Males |
16
|
15
|
| - Females |
11
|
10
|
| Duration of symptoms |
15.6 +/- 2.7 months
|
14.4 +/- 2.2 months
|
| Presentation (%): |
| - Radiculopathy only |
0
|
28%
|
| - Myelopathy only |
70%
|
56%
|
| - Radiculo-myelopathy |
30%
|
16%
|
| Pre-operative NCSS score |
8.4 +/- 2.7
|
10.3 +/- 2.4
|
| Surgical Approach |
| - Anterior |
12
|
12
|
| - Posterior |
12
|
13
|
| - Both anterior and posterior |
3
|
|
| Follow-up period: |
20.3 +/- 2.5
|
21.7 +/- 2.7
|
| NCSS score at follow-up |
10.2 +/- 2.4
|
12.7 +/- 2.1
|
| Average degree of neurological recovery |
0.38 +/- 0.12
|
0.65 +/- 0.13*
|
Table 3 - Relationship between duration of symptoms and the degree of neurological recovery.
| Duration of symptoms |
Degree of neurological recovery (%)
|
Total
|
|
>0.25-0.5
|
>0.5
|
|
||
|
1
|
9
|
12
|
22
|
|
|
>6 – 24 months
|
6
|
9
|
3
|
18
|
|
> 24 months
|
3
|
5
|
4
|
12
|
|
Total
|
10
|
23
|
19
|
52
|
The average number of bulging cervical discs in Group A (2.6 +/- 0.4) was not significantly different from that in Group B (2.3 +/- 0.3). The correlation between the number of bulging discs and each of the pre-operative NCSS and the post-operative neurological recovery were not statistically significant. On the other hand, the mean pre-operative spinal cord deformity ratio in Group A (0.23 +/- 0.07) was significantly lower than in Group B (0.33 +/- 0.08, p=0.02). The correlation between the degree of spinal cord deformity and each of the pre-operative NCSS and the post-operative neurological recovery were not statistically significant.
In patients with spinal cord hyperintense signal (Group A), spinal cord hyperintensity signal in MRI involved one segment of the cervical cord in 18 out of 27 patients (66.7%), two segments in 5 patients (18.5%) and 3 or more segments in 4 patients (14.8%). The correlation between the length of spinal cord hyperintensity and the degree of neurological recovery was not statistically significant.
The late post-operative MRI studies that were performed in 17 patients in Group A demonstrated that the spinal cord deformity significantly improved after decompressive surgery (mean pre-operative cord deformity ratio 0.24 +/- 0.09, mean post-operative cord deformity ratio 0.43 +/- 0.07; p=.003). The intramedullary hyperintense signal on pre-operative T2-weighted MRI was still observed on post-operative MRI in all 17 patients, although all those patients showed neurological improvement after surgery (Fig. 2 and 3). However, in 2 patients with pre-operative spinal cord hyperintense signal involving 3 segments, the intramedullary hyperintense signals greatly decreased to involve one cervical cord segment only in the post-operative MRI (Fig. 4).
Table 4 - Relationship between initial NCSS score and the degree of neurological recovery.
|
Initial NCSS score
|
Degree of neurological recovery (%)
|
Total
|
|
>0.25-0.5
|
>0.5
|
|
||
|
3-5
|
2
|
5
|
0
|
7
|
|
6-8
|
6
|
8
|
4
|
18
|
|
9-11
|
2
|
8
|
8
|
18
|
|
12-14
|
0
|
2
|
7
|
9
|
|
Total
|
10
|
23
|
19
|
52
|
|
|
|||||||
| Figure 3 — Preoperative (a and b) and late postoperative (c and d) MRI demonstrates the persistense of spinal cord hyperintensity regardless of adequate anterior decompression. |
|
|
DISCUSSION
|
The importance of preoperative clinical evaluation in planning and predicting the outcome of surgical management in cervical spondylosis is in no way diminished. Preoperative clinical assessment is of particular significance for patients with multilevel cervical spondylosis to determine the levels responsible for their symptoms.(21) In addition, Lee, et al., (1993) found that age greater than 60 years at the time of presentation, duration of symptoms more than 18 months prior to surgery, pre-operative bowel or bladder dysfunction, and lower-extremity dysfunction were associated with poorer surgical outcome, although even when these conditions were present, gait improvement was noted in at least 70% of patients in their study.(12) The clinical results of our study agreed with many of these observations and showed that the degree of postoperative neurological recovery was inversely correlated with both the duration of symptoms and the severity of the preoperative neurological deficit. However, the degree of postoperative clinical improvement in our study was not significantly correlated with age.
In the evaluation of cervical spondylosis, MRI is the preferred initial diagnostic study because of its superior depiction of soft tissue anatomy, including intervertebral discs and spinal cord changes.(8,14) This has been related for several reasons. Firstly, the intervertebral discs can be easily evaluated for internal degenerative changes non-invasively with T2-weighted images. Secondly, disc contour changes can be assessed easily in various planes with no loss of resolution. Thirdly, neural structures and soft tissue structures are well delineated in detail that is superior to other imaging modalities without the need for contrast material. Frequently, the T2-weighted MRI shows the presence of cervical cord hyperintense signal in association with bulging cervical disc. The clinical significance and origin of this spinal cord hyperintense signal is not clear.
Several authors have tried to analyse the clinical significance of intramedullary hyperintensity on images with cervical spondylosis or ossified posterior longitudinal ligament.(11,15,17,22) Unfortunately, these studies have failed to show a significant relationship between intramedullary hyperintensity and clinical symptoms. In the present study, the mean NCSS score of neurological deficits in patients exhibiting intramedullary hyperintensity on MRI was found to be significantly lower than that in patients who did not exhibit intramedullary abnormality, although this finding did not indicate poor functional prognosis. Also in our study, 7 of our patients with intramedullary hyperintensity presented with “fried egg-like” hyperintensity in the spinal cord on T2-weighted images. This characteristic pattern of intramedullary abnormality is known to represent cystic changes in the grey matter.(1,6,11,18 )
From our results and those of others, it is likely that this intramedullary hyperintensity represents histopathological changes in the spinal cord; however some of these changes may not result in clinical symptoms.(11,18,20) Kameyama, et al., found that in cases with ossified posterior longitudinal ligament, if the spinal cord is mildly affected, the spinal cord showed cystic changes in grey matter while the surrounding white matter is preserved.(10) Such a patient will show only mild neurological deficits with intramedullary hyperintense signal appearing in T2-weighted MRI. On the other hand, when the spinal cord is severely affected by cervical spondylosis or ossified posterior longitudinal ligament, it will show degenerative changes in the white matter tracts, including Wallerian degeneration, with ascending and descend-ing degeneration of the white matter tracts.(10,11,18) Patients with degenerative changes in the white matter will present with various degrees of neuro-logical deficits.
Al-Mefty and his colleagues (1993) designed a canine model simulating both cervical spondylosis and its results in delayed progressive myelopathy.(2) Subclinical cervical cord compression was achieved in dogs by placing a Teflon washer posteriorly and a Teflon screw anteriorly, producing an average of 29% stenosis of the spinal canal. Twelve of 14 animals developed delayed and progressive clinical signs of myelopathy, with a mean latent period to onset of myelopathy of 7 months. MRI revealed intramedullary changes. Histopathological studies showed abnormalities overwhelmingly within the grey matter, including changes in vascular morphology, loss of large motor neurons, necrosis, and cavitation. Axonal degeneration and obvious demyelination were rarely seen. The most profound morphological changes occurred at the site of compression. They proposed that a momentary arrest of microcirculation would occur during extension of the neck because of loss of the reserve space in the compromised spinal canal. This microcirculatory disturbance was predominant in the watershed area of the cord and mainly affected the highly vulnerable anterior horn cells, leading to neuronal death, necrosis and eventual cavitation at the junction of the dorsal and anterior horns.
Kameyama, et al. (1998), reported similar clinical observations in 3 cases with cervical spondylotic amyotrophy syndrome characterised by severe muscular atrophy in the upper extremities, with an absent or insignificant sensory deficit.(9) Sagittal T2-weighted MR images in those 3 patients showed multi-segmental linear high-signal intensity within the compressed spinal cord. These high-signal intensity lesions appeared to be located at the anterior horns on axial images. The spinal cord compression was less severe in the neck-neutral position, but spinal canal stenosis increased when the neck was extended. They suggested multi-segmental damage to the anterior horns caused by dynamic cord compression, possibly through circulatory insufficiency, as a pathophysiologic mechanism of this syndrome.(9)
Tanaka, et al., (1997) studied the cranio-caudal motion velocity in the cervical spinal cord in degenerative diseases using MRI.(19) In this study, the spinal cord showed a higher motion velocity at the compression level than at non-compression levels in patients with impaired lower extremity motor function but not in patients with normal lower extremity motor function.
The other possible origin of the intramedullary hyperintense signal will be congestive or oedematous changes in the spinal cord caused by disturbed venous drainage due to spinal cord compression. However, this possibility is unlikely because the pre- and postoperative MRI studies in our series and in that of others, revealed that intramedullary hyperintensity was still observed after satisfactory decompressive surgery in all patients.(11) However, congestive or oedematous changes might play a contributing role in the appearance of intramedullary hyperintensity in a small group of our patients in whom the extent of the spinal cord hyperintense signals were found to be reduced from three segments in the preoperative MRI to only one segment in the postoperative MRI.
In our study, we used a simple method (cord deformity ratio = sagittal diameter/transverse diameter) to evaluate deformity of the spinal cord caused by anteroposterior compression of the spinal canal.(11) Similar methods have been used by others.(17,22) Yone, et al., (1992) have reported that the minimum anteroposterior diameter of the cervical cord measured on T1-weighted sagittal images tended to decrease with the increasing severity of myelopathy in cases with ossified posterior longitudinal ligament, whereas there was no such relationship in cases with cervical spondylosis.(22) Okada, et al., (1994) reported that the ratio of sagittal diameter/transverse diameter of the spinal cord was not correlated with the severity of myelopathy in cases with either cervical spondylosis or ossified posterior longitudinal ligament.(17) In our series, the group of patients with intramedullary hyperintensity (Group A) was found to have a significantly lower mean cord deformity ratio than those without intramedullary hyperintensity (Group B). However, there was no direct correlation between the cord deformity ratio and the preoperative neurological deficits or the degree of postoperative neurological recovery.
The results of our study and those of others demonstrated that MRI played an important role in the planning of surgical treatment for cervical spondylosis.(5,14) MRI nicely demonstrated the number of bulging discs and the level of maximum cord compression, in addition to the location and extent of intramedullary hyperintensity. These data are vital for making decisions of the approach (anterior or posterior) that should be used and the extent of surgical decompression.
Many surgical procedures, including several aggressive anterior and posterior approaches, have been introduced for surgical decompression of the spinal cord as a result of cervical spondylosis. Ou, et al., (1994) introduced the extensive anterior decompression technique that included the resection of intervertebral joints, neural and transverse foraminotomy, subtotal corpectomy, and fusion with strut graft.(16) Others suggested aggressive posterior decompressive techniques for cervical spondylosis as the transdural approach to the anterior spinal canal and the modified open-door cervical expansive laminoplasty.(4,12) Nevertheless, the functional outcome was not related to the approach used and declined with long-term follow-up.(14) Prospective randomised studies are certainly needed for better evaluation of these techniques.
The personal experience of Prof. Gamal Azab and the results of our study have demonstrated that, in cases with intramedullary hyperintensity, the degree of post-operative neurological recovery was not different whether surgical decompression was limited to the level(s) of intramedullary hyperintensity or it was extended to involve all levels of bulging discs regardless of the extent of intramedullary hyperintensity.(3) These observations infer that surgical decompression should be limited to the area of intramedullary hyperintensity. However, double-blinded controlled studies of longer follow-up should be performed before we reach a final conclusion.
|
CONCLUSION
|
|
REFERENCES
|
| 1. | Al-Mefty O, Harkey LH, Meddleton TH: Myelopathic cervical spondylotic lesions demonstrated by magnetic resonance imaging. J Neurosurg 1988, 68: 217-222 |
| 2. | Al-Mefty O, Harkey LH, Marawi I, Haines DE, Peeler DF, Wilner HI, Smith RR, Holaday HR, Haining JL, Russell WF: Experimental chronic compressive cervical myelopathy. J Neurosurg 1993, 79: 550-561 |
| 3. | Azab GM: Personal communication |
| 4. | Fox MW, Onofrio BM: Transdural approach to the anterior spinal canal in patients with cervical spondylotic myelopathy and superimposed central soft disc herniation. Neurosurg 1994, 34: 634-641 |
| 5. | Houser OW, Onofrio BM, Miller GM, Fogler WN, Smith PL: Cervical stenosis and myelopathy: evaluation with computed topographic myelography. Mayo Clin Proc 1994, 69: 557-563 |
| 6. | Iwasaki Y, Abe H, Isu T: CT myelography with intramedullary enhancement in cervical spondylosis. J Neurosurg 1985, 63: 363-366 |
| 7. | Kadoya S: Grading and scoring system for neurological function in degenerative cervical spine disease – Neurosurgical Cervical Spine Scale. Neurol Med Chir 1992, 32: 40-41 |
| 8. | Kaiser JA, Holland BA: Imaging of the cervical spine. Spine 1998, 23: 2701-2712 |
| 9. | Kameyama T, Ando T, Yanagi T, Yasui K, Sobu G: Cervical spondylotic amyotrophy. Magnetic resonance imaging demonstration of intrinsic cord pathology. Spine 1998, 23: 448-452 |
| 10. | Kameyama T, Hashizume Y, Ando T: Spinal cord morphology and pathology in ossification of the posterior longitudinal ligament. Brain 1995, 118: 263-278 |
| 11. | Koyanagi I, Iwasaki Y, Hida K, Imamura H, Abe H: Magnetic resonance imaging findings in ossification of the posterior longitudinal ligament of the cervical spine. J Neurosurg 1998, 88: 247-254 |
| 12. | Lee TT, Manzano GR, Green BA: Modified open-door cervical expansive laminoplasty for spondylotic myelopathy: operative technique, outcome and predictors for gait improvement. J Neurosurg 1997, 86: 64-68 |
| 13. | Long DM: Lumbar and cervical spondylosis and spondylotic myelopathy. Curr Open Neurol Neurosurg 1993, 6: 576-580 |
| 14. | McCormack BM, Weinstein PR: Cervical spondylosis. An update. West J Med 1996, 165: 43-51 |
| 15. | Morio Y, Yamamoto K, Kuranobu K: Does increased signal intensity of the spinal cord on MR images due to cervical spondylosis predict prognosis? Arch Orthop Trauma Surg 1994, 113: 254-259 |
| 16. | Ou Y, Lu J, Mi J, Cheng Li, Zhang J, Li Y, Sheng N: Extensive anterior decompression for mixed cervical spondylosis. Resection of uncovertebral joints, neural and transverse foraminotomy, subtotal corpectomy, and fusion with strut graft. Spine 1994, 19: 2651-2656 |
| 17. | Okada Y, Ikata T, Yamada H: Magnetic resonance imaging study on the results of surgery for cervical compression myelopathy. Spine 1994, 18: 2024-2029 |
| 18. | Ono K, Ota H, Tada K: Cervical myelopathy secondary to multiple spondylotic protrusions. A clinicopathologic study. Spine 1977, 2: 109-125 |
| 19. | Tanaka H, Sakurai K, Iwasaki M, Harada K, Inaba F, Hirabuki N, Nakamra H: Craniocaudal motion velocity in the cervical spinal cord in degenerative disease as shown by MR imaging. Acta Radiol 1997, 38: 803-809 |
| 20. | Terae S, Taneichi H, Abumi K: MRI of Wallerian degeneration of the injured spinal cord. J Comp Assist Tomogr 1993, 17: 700-703 |
| 21. | Watson JC, Broaddus WC, Smith MM, Kubal WS: Hyperactive pectoralis reflex as an indicator of the upper cervical spinal cord compression. Report of 15 cases. J Neurosurg 1997, 86: 159-161 |
| 22. | Yone K, Sakou T, Yanase M: Preoperative and post operative magnetic resonance image evaluations of the spinal cord in cervical myelopathy. Spine 1992, 17(Suppl 10): S388-S392 |
 |
