British Journal of Radiology (2003) 76, 525-531
© 2003 British Institute of Radiology
doi: 10.1259/bjr/99259611
Evaluation of magnetic resonance myelography in the investigation of cervical spondylotic radiculopathy
D Birchall, FRCR,
D Connelly, FRCR,
L Walker, MRCP and
K Hall, FRCR
Department of Neuroradiology, Regional Neurosciences Centre, Newcastle General Hospital, Westgate Road, Newcastle upon Tyne NE4 6BE, UK
Correspondence: Dr Daniel Birchall
 |
Abstract
|
|---|
The objective of this study was to prospectively evaluate the accuracy of MR myelography for the demonstration of foraminal nerve root impingement in cervical spondylotic radiculopathy. 40 patients with cervical spondylotic radiculopathy were imaged with conventional MRI and with MR myelography. The diagnostic accuracy of these imaging strategies for the demonstration of exit foraminal stenosis was calculated relative to a gold standard of the combination of conventional MRI and MR myelography. Conventional MRI had a sensitivity of 88.9%, specificity of 99.1%, and diagnostic accuracy of 94.5% for the demonstration of exit foraminal disease (p<0.001). MR myelography alone had a sensitivity of 84.4%, a specificity of 90.1%, and diagnostic accuracy of 88% (p<0.001). However, the addition of MR myelography increased the diagnostic yield of the MR examination for the detection of foraminal stenotic disease. MR myelography is a useful adjunct to conventional MRI in the investigation of cervical spondylotic radiculopathy.
|
Introduction
|
|---|
Symptomatic cervical spondylotic radiculopathy is a prevalent condition worldwide and creates an enormous burden on medical and imaging resources. An effective imaging strategy for the demonstration of clinically significant cervical degenerative disease is a fundamental requirement for the management of this condition. MRI has replaced CT myelography in many centres for the imaging of cervical spondylotic radiculopathy because of its high soft tissue discrimination and its multiplanar capability [1–4]. MR is, in contrast to CT myelography, non-invasive and does not utilize ionizing radiation. As a result, CT myelography is generally reserved for the evaluation of patients who cannot undergo MR examination, and for the further elucidation of equivocal MR findings and the delineation of osseous foraminal stenosis [5, 6].
It is widely recognised, however, that the diagnostic accuracy of MRI in cervical spondylotic radiculopathy is limited, especially in the assessment of foraminal nerve root impingement [7, 8]. In particular, MRI has a false negative rate for the detection of foraminal nerve root compression because of suboptimal demonstration of foraminal disc and osteophyte. CT myelography, in contrast, is highly sensitive to foraminal disease because of the superior differentiation of bone and soft tissue [9], and the thin axial sections routinely acquired using spiral CT technology allow the clear demonstration of nerve root impingement and compression. Many centres consider MRI and CT myelography to be complementary techniques, using MRI as an initial screening technique. A previous study reported the superiority of CT myelography for the pre-operative evaluation of cervical spondylotic radiculopathy [9] over conventional MRI.
MR myelography is an imaging technique which utilizes heavily T2 weighted sequences to produce high signal from fluid (including cerebrospinal fluid (CSF)) [10, 11]. This results in highly contrasted images that are similar in appearance to conventional myelograms. This technique has been applied to the imaging of lumbar degenerative disease and several authors have reported it to be a valuable supplement for the demonstration of the lumbar thecal sac and dural sleeves [12–15].
We are not aware that there has been a previous prospective study of the efficacy of MR myelography in the setting of degenerative cervical radiculopathy. The objective of this study was to prospectively evaluate the diagnostic effectiveness of MR myelography for the demonstration of foraminal nerve root impingement in cervical spondylotic radiculopathy.
|
Methods
|
|---|
40 patients (22 female, 18 male, age range 28–76 years, average age 48 years) with a diagnosis of cervical spondylotic radiculopathy were included in the study. Clinical criteria for inclusion included brachalgia, upper limb motor weakness or sensory dysfunction, in the absence of signs of myelopathy and without suspicion of alternative diagnoses (such as demyelination), in association with evidence of corroborative degenerative change on plain cervical radiographs. Patients with a history of previous cervical surgery or with evidence of myelopathy were excluded from the study.
MRI was performed on a 1.5 T Intera scanner (Philips Medical Systems, Best, Netherlands) using a phased array spine coil. Each patient was prospectively imaged using the following MR protocol: sagittal turbo spin echo (TSE) T1 weighted images (repetition time (TR) 397 ms, echo time (TE) 14 ms, TSE factor 3, acquisition time 4:09); sagittal TSE T2 weighted images (TR 2498 ms, TE 110 ms, TSE factor 16, acquisition time 4:02); axial T2 weighted images (Driven equilibrium spin echo, TR 1200 ms, TE 120 ms, TSE factor 34, acquisition time 6:20, slice thickness 1.7 mm) from C3 to T1; MR myelography (single shot TSE, TR 8000 ms, TE 1000 ms, TSE factor 256, acquisition time 3:36). MR myelographic images were automatically reconstructed as maximum intensity projections and presented as 9 consecutive images at 22.5° intervals, extending from a right lateral to a left lateral projection. This process did not significantly add to the examination duration. Normal MR myelographic appearances are illustrated in Figure 1
.
View larger version (73K):
[in this window]
[in a new window]
|
Figure 1. (a) Normal anteroposterior and (b) 45° left anterior oblique MR myelogram demonstrating cervical nerve rootlets extending from the cord surface into the exit foramina.
|
|
The films were reviewed by two experienced neuroradiologists (DB and KH), without knowledge of the clinical presentation. A consensus opinion was reached regarding the presence or absence of foraminal compressive disease, which was diagnosed in the presence of nerve root compression. In each case, the left and right exit foramina at all levels from C3/4 to C7/T1 were assessed, resulting in 10 separate observations. Review of 40 consecutive examinations resulted in 400 foramina for evaluation.
On further separate occasions, the two observers were presented with the conventional MRI alone and subsequently with only the MR myelographic images. In a similar fashion, a consensus opinion was reached with respect to the presence of foraminal nerve root compression.
The combination of conventional MRI (sagittal T1 and sagittal and axial T2 images) and MR myelography was considered as a gold standard for the purposes of this study. The relative sensitivity, specificity and diagnostic accuracy of conventional MRI and of MR myelography when examined in isolation was analysed using contingency tables and the chi-squared statistic. Positive and negative predictive values were also calculated.
|
Results
|
|---|
All 40 MR examinations were considered to be technically satisfactory. This resulted in a total of 400 exit foramina available for review.
The contingency table comparing conventional MRI with the standard of the combination of conventional and myelographic MR examination is shown in Table 1.
View this table:
[in this window]
[in a new window]
|
Table 1. Contingency table comparing conventional MRI (CMR) with the standard of the combination of conventional and myelographic MR examination (CMR+MRM)
|
|
The sensitivity of conventional MRI relative to the chosen standard was calculated at 88.9%, with a specificity of 99.1%. The positive predictive value of conventional MRI was 98.8% and the negative predictive value was 91.6%, relative to the standard. The diagnostic accuracy was calculated at 94.5%. The chi-squared statisic was 159 on 1 degree of freedom and the p-value was <0.001.
The contingency table comparing MR myelography with the standard of the combination of conventional and myelographic MR examination is shown in Table 2.
View this table:
[in this window]
[in a new window]
|
Table 2. Contingency table comparing MR myelography (MRM) with the standard of the combination of conventional and myelographic MR examination (CMR+MRM)
|
|
The sensitivity of MR myelography relative to the chosen standard was calculated at 84.4%, with a specificity of 90.1%. The positive predictive value of MR myelography was 88.4% and the negative predictive value was 87.7%, relative to the standard. The diagnostic accuracy was calculated at 88%. The chi-squared statistic was 115 on 1 degree of freedom and the p-value was <0.001.
Three illustrative cases are shown in Figures 2–4.
View larger version (77K):
[in this window]
[in a new window]
|
Figure 2. (a) Left anterior oblique MR myelogram in a patient with left cervicobrachalgia shows abnormal tissue relating to the left C6/7 exit foramen, with compression of the exiting left C7 nerve root (arrow). Multilevel transverse indentations on the cerebrospinal fluid (CSF) column correspond to the presence of disc complexes indenting the ventral CSF space. (b) Axial T2 image at C6/7 confirms the presence of left-sided osteophytic foraminal narrowing and compression of the left C7 nerve root.
|
|
View larger version (87K):
[in this window]
[in a new window]
|
Figure 3. (a) 45° right anterior oblique MR myelogram in a patient with right cervicobrachalgia showing prominent compression of the exiting right C6 nerve root (large arrow). There is a smaller indentation on the exiting right C5 nerve root (small arrow). (b) Axial T2 image at C5/6 confirms the presence of right-sided osteophytic foraminal stenosis. (c) Axial T2 image at C4/5 demonstrates a smaller broad-based disc complex which is mildly impinging on the exiting C5 nerve roots. This finding was overlooked on review of conventional imaging alone and was detected because of the MR myelographic appearances.
|
|
View larger version (117K):
[in this window]
[in a new window]
|
Figure 4. (a) Axial T2 image at C5/6 in a patient with left-sided brachalgia shows bilateral exit foraminal stensosis and C6 nerve root compression. Similar disease was present at C6/7. (b) Axial T2 image at C4/5 showing possible narrowing of the left and right exit foramina. This was interpreted as bilateral foraminal stenosis on review of conventional MRI alone. (c) MR myelogram confirms compression of the exiting left and right C6 and C7 nerve roots. The exiting C5 nerve roots however are clearly seen and are not compressed (arrows).
|
|
|
Discussion
|
|---|
Cervical spondylotic radiculopathy is a common condition and creates an enormous demand on spinal and medical imaging services. In the past, CT myelography has been considered the gold standard for the demonstration of foraminal nerve root impingement, with excellent bone and soft tissue contrast and high spatial resolution. However, MRI has generally replaced CT myelography as the primary diagnostic tool in this condition, primarily because as a technique it is non-invasive and less time- and resource-intensive and because there is no exposure to ionizing radiation [1–4]. However, it is recognised that although MR provides superior diagnostic information regarding the spinal cord and spinal canal, the relatively poor contrast observed between bone, disc material and the normal contents of the cervical exit foramina results in suboptimal delineation of foraminal pathology and a significant false negative rate [5–8]. Conversely, the susceptibility artefacts at soft tissue–bone interfaces that may be observed with gradient echo imaging are associated with a tendency to overestimate foraminal stenosis in a proportion of patients [16].
Several papers have correlated MR appearances with surgical findings in patients with compressive cervical spondylotic radiculopathy, and have reported diagnostic accuracy close to 90% [1, 6]. However, other groups have reported considerably lower diagnostic sensitivity for MR, with several studies confirming that CT myelography has greater sensitivity for the demonstration of foraminal entrapment. For example, Modic et al demonstrated that MRI corresponded to surgical findings in a patient group with cervical spondylotic radiculopathy in only 74% of cases, whereas CT myelography correctly predicted the surgical findings in 85% [7]. In this report, a combination of MRI and CT myelography increased the diagnostic accuracy to over 90%. It has been suggested that use of thin section three-dimensional sequences on high performance scanners may increase the diagnostic sensitivity of MR. Bartlett et al report a diagnostic accuracy of 89% when comparing three-dimensional T2 weighted images with CT myelography for the detection of foraminal nerve root compression [17]. However, Yousem et al reported a diagnostic accuracy rate ranging from 73% to 82%, despite using three-dimensional axial T2 imaging and 1.5 mm sections [18].
As a consequence, CT myelography is used as a complimentary investigation in many centres, and is used in a significant proportion of patients in whom diagnostic uncertainty remains after MRI [5, 6]. Indeed, a recent review of trends in imaging of the spine in the USA demonstrated that there has been an increase in the use of myelography by over 50% in the years between 1993 and 1998 [19]. CT myelography is, however, invasive and is often poorly tolerated, and there is therefore a continuing effort to improve the diagnostic accuracy of MRI for this patient group in order to reduce the necessity for subsequent CT myelography. This is a particularly important factor in those countries in which managed systems of healthcare promote the use of more streamlined diagnostic pathways.
MR myelography represents a relatively recent development in MRI [10, 11] and has theoretical advantages in the visualization of the nerve roots within the entrance and proximal canal of the exit foramina. Several papers have described the use of MR myelography in the investigation of lumbar degenerative disease [12–15]. The diagnostic accuracy of MR myelography is reported to be insufficient to justify its use as an independent diagnostic technique. Furthermore, the addition of MR myelography to conventional MRI does not significantly improve the diagnostic accuracy of MR in the investigation of lumbar degenerative disease. The limited diagnostic efficacy of the technique has been attributed to inconsistent definition of nerve roots and to the observation that lumbar disc protrusions may displace only the epidural fat and not the thecal sac [12]. However, when used as an adjunct to conventional MRI, MR myelography has been shown to be useful for the further characterization of equivocal findings in a proportion of cases and to increase the diagnostic confidence in this setting [12, 13]. Otherwise, there is little evidence to support the use of MR myelography in the routine imaging of lumbar degenerative disease.
In this study, MR myelography altered the interpretation of the conventional MR images in 22 of 400 exit foramina (5.5%) when viewed in combination. The addition of MR myelography to conventional MRI of cervical spondylotic radiculopathy increased the number of compressive foraminal stenoses positively identified. This generally occurred with the presence of a soft lateral or foraminal disc protrusion, the suboptimal contrast with the normal foraminal contents being insufficient to allow its ready detection, as demonstrated in Figure 3
. Although the most radiologically obvious foraminal stenoses were identifiable on conventional MRI in all cases, the addition of MR myelography improved the detection of concomitant foraminal stenoses in multilevel disease.
However, when the MR images were reviewed with the relevant clinical details, conventional MRI detected the most radiologically significant abnormalities in all 40 patients.
MR myelography helped in the detection of foraminal disease in two ways. First, the heavily T2 weighted nature of the technique resulted in high contrast between the CSF and all other soft tissue structures, including disc material. As a consequence, this technique allowed the demonstration of foraminal disc protrusions that would otherwise have been less easily detected on conventional axial T2 weighted images. Second, any possible abnormality seen on MR myelography prompted a re-evaluation of the relevant axial and sagittal sequences and produced an increased rate of detection of foraminal compression as a result of this review process.
MR myelography when viewed in isolation had an insufficient diagnostic accuracy to justify its use as an independent imaging technique for the evaluation of cervical foraminal disease. MR myelography had a greater tendency to underestimate, rather than overestimate, the severity of foraminal disease. This is attributed to a lack of sensitivity to foraminal compression in the more lateral aspects of the exit foramen, where the dural sleeve narrows and where the amount of epidural fat increases.
It should be noted that the diagnostic accuracies quoted in this paper are relative to a gold standard of the combination of conventional imaging and MR myelography. It is well established that MRI has a suboptimal diagnostic accuracy when compared with surgical findings or with CT myelography, and no such standard has been applied in this study. CT myelography is not routinely performed at our institution and its use in this study was not justified ethically or clinically. As a result, it is not possible to comment on the true diagnostic accuracy of the various MR techniques. This does not, however, detract from the validity of observing the effect of including MR myelographic sequences in the interpretation of the MR examination. Similarly, the observations were made by two experienced observers in conjunction and as such no evaluation of observer variability was made. It is likely that the subjective nature of the observations will introduce a degree of observer variability and this is currently under evaluation at our institution.
In conclusion, this study suggests that MR myelography may be a useful adjunct to conventional axial and sagittal imaging in the investigation of cervical spondylotic radiculopathy. The use of MR myelography increased the number of compressive foraminal stenoses positively identified and it has the potential to reduce the need for subsequent CT myelographic examination in a proportion of this patient group. Further study is required to define the diagnostic accuracy of this technique by comparison with conventional CT myelography. However, MR myelography is a robust sequence and adds only a short time to the overall examination, and it is arguable that routine use should be made of MR myelography in the imaging of patients with cervical spondylotic disease.
Received for publication October 12, 2002.
Revision received February 13, 2003.
Accepted for publication April 29, 2003.
|
References
|
|---|
- Brown BM, Schwartz RH, Frank E, et al. Preoperative evaluation of cervical radiculopathy and myelopathy by surface coil MRI. AJR Am J Roentgenol 1988;151:1205–12.[Abstract/Free Full Text]
- Shafaie FF, Wippold J, Gado M, Pilgram T, Riew KD. Comparison of computed tomography myelography and magnetic resonance imaging in the evaluation of cervical spondylotic myelopathy and radiculopathy. Spine 1999;24:1781–5.[CrossRef][Medline]
- Boutin RD, Steinbach LS, Finnesey K. MRI of degenerative diseases of the cervical spine. Magn Reson Imaging Clin N Am 2000;8:471–90.[Medline]
- Tsuruda JS, Norman D, Dillon W, Newton TH, Mills DG. Three-dimensional gradient-recalled MRI as a screening tool for the diagnosis of cervical radiculopathy. AJNR Am J Neuroradiol 1989;10:1263–71.[Abstract]
- Kaiser JA, Holland BA. Imaging of the cervical spine. Spine 1998;23:2701–12.[CrossRef][Medline]
- Van de Kelft E, van Vyve M. Diagnostic imaging algorithm for cervical soft disc herniation. Acta Chirurgica Belgica 1995;95:152–6.[Medline]
- Modic MT, Masaryk TJ, Mulopulos GP, Bundschuh C, Han JS, Bohlman H. Cervical radiculopathy: prospective evaluation with surface coil MRI, CT with metrizamide, and metrizamide myelography. Radiology 1986;161:753–9.[Abstract/Free Full Text]
- Bartlett RJ, Rowland Hill CA, Devlin R, Gardiner ED. Two-dimensional MRI at 1.5 and 0.5 T versus CT myelography in the diagnosis of cervical radiculopathy. Neuroradiology 1996;38:142–7.[Medline]
- Karnaze MG, Gado MH, Sartor KJ, et al. Comparison of MR and CT myelography in imaging the cervical and thoracic spine. AJNR Am J Neuroradiol 1987;8:983–9.
- Krudy AG. MR myelography using heavily T2-weighted fast spin-echo pulse sequences with fat presaturation. AJR Am J Roentgenol 1992:1315–20.
- el-Gammal T, Brooks BS, Freedy RM, Crews CE. MR myelography: imaging findings. AJR Am J Roentgenol 1995;164:173–7.[Abstract/Free Full Text]
- Thornton MJ, Lee MJ, Pender S, McGrath FP, Brennan RP, Varghese JC. Evaluation of the role of magnetic resonance myelography in lumbar spine imaging. Eur Radiol 1999;9:924–9.[CrossRef][Medline]
- Pui MH, Husen YA. Value of magnetic resonance myelography in the diagnosis of disc herniation and spinal stenosis. Austral Radiol 2000;44:281–4.[CrossRef][Medline]
- Kuroki H, Tajima N, Hirakawa S, Kubo S, Tabe R, Kakitsubata Y. Comparative study of MR myelography and conventional myelography in the diagnosis of lumbar spine diseases. J Spinal Disord 1998;11:487–92.[Medline]
- Hergan K, Amann T, Vonbank H, Hefel C. MR myelography: a comparison with conventional myelography. Eur J Radiol 1996;21:196–200.[CrossRef][Medline]
- Tsurada JS, Remley K. Effects of magnetic susceptibility artefacts and motion in evaluating the cervical neural foramina on 3DFT gradient-echo MRI. AJNR Am J Neuroradiol 1991;12:237–41.[Abstract]
- Bartlett RJ, Rowland Hill C, Gardiner E. A comparison of T2 and gadolinium enhanced MRI with CT myelography in cervical radiculopathy. Br J Radiol 1998;71:11–9.[Abstract]
- Yousem DM, Atlas SW, Goldberg HI, Grossman RI. Degenerative narrowing of the cervical spine neural foramina: evaluation with high-resolution 3DFT gradient-echo MRI. AJNR Am J Neuroradiol 1991;12:229–36.[Abstract]
- Rao VM, Parker L, Levin DC, Sunshine J, Bushee G. Use trends and geographic variation in neuroimaging: nationwide medicare data for 1993 and 1998. AJNR Am J Neuroradiol 2001;22:1643–9[Abstract/Free Full Text]
This article has been cited by other articles:
Top
Abstract
Introduction
