Acta Radiologica 30 (1989) Fasc. 3

FROM THE DEPARTMENTS OF DIAGNOSTIC RADIOLOGY, SECTION OF NEURORADIOLOGY, AND THE DEPARTMENT OF ORTHOPAEDICS, UNIVERSITY HOSPITAL, S-90185 UMEA, SWEDEN.

COMPARISON OF MYELOGRAPHY AND COMPUTED TOMOGRAPHY IN ESTABLISHING LUMBAR DISC HERNIATION

M. K. J. FAGERLUND and U. E. THELANDER

Abstract The results of lumbar myelography and computed tomography (CT) were compared in 51 consecutive patients with clinically suggested lumbar disc herniation. A total of 100 intervertebral levels were examined. At 62 levels, either L4/L5 or L5/S I, myelography was normal. CT showed no pathologic changes at 55 levels. The results concurred between myelography and CT in 89 per cent of the patients with normal findings. Four cases of disc herniation and one bulging disc, which had been missed at myelography because of a large epidural space at L5/S I, were picked up by CT. Two of these were verified by surgery and two were treated conservatively. There was one possible false negative case with CT as well. Abnormalities were shown at 38 intervertebrallevels, 22 in the bulging and 16 in the herniated disc group. The pathologic changes concurred in 84 per cent between the two investigations. For intervertebral disc herniation the true negative rate was, for myelography 88 per cent and for CT 97 per cent. The treatment strategy could have been based on CT alone at 37 out of 38 levels (97%), and on myelography alone at 34 out of 38 levels (89%). Furthermore, at CT the imaging of soft tissues and intervertebral joints was superior to that at myelography. It is concluded that CT should be the primary examination method of imaging for lumbar disc herniation. Myelography is, however, to be preferred where the level of the lesion is clinically unclear or when the entire lumbar region and thoraco-lumbar junction are to be examined. Key words: Spine, intervertebral discs; disc herniation; computed tomography; myelography.

The imaging of the lumbar spine is a field undergoing rapid change. Magnetic resonance imaging is predicted to be the most suitable method of examination for establishing a herniated lumbar disc (1, 15), but as long as only a few institutions can afford the costly equipment, computed tomography (CT) and myelography will remain the standard examination methods. Despite the recent development of high resolution CT 16-898083

scanners, there still exists some controversy as to the best standard method of examination for clinically suggested herniation of a lumbar intervertebral disc (13). CT, being non-invasive, is preferred in many institutions as the first examination. In several previous reports, with relatively large numbers of patients, CT was found to be as good as (10,20, 23, 26) or better (9, 25, 30) than myelography. In contrast BELL et colI. (2), however, preferred myelography to CT. The study design presents a problem because the patient material has to be selected, all having suggested lumbar disc herniation. The only objective criteria for both CT and myelography would be verification by surgery. Unfortunately, the treatment strategies vary from surgeon to surgeon, and furthermore, even surgery may have a certain incidence of false positive or negative results (12). The information derived from these methods is basically different. The intervertebral discs are visualized indirectly in myelography, while CT gives a direct axial imaging of the disc. With the development of a new method of examination there is usually a tendency to overestimate the role of the new modality. A blind randomized study is difficult if not impossible as the examinations are carried out on the clinically suggested disc pathology, and also because the examiner may well have knowledge of previous tests. Finally, patients with normal myelography are seldom operated upon. In this study the 2 first difficulties were eliminated, as far as possible, by analysing the myelography and CT studies 2 to 3 years after the examina-

Accepted for publication 26 November 1988.

241

242

M. K. 1. FAGERLUND AND U. E. THELANDER

tions were performed. In that way the error from inexperience in the use of the new technique was also eliminated. Since its introduction contrast myelography has enjoyed the position as the standard pre-operative radiologic method of examination. If this standard is to be replaced by a new examination modality, this must be able to show both normal and pathologic conditions with equal or better accuracy than the old method. Therefore, the aim of this study was to analyse the findings at CT when myelography was normal, as well as when it showed disc bulging or herniation. The results were then compared, when available, with the surgical findings.

Table 1 Level by level comparison of myelography with computed tomography Myelography

Normal disc (n=62) Bulging disc (n=22) Herniated disc (n= 16)

Computed tomography Normal Slight bulgdisc ing

Broad bulging

32 I

10

23 6

3

Herniated disc 4 5 16

Material and Metholds

Table 2

Fifty-one consecutive patients with clinically suggested intervertebral disc herniation of the 2 lowest lumbar levels were examined by CT and myelography, resulting in an analysis of 100 intervertebral levels, two patients having only one level examined. The interval between the two examinations was no more than 2 hours. There were 33 males (mean 40.1 years) and 18 females (mean 37.8 years). The CT (Siemens Somatom 2) was performed as the first examination using 125 kV, 260 mAs and with a slice thickness of 4 mm. The increments between the slices were 3 mm in order to facilitate sagittal and coronal image reformatting. The gantry angulation was adjusted so as to lie parallel to the plane of the intervertebral discs. This was, however, not always possible at the lumbosacral interspace since the maximum obtainable gantry tilt was 20 degrees. The patients were lying supine on a hard cushion with knees and hips flexed. The intervertebral space was examined from the intervertebral foramen above, to the foramen below the disc being examined. The data analysis was based on images reproduced on films with a standard window of 256 Hounsfield units (HU) and a level of 30 HU. Lumbar myelography (15-18 ml metrizarnide, 170 mg l/rnl, Nycomed, Norway) included a lateral, a posteroanterior and right and left oblique views in a prone position. Finally, lateral flexion and extension views in the sitting position were obtained. The myelograms were divided into 3 groups, based on the grade of soft tissue indentation in the lateral flexion/ extension views and on the degree of deformity of the dural sac or nerve root sheaths: 1) No indentation or slight indentation, disappearing in the forward flexion view. This group was called the normal group. 2) Indentation not reduced and/or a slightly rounded deformity of one or several nerve root sheaths. This was the bulging disc group. 3) Widening and shortening or amputation of a root sheath and/or an angular deformity of the dural sac with ventral indentation. This corresponded to the disc herniation group.

Treatment of patients showing 62 normal levels, 22 levels with a bulging disc and 16 levels with disc herniation at myelography Myelography

Treatment

Normal group

Conservative, n=58 Hernia operations, n=4 Conservative, n= 15 Laminectomies, narrow canal, n=4 Hernia operations, n=2 Stabilizing operation, n= I Hernia operations, n= II Laminectomy, spinal stenosis, n= 1 Conservative treatment, n=2 Refused operation, n= I File lost, n= 1

Bulging group

Herniations

Later on, without any prior knowledge of the clinical findings, the CT examinations were classified into 4 groups: 1) No bulging of the intervertebral disc. 2) Slight bulging of the disc without displacement of the epidural fat. 3) Broad disc bulging with fat displacement or dural sac indentation. 4) Focal disc protrusion or a broad disc extension visible at least on 2 consecutive slices. This group was called the herniation group. Futhermore, the CT evaluation also included displacement of the epidural fat, obliteration of the lateral recesses, degenerative facet joint arthropathy and disc degeneration. A narrow spinal canal was defined as a diminution of the overall size of the canal. The criteria included bony overgrowth of the facet joints, hypertrophied flaval ligaments and obliteration of the epidural fat. Results

For a comprehensive analysis the myelography was compared with CT, level by level. There were 62 normal intervertebral discs at myelography (Table 1). Pathologic changes were shown at 38 intervertebral levels, 22 in the

LUMBAR DISC HERNIATION

a

b

Fig. I. a) Myelography with a short 51 root (_) but without signs of disc herniation. b) Medial focal disc extension L5/51 consistent with a disc herniation (_).

J a

b

Fig. 2. a) At myelography. compression of the left L5 root sheath (_). Other projections revealed a narrow canal at this level. The 51 root sheath without pathologic changes. b) CT showing focal disc extension with dislocation of the SI root on the left (-). Surgically confirmed.

a

b

Fig. 3. a) Normal myelography. b) CT shows a deep L5/SI disc herniation on the left (-).

243

bulging and 16 in the herniated disc group. Operative verification was available for 23 levels (Table 2). Normal myelography. At 55 intervertebral levels CT was normal or showed only slight disc bulging. Myelography and CT concurred in 89 per cent of cases when the findings were normal. Seven abnormal discs were picked up by CT, four of them revealing a focal disc herniation. Two patients were operated upon and the herniations shown at CT were verified in both cases (Figs I, 2). Two patients were treated conservatively despite clear clinical evidence of root compression (Fig. 3). The three remaining cases showed one extraforaminal disc protrusion with an osteophyte compressing the right L4 ganglion, one broadly bulging disc without root compression, and one narrow spinal canal with a large posterior osteophyte, or perhaps a calcification of the posterior longitudinal ligament, dislocating the left SI root. At myelography the four hernias and one bulging disc escaped detection due to a large epidural space at the L5/S1 level. Two other patients from the normal group were operated upon. In the first case myelography was normal, and CT showed a slight bulge with calcification of the posterior longitudinal ligament. At surgery a slightly bulging L5/S1 disc, with an intact posterior longitudinal ligament was found. In the second case there was a surgically confirmed disc herniation at the L4/L5 level, but at the L5/S1 level both myelography and CT showed normal conditions. At operation, however, a large subligamentar disc protrusion was found at that level. This discrepancy was probably caused by a caudad migration of the large L4/L5 herniation. These patients were considered to be false negatives. Abnormal myelography. CT confirmed all herniations seen at myelography, and disclosed five further herniations, which had been considered to be bulging discs at myelography. Two of these were verified at surgery, and two were included in a conservative treatment trial of disc herniation. The remaining patient could not be operated upon because of other medical problems. A further discrepancy in the interpretation by myelography and by CT was found in two other cases. In the first case myelography showed a deformity of the right S1 root sheath giving rise to suggestion of a herniated disc with root compression. CT, on the other hand, showed no changes at this level, but a disc herniation at the level above (Fig. 4). This was verified by surgery. With the myelogram alone, surgery of two different levels would have been indicated. The second patient showed a bulging L4/L5 disc when examined by both techniques, but CT missed an instability at that level which was picked up by the flexion/extension views at myelography; a stabilizing operation was carried out as a result. As far as pathologic changes were concerned, the two methods concurred in 82 per cent (31 of 38 levels) of the cases. The treatment strategy could have been based on CT alone at 37 of 38 levels (97%), and on myelography

244

M. K. J.

a

FAGERLUND AND U. E. THELANDER

b

c

Fig. 4. a) Myelography shows a disc herniation at the level L41L5 and partial amputation and thickening of the right root

sheath 51 (.....). b) CT at the level of U/LS confirms the herniation (-+). c) CT at the level LS/51 reveals no pathology.

alone at 33 of 38 levels (87%). In the case of disc herniation, both at myelography and CT, the site of the herniation was depicted correctly by both methods. Furthermore, the imaging by CT, of the spinal canal width, soft tissues and intervertebral joints, was superior to that by myelography. A narrow spinal canal was seen at 7 levels by myelography and at 10 levels by CT. General comparison. Based on findings at surgery and clinical signs, the true negative rate for myelography, regarding disc herniation, was 87 per cent (73 of 84 levels). For CT the rate was 97 per cent (73/75). One false positive intervertebral level was found at myelography, CT showed none. At myelography facet joint arthropathy was difficult to detect, whereas CT revealed this at 14levels (23%). In one case the flexion/extension views at myelography showed an instability, whereas CT could not detect this instability.

35). The accuracy of CT is considered to be superior to myelography in detecting disc herniations (9, 25, 30) and CT is advocated as the primary examination method instead of myelography (20, 26, 29). However, in one critical report BELL et coli. (2) concluded that myelography was still more accurate than CT. In that study the findings at myelography and at CT were analyzed on the basis of the findings at surgery. Cases with normal findings and conservative treatment were not included in the analysis. Also, the study suffered from some inherent limitations, such as inclusion of CT examinations from different types of scanners with varying image quality, and perhaps limited experience in interpreting CT of the spine at the beginning of the high resolution CT era (II, 31, 32). The time between myelography and CT was not mentioned, in contrast to the present study, which was designed with a defined time interval of less than two hours. In a recent study (26) myelography was found to be slightly more sensitive than CT, when compared with findings at surgery. The false negative rate, however, was relatively high, in the myelography group, compared with CT. In both reports (2, 26) the conclusions were based on a selected group of patients who had undergone surgery. Such a selection does not take into account the group of patients with normal findings at examination who do not come to surgery. The treatment strategies of non-operated patients with a positive test outcome, were not analysed either. The conclusions in the present study could be criticized because surgical verification was not available for all levels with disc herniation. However, because of a parallel study focusing on conservative treatment methods for disc herniation, the protocol was designed especially for that purpose and carefully followed. The importance of detailed clinical findings in the management of patients with possible disc herniation is stressed by other investi-

(-+)

Discussion Myelography has been the standard radiologic examination method for suggested disc herniation for more than 30 years. The accuracy of myelography compared with surgery is reported to be approximately 85 to 90 per cent at the UIL51evei and 75 to 80 per cent at the LS/SI level (5, 9, 14, 27, 28, 36). The main limitations of myelography include poor visualization of extreme lateral disc herniation (21, 22, 33), and that protrusion of the L5/SI disc escapes detection because of a large lumbosacral epidural space (9, 35). CT, being a non-invasive method of examination, has several advantages over myelography. The intervertebral space LS/SI is imaged more accurately than at myelography (9, 35), the soft tissues are clearly visible, and the radiation dose is lower than that at myelography (8, 24,

LUMBAR

DISC

gators as well (19). Reliance on surgery alone may have certain limitations (12). In the present study it is suggested that one level with false negative findings both at myelography and CT might be explained by confusion at surgery due to a caudad extension of a herniation from the level above. This is indicated also by the fact that actually two CT examinations were performed on this patient on the same day. For the purpose of a different study (7) a postmyelography CT showed the same findings as those seen at the CT conducted before myelography. In two earlier studies (20, 23) a comparison of normal myelograms with CT were also made. The criteria for a normal myelogram, however, were not clearly defined or were not defined at all. This tends to favor CT in the comparison with myelography. The clinical distinction between nerve root compression due to herniated disc and facet joint arthropathy can be difficult. Patients with abnormal intervertebral discs frequently also have abnormal facet joints combined with lateral recess entrapment (24). CT offers an excellent tool for the study of the facet joints. For that purpose a wide bone window setting is needed. It is sometimes possible at myelography to observe intervertebral facet joint arthropathy in the oblique views, but the real extent of the lesion is demonstrated better by CT. The usual findings include hypertrophy and osteophyte formation at the facet joints, subchondral sclerosis, cartilage narrowing and erosions at the facets (3, 4). The distinction between disc herniation and facet joint arthropathy is important because it is possible to treat the facet joints with corticosteroid injections (6). The treatment strategy for the patients examined could have been based on CT findings alone, except in one case, where an instability was demonstrated by myelography only. The patients in our study have now been followed up for approximately 3 to 5 years. Two patients with conservatively treated disc herniations are awaiting operation; none of the patients with normal or bulging discs have undergone surgery. The transaxial imaging mode, limiting the examination to the two or three lowest lumbar levels, is considered to be a potential disadvantage of CT (35). However, 90 to 95 per cent of all disc herniations occur at the two lowest lumbar levels (27, 28, 31). With abnormalities at several levels, as in spinal stenosis, or at the conus region, myelography is still the preferred diagnostic method. In the future, however, magnetic resonance imaging (MRI) may become the primary examination modality for lumbar disc disease (1, 16). MRI also seems to be the best method for the imaging of degenerative disc disease (17). In the present material, only a few disc degenerations were diagnosed at CT. It is assumed that CT underestimated the frequency of disc degeneration (34). The present study showed that myelography could safely be replaced by CT as the primary examination method for suggested lumbar disc herniation. CT was capable of

245

HERNIATION

detecting disc herniation more accurately than myelography and could exclude a herniation at least as well as myelography. Request for reprints: Dr Markku FagerIund, Department of Diagnostic Radiology, Section of Neuroradiology, University Hospital, S-90185 Umea, Sweden.

REFERENCES 1. AGUET J. and LANE B.: Imaging of degenerative disk disease of the lumbar spine. Postgrad. Radiol. 7 (1987), 241. 2. BELL G. R., ROTHMAN R. H., BOOTH R. E. et coli.: A study of computer-assisted tomography. II. Comparison of metrizamide myelography and computed tomography in the diagnosis of herniated lumbar disc and spinal stenosis. Spine 9 (1984), 552. 3. CARRERA G. F.: Computed tomography of the lumbar facet joints. In: Computed tomography of the spine, p. 485. Edited by M. 1. D. Post. Williams & Wilkins, Baltimore, London 1984. 4. - HAUGHTON V. M., SYVERTSEN A. and WILLIAMS A. L.: Computed tomography of the lumbar facet joints. Radiology 134 (1980), 145. 5. COOK P. L. and WISE K.: A correlation of the surgical and radiculographic findings in lumbar disc herniation. Clin. Radiol. 30 (1979), 671. 6. DESTOUET J. M., GILULA L. A., MURPHY W. A. and MONSEES B.: Lumbar facet injection. Indication, technique, clinical correlation, and preliminary results. Radiology 145 (1982), 321. 7. FAGERLUND M. K. J.: Computed tomography in low back pain before and after myelography. A qualitative comparison. Acta Radiol. 29 (1988), 353. 8. FIROOZNIA H., BENJAMIN V., KRICHEFF I. I., RAFII M. and COLIMBU C.: CT of lumbar spine disk herniation. Correlation with surgical findings. Amer. J. Neuroradiol. 5 (1984), 91. 9. FRIES J. W., ABODEELY D. A., VIJUNGCO J. G., YEAGER V. L. and GAFFEY W. R.: Computed tomography of herniated and extruded nucleus pulposus. J. Comput. Assist. Tomogr. 6 (1982), 874. 10. HAUGHTON V. M., ELDEVIK O. P., MAGNAES B. and AMUNDSEN P.: A prospective comparison of computed tomography and myelography in the diagnosis of herniated lumbar disks. Radiology 142 (1982), 103. II. HEiTHOFF K. B.: Editorial: On CT scanning and metrizamide myelography. Spine 10 (1985),692. 12. HIRSCH D. and NACHEMSON A.: The reliability of lumbar disk surgery. Clin. Orthop. 29 (1963), 189. 13. KIEFFER S. A., CACAYORIN E. D. and SHERRY R. G.: The radiological diagnosis of herniated lumbar intervertebral disk. A current controversy. J. Amer. Med. Ass. 251 (1984), 1192. 14. - SHERRY R. G., WELLENSTEIN D. E. and KING R. B.: Bulging lumbar intervertebral disk. Myelographic differentiation from herniated disk with nerve root compression. Amer. J. Roentgenol. 138 (1982), 709. 15. LUKIN R. R., GASKILL M. F. and WIOT J. G.: Lumbar herniated disc and related topics. Sem. Roentgenol. 23 (1988), 100. 16. MODIC M. T., MASARYK T., BOUMPHREY F., GOORMASTIC M. and BELL G.: Lumbar herniated disk disease and canal stenosis. Prospective evaluation by surface coil MR, CT, and myelography. Amer. J. Neuroradiol. 7 (1986),709. 17. - PAVLICEK W., WEINSTEIN M. A. et coli.: Magnetic resonance imaging of intervertebral disc disease. Clinical and pulse sequence considerations. Radiology 152 (1984), 103. 18. MORINGLANE J. R., VOIGT K. und SEEGER W.: Vergleich

246

19.

20.

21.

22.

23.

24.

25.

26.

27.

M. K. 1. FAGERLUND AND U. E. THELANDER

myelographisher und intraoperativer Befunde beim lumbalen Bandscheibenvorfall. Neurochirurgia 20 (1977), 199. MORRIS E. W., DI PAOLA M., VALLANCE R. and WADDELL G.: Diagnosis and decision making in lumbar disc prolapse and nerve entrapment. Spine 11 (1986), 436. MOUFARRIJ N., HARDY R. and WEINSTEIN M.: Computed tomographic, myelographic, and operative findings in patients with suspected herniated lumbar discs. Neurosurgery 12 (1983), 184. NELSON M. J. and GOLD L. H. A.: CT evaluation ofintervertebral foramina lesions with normal or non-diagnostic myelograms. Report of ten cases. Comput. Radiol. 7 (1983), 155. NOVETSKY G. J., BERLIN L., EpSTEIN A. J., LOBO N. and MILLER S. H.: The extraforaminal herniated disk. Detection by computed tomography. Amer. 1. Neuroradiol. 3 (1982), 653. RASKIN S. P. and KEATING J. W.: Recognition of lumbar disk desease. Comparison of myelography and computed tomography. Amer. J. Neuroradiol. 3 (1982), 215. RISIUS B., MODIC M. T., HARDY JR R. W., DUCHESNAU P. M. and WEINSTEIN M. A.: Sector computed tomographic spine scanning in the diagnosis of lumbar nerve root entrapment. Radiology 143 (1982), 109. SACHSENHEIMER W., HAMER J. and MULLER H. A.: The value of spinal computed tomography in diagnosis of herniated lumbar disks. Acta Neurochir. 60 (1982), 107. SCHIPPER J., KARDAUN J. W. P. F., BRAAKMAN R., VAN DONGEN K. J. and BLAAUW G.: Lumbar disk herniation. Diagnosis with CT or myelography? Radiology 165 (1987), 227. SHAPIRO R.: Myelography. Fourth edition, p. 422. Year Book Medical Publishers, Chicago 1984.

28. SPANGFORT E.: The lumbar disk herniation. A computer-aided analysis of 2504 operations. Acta Orthop. Scand. (1972) Suppl. No. 142. 29. STRATEMEIER P. H.: Evaluation of the lumbar spine. A comparison between computed tomography and myelography. Radiol. Clin. N. Amer. 21 (1983), 221. 30. TCHANG S., HOWIE J., KIRKALDy-WILLIS W., PAINE K. and MOOLA D.: Computed tomography versus myelography in diagnosis of lumbar disk herniation. J. Canad. Assoc. Radiol. 33 (1982), 15. 31. TEPLICK J. G. and HASKIN M. E.: CT and lumbar disc herniation. Radiol. Clin. N. Amer. 21 (1983), 259. 32. WIESEL S. W., TSOURMAS N., FEFFER H. L., CITRIN C. M. and PATRONAS N.: A study of computer-assisted tomography. 1. The incidence of positive CAT scans in an asymptomatic group of patients. Spine 9 (1984), 549. 33. WILLIAMS A. L., HAUGHTON V. M., DANIELS D. L. and THORNTON R. S.: CT recognition of lateral lumbar disk herniation. Amer. J. Neuroradiol. 3 (1982), 211. 34. - - MEYER G. A. and Ho K. C.: Computed tomographic appearance of the bulging annulus. Radiology 142 (1982), 403. 35. - - and SYVERTSEN A.: Computed tomography in the diagnosis of herniated nucleus pulposus. Radiology 135 (1980), 95. 36. WRIGHT F. W., SANDERS R. C., STEEL W. M. and O'CONNOR B. T.: Some observations on the value and techniques of myelography in lumbar disc lesions. The results over a five year period at the Nuffield Orthopaedic Centre, Oxford. Clin. Radiol. 22 (1971), 33.