MRI AND DISCOGRAPHY OF ANNULAR TEARS AND INTERVERTEBRAL DISC DEGENERATION

MRI AND DISCOGRAPHY OF ANNULAR TEARS AND INTERVERTEBRAL DISC DEGENERATION A PROSPECTIVE CLINICAL COMPARISON 0. L. OSTI, R. D. FRASER From the Royal Adelaide Hospital, South Australia We attempted to correlate the findings of MRI and discography in patients with low back pain, examining 108 lumbar intervertebral discs in 33 consecutive patients. MRI results were assessed from the intensity and shape of the signal obtained from the central part of the disc. Discography was classified according to the pattern of contrast material, the pressure accepted and the pain reproduced. All discs which were abnormal on MRI had altered patterns on discography, but 18 of the 60 discs with normal MRI had abnormal discograms. Of 39 asymptomatic discs, 33 had normal MRI signals and 24 had normal discograms. None of the 15 discs showing severe degeneration on MRI sustained high levels of intradiscal pressure, but only six of the 60 discs giving normal MRI had low pressure. With current techniques, discography is more accurate than MRI for the detection of annular pathology: a normal MRI does not exclude significant changes in the peripheral structure ofthe intervertebral disc which can produce low back pain. Tears within the annulus fibrosus are one of the main features of intervertebral disc degeneration (Schmor! and Junghanns 1971 ; Vernon-Roberts 1987). Discography has been used extensively to identify such s (Holt 1968 ; North American Spine Society 1988 ; Esses, Botsford and Kostuik 1989; Nachemson 1989) and to assess their clinical significance in patients with low back pain (Crock 1983; Co!houn et a! 1988). Magnetic resonance imaging (MRI) can show degeneration of the intervertebral disc by the analysis of signal intensity obtained from the nucleus pulposus and the inner annulus (Modic and Weinstein 1984; Pech and Haughton 1985; Gibson et a! 1986; Schneiderman et a! 1987), and it has been suggested that special techniques may reveal s within the annulus fibrosus (Yu et a! 1988). We have made a prospective analysis of the correlation between MR appearance and the results of discography in a consecutive series of patients with low back pain. PATIENTS AND METHODS We reviewed patients investigated for low back pain from May 1987 to December 1988, excluding those with 0. L. Osti, MD, PhD, Orthopaedic Registrar R. D. Fraser, MD, FRACS, Head ofthe Spinal Unit Spinal Unit, Royal Adelaide Hospital, North Terrace, Adelaide, South Australia 5000. Correspondence should be sent to Dr 0. L. Osti. 1992 British Editorial Society ofbone and Joint Surgery 030l-62OX/92/33S1 $2.00 JBoneJoint Surg[Br] 1992; 74-B:431-S. spinal osteomyelitis, discitis or other infectious diseases, trauma or tumour. There were 33 patients, in whom we studied a total of 1 14 lumbar intervertebral discs. The mean age of the patients was 35 years (24 to 64). For MRI we used a Siemens Magnetom 1.0 Tes!a machine (Siemens AG, Erlangen, Germany). T2- weighted spin echo sagitta! images of 5 mm thickness were obtained using a TR (repetition time) of 2.5 sec and a TE (echo time) of 90 msec. The images of the intervertebral discs were classified according to the signal intensity from the central zone. This was recorded as normal, reduced, or absent when no difference could be seen between the central and peripheral zones in central sagittal slabs ( 1). We tried to classify the shape of the central signal as normal, enlarged anteriorly, enlarged posteriorly, or both. This distinction proved to be extremely difficult in most cases, and we therefore limited the MRI classification to intensity only. We performed discography after MRI, using a posterolatera! approach and a two-needle technique with stilettes. The radiographic contrast material was mixed with cefazolin as a prophylactic measure to prevent discitis (Osti, Fraser and Vernon-Roberts 1990a). Discographic patterns were classified as follows: Type 1. Normal : the nucleus was well outlined in a roundish, regular, simple or bilobulated pattern and no annular s were seen. Type 2. The nucleus was we!! outlined but an anterior or posterior extended to the inner or outer annulus, with no leaking into the epidura! space. Type 3. The nucleus was well outlined, but one or more s extended to the outer annulus anteriorly or VOL. 74-B, No. 3, MAY 1992 431

R. D. FRASER posterio space Type4. extended leakage (when clearly (when or Pain T in in d ir Sagittal T2-weighted image of human lumbar spine (TR 2.5 : TE 90) to show grading by the signal intensity from the disc. The upper disc is graded normal, the middle reduced and the lower absent. Discogram patterns. Figure 2a - L4-L5 is normal, type 1 ; LS-S1 shows a posterior annular, type 2. Figure 2b - At L3-L4, there is physiological pooling in the centre of the disc but the contrast is also outlining the inner and middle fibres of the anterior annulus. There is a cleft running towards the upper end-plate of the lower vertebra. This is a type-2 pattern. Figure 2c - At L4-LS there is extravasation of contrast posteriorly along the perispinal ligaments and into the epidural space. This is a type-3 pattern. Figure 2d - At LS-S1 there is no pooling of contrast, and multiple s anteriorly and posteriorly. This is a type-4 pattern. 2d THE JOURNAL OF BONE AND JOINT SURGERY

MRI AND DISCOGRAPHY OF ANNULAR TEARS AND INTERVERTEBRAL DISC DEGENERATION 433 intradiscal pressure, while only six of the 60 with normal MRI had low pressures. Of the 39 discs which caused typical pain on discography, 27 had abnormal signals on MRI. All the Table I. Comparison of results of MRI and discography in 114 lumbar discs symptomatic discs had degenerative contrast patterns at discography. Conversely, 33 of the 39 asymptomatic discs had normal MRI signals, and only 24 had normal discographic patterns. Only six of the 46 discs diagnosed as degenerate on MRI were asymptomatic at discography. MRI signal intensity Number ofdiscs Discogra jthy type5 1 2 3 4 Normal Inner Outer Normal 60 27 15 12 6 Decreased 39-9 15 15 Absent 15 - - 6 9 see text Marked degeneration DISCUSSION Cloward (1952) first discussed the diagnostic value of discography for highlighting s and disorganisation of the annu!us fibrosus in cases of ruptured lumbar intervertebral disc. Brodsky and Binder (1979) suggested that pain produced by intradiscal injection was due to the transmission of pressure through a torn or weakened annulus to stretch its outer fibres. Park et a! (1979) 3a 3b Comparison of discograms and MRI at L4-L5 and L5-Sl. Figure 3a - The discogram shows marked extravasation of contrast anteriorly and posteriorly at L4-L5 reaching the epidural space and outlining the most peripheral anterior and posterior layers ofthe annulus fibrosus (type 3). At L5-S1 there is a type-4 pattern. Figure 3b - On MRI ofthe same case, the L4-L5 level shows some distortion ofthe central signal, and the L5-51 disc shows absence of signal, except for a small area in the posterior annulus. This is consistent with the type-4 pattern seen at discography. 4a 4b Comparison o and L4-L5 levels. there is e posterolateral. Figure 4b - The? any of the three levels. L.?L.J a e 4a - vasation of contrast into 1 posterior I scan shows no obvious abnormality in signal VOL. 74-B, No. 3, MAY 1992

434 0. L. OSTI, R. D. FRASER Table II. Comparison of results of MRI and discography in 1 14 lumbar discs Discogra thy type Pain at discography Number ofdiscs 1 2 3 4 MRI signal intensity Normal Inner Outer Marked degeneration Nonnal Decreased Absent None 39 24 6-9 33 3 3 Atypical 36 3 15 6 12 15 18 3 Typical 39-3 27 9 12 18 9 reported on 14 young adult patients with discographic evidence of posterior annular fissures extending from an otherwise normally outlined nucleus, and postulated that radiating fissures could be early evidence of disc injury. They also considered that pain provocation could be the result of transference of increased nucleus tension to the outer annulus, which would not occur if there was a complete radial rupture. Adams, Dolan and Hutton (1986) correlated discographic patterns with stages of degeneration in cadaver material. They also considered that radiating clefts with no leakage of contrast material into the epidura! space would be commonly found in patients with back pain and would often be associated with typical symptoms during the injection. In our series 36 of the 39 discs with typical pain at injection showed s extending to the outer annulus. None of the 27 discograms showing a normal pattern produced typical pain : 24 of them caused no discomfort. We were unable to correlate intradiscal pressure with pain. Most normal discs accepted a high intradiscal pressure with no pain, but three of the nine discograms with a low pressure reproduced the patients typical pain. MRI is now an alternative method of obtaining information on the state ofthe intervertebraldisc. Gibson et a! (1986) compared the findings of MRI and discography in five discs, and concluded that MRI was more accurate in the diagnosis of degeneration. Schneiderman et a! (1987) correlated MRI and discography of 101 discs in 36 patients. MRI was accurate in predicting normal or abnormal disc morphology at 100 levels, but MRI signal intensity alone could not determine the exact pattern seen at discography. They also found that in patients with early symptoms due to acute herniation or s of the annulus, MRI may not show changes in signal intensity. Zucherman et a! (1988) reporting on 18 patients also considered that MRI did not accurately reflect the internal disc morphology, and may not detect significant pathology. Our results suggest that, using current standard techniques, MRI fails to demonstrate some of the structural changes in the annulus which would be seen on discography. A normal MRI signal intensity from the central zone is not proof of normal disc morphology. In an animal model of disc degeneration we found that MRI using standard T2 spin echo sequences failed to detect changes in signal intensity until 1,2 months after the initial lesion (Osti, Fraser and Vernon-Roberts 1990b). This confirmed previous experimental studies which showed that after an acute failure of the annu!us the content of nuclear proteoglycan of the disc shows a temporary increase (Lipson and Muir 1981), which is associated with high values of hydration and therefore with a normal MRI signal. We found no false-positive results on MRI, and it may therefore be a useful screening procedure for disabling disc disease. Discography should be considered, as suggested by Zucherman et a! (1988), for patients with normal MRI and continuing severe symptoms. Yu et a! (1988) comparing anatomical specimens and MRI of human intervertebral discs, concluded that T2-weighted image sequences showed different types of within the annulus fibrosus. This study was based on cadaver material, but it is likely that continuing technological improvements and the use of enhancing contrast agents may soon improve the quality of the MRI signal from the annulus. This may allow the visualisation of annular s in the absence of major degenerative changes in the nucleus pulposus. We conclude that discography is, at present, the more accurate investigation for annular s which are likely to produce low back pain. The authors wish to acknowledge the Adelaide Bone and Joint Research Foundation and the South Australian Government Insurance Commission for their financial support and Ms Christine Corke and Mrs Chrissie Muecke for their secretarial help. Although none of the authors have received or will receive benefits for personal or professional use from a commercial party related directly or indirectly to the subject of this article, benefits have been or will be received but are directed solely to a research fund, foundation, educational institution, or other non-profit institution with which one or more of the authors is associated. REFERENCES Adams MA, Dolan P, Hutton WC. The stages of disc degeneration as revealed by discograms. J Bone Joint Surg [Br] 1986; 68-B :36-41. Brodsky AE, Binder WF. Lumbar discography : its value in diagnosis and treatment oflumbar disc lesions. Spine 1979; 4:110-20. Ooward RB. Anterior herniation of a ruptured lumbar intervertebral disk. Arch Surg 1952; 64:457-63. Cothoun E, McCall 1W, Williams L, Cassar Pullicino VN. Provocation discography as a guide to planning operations on the spine. J Bone Joint Surg [Br] 1988 ; 70-B :267-71. Crock HV. Practice ofspina/surgery. Wien etc : Springer-Verlag, 1983. THE JOURNAL OF BONE AND JOINT SURGERY

MRI AND DISCOGRAPHY OF ANNULAR TEARS AND INTERVERTEBRAL DISC DEGENERATION 435 Esses SI, Botsford DJ, Kostuik JP. The role of external spinal skeletal fixation in the assessment of low-back disorders. Spine 1989; 14:94-601. Gibson MJ, Buckley J, Mawhinney R, Mulliolland RC, Worthington BS. Magnetic resonance imaging and discography in the diagnosis of disc degeneration. J Bone Joint Surg [Br] 1986; 68-B :369-73. Holt EP Jr. A question of discography. J Bone Joint Surg [Am] 1968; 50-A :720-6. Lipson SJ, Muir H. Proteoglycans in experimental intervertebral disc degeneration. Spine 1981 ; 6:194-210. Medic MT, Weinstein MA. Nuclear magnetic resonance of the spine. Br MedBu// 1984; 40:183-6. Nachemson A. (Editorial comment) lumbar discography : where are we today? Spine 1989; 14:555-7. North American Spine Society. Position statement on discography. Spine 1988; 13:1343. Osti OL, Fraser RD, Vernon-Roberts B. Discitis after discography : the role of prophylactic antibiotics. J Bone Joint Surg [Br] 1990a; 72- B :271-4. Osti OL, Fraser RD, Vernon-Roberts B. An animal model of intervertebral disc degeneration. Procs ofthe International Society for the Lumbar Spine, Boston l990b:34. Park WM, McCall 1W, O Brien JP, Webb JK. Fissuring of the postenor annulus fibrosus in the lumbar spine. Br J Radiol 1979 ; 2: 382-7. Pech P, Haughton VM. Lumbar intervertebral disk : correlative MR and anatomic study. Radio/ogy 1985; 156 :699-701. Schmorl G, Junghanns H. The human spine in health and disease. Fifth German edition by Junghanns H. Translated and edited by Besemann EF. New York, etc: Grune & Stratton, 1971. Schneiderman G, Flannigan B, Kingston 5, et al. Magnetic resonance imaging in the diagnosis of disc degeneration : correlation with discography. Spine 1987; 12:276-8 1. Vernon-Roberts B. Pathology of intervertebral discs and apophyseii joints. In : Jayson MIV, ed. The lumbar spine and back pain. Third edition. Edinburgh, etc: Churchill Livingstone, 1987 :37-55. Yu SW, Sether LA, Ho PS, Wagner M, Haughton VM. Tears of the annulus fibrosus : correlation between MR and pathologic findings in cadavers. AJN 1988; 9:367-70. Zucherman J, Derby R, Hsu K, et al Normal magnetic resonance imaging with abnormal discography. Spine 1988 ; 13:1355-9. VOL. 74-B, No. 3, MAY 1992