Neck Solutions

Painful Disc Lesion: Can Modern Biplanar Magnetic Resonance Imaging Replace Discography?

From: J Spinal Disord Tech 2008; 21:430–435

Internal disc disruption’’ is one of the prominent somatic sources of low back pain. MRI is the most common investigation performed to evaluate discogenic pain. Though MRI has advantages of being a sensitive investigation for identifying pathologic anatomy of disc degeneration, its value is limited by its inability to evaluate the physiologic status of the disc. Studies have questioned the specificity of MRI in diagnosing discogenic pain and have reported degenerative changes in 26% to 57% of asymptomatic volunteers. Discography is a useful tool to identify the painful disc responsible for patient’s symptoms. In evaluation of discogenic low back pain there has been a debate as to whether discography should replace MRI. Proponents of discography believe that it is an invaluable tool for identifying the pathologic disc producing pain. Critics believe that discography is an invasive investigation and has no place in modern practice.

Over the last decade there has been a vast improvement in the quality of MR images. Attempts have been made to identify features on MRI scans that would correlate with patient’s symptoms and would potentially eliminate the need for discography. Such features are the vertebral end plate changes described by Modic and the high intensity zone described by Aprill and Bogduk. The implications of these findings have been a subject of controversy.

This is a prospective study aimed at correlating the results of discography with new MRI classification in a consecutive series of patients with disabling low back pain considered for spinal surgery. In addition, we also aimed to compare vertebral end plate changes and high intensity zones seen on MRI with discography findings.

Woodend MRI Classification
Grade 1: White nuclear signal, normal height bean shape nucleus, annular margins well defined, no tears.
Grade 2: Speckled nuclear signal, height reduced less than 10%,distortion of nuclear shape, small radial tears not reaching the PLL on axial views.
Grade 3: Speckled or dark nucleus, height reduced by 10%-50%, radial tears extending upto or torn PLL on sagittal/axial views.
Grade 4: Dark nucleus, height reduced by greater than 50%, no difference between appearance of annulus and nucleus ± complex tears.

Disc Morphology Grading
Grade 1: cotton ball or lobular appearance of disc with no tears.
Grade 2: disc with fissures or clefts in the nucleus or inner annulus.
Grade 3: degenerate disc with radial fissures leading to outer edge of the annulus.
Grade 4: disc with complete radial fissure.

Internal disc disruption was described by Crock as a condition causing back pain without nerve root compression. Tears within the annulus fibrosus are a major hallmark of degenerative intervertebral disc disease. Tears within the annulus evoke an inflammatory response, which leads to degradation of disc matrix and chemical irritation of nerve fibers within the outer annulus. Discogenic pain is a mechanical, nonradicular pain with a somatic pattern. Therefore, the use of symptoms and signs to identify the symptomatic degenerative disc level has limited value. A method for investigating discogenic pain cannot be considered reliable unless there is a confirmed relationship between its findings and a patient’s symptomatology

Lindbolm first introduced discography in 1944. Holt raised the first major challenge to this technique in 1968. Holt reported false-positive findings in 37% of patients in a study on asymptomatic prison inmates. However, following a study carried out by Walsh et al in 1990 under more stringent conditions, doubts were raised regarding Holt’s findings. Walsh et al carried out discography in 10 asymptomatic volunteers and reported no false-positive result. Adams et al performed a cadaveric study and proposed a discography classification, which has been used, in a modified form, in this study. The Dallas Discogram Description described by Sachs et al is based on computed tomography scan after discography. We believe that this exposes patient to unnecessary radiation and increases cost without any additional benefit. The morphology of the disc as revealed by discography is irrelevant. In our study, we have used concordant pain on discography as a guide to the pain source. The analysis for sensitivity and specificity has been performed after excluding discs with discordant pain. As the investigation is performed under local anesthetic we believe that the patient’s response can be reliable.

In the late 1980s, when MRI became popular among spinal surgeons, a number of studies questioned the utility of discography in discogenic pain. Gibson et al reported 88% correlation between findings of MRI and discography in 50 patients. Discography was carried out under sedation and the authors believed that this was an unreliable sign. Schneiderman et al concluded that MRI was 99% accurate as compared with discography. In both these studies the correlation of MRI was with disc morphology rather than pain. Simmons et al reported that 37% of the discs reported abnormal on MRI were asymptomatic on awake discography. Linson and Crowe reported 94% correlation between MRI and discography. In their study, only 53% of the degenerative discs were found to be symptomatic. They concluded that discography remains the examination of choice for differentiating between symptomatic and asymptomatic discs. Collins et al reported morphologic correlation between discography and MRI in 89.5% cases. Only 13 of 73 discs were symptomatic and 10 had degeneration with annular bulges on MRI. Osti et al in a study on 108 intervertebral discs found that of the 39 symptomatic discs only 27 had abnormal signals on MRI. They concluded that discography is a more accurate investigation to detect symptomatic disc. Horton and Daftari correlated disc signals and annular bulges with pain reproduced at discography and concluded that MRI does not reliably predict or replace discography.

In this study, we have attempted to use an MRI classification system that will help predicting pain and allow comparison of various disc appearances. This classification embraces various aspects of disc morphology such as MRI signal, disc height, and annular tear in combination rather than considering them as individual entities. The classification uses the carefully aligned axial and sagittal images on MRI, unlike most of the previous studies in which only sagittal scans were used. We now use these scans routinely. The interobserver and intraobserver reliability of the classification system is good. The classification system can be used to communicate status of disc morphology and has a potential as a research tool. There is good correlation between the grades of degenerative disc on discography and the MRI classification system. In our study, we also found that grades 3 and 4 disc morphology has a 81% sensitivity and 80% specificity in predicting discogenic pain. We postulated that annular tears would predict a disc to be painful and included these as one of the criteria for pain prediction on MRI. These tears are best visualized on axial scans. The majority of the discs with annular tears were grade 3 or grade 4. Of the 25 grade 2 discs 14 had annular tears and 12 produced concordant or discordant pain. The presence of grade 3 or grade 4 disc on MRI scan or grade 2 with annular tear increased the sensitivity of MRI scan in detecting a painful disc to 94%.

There has been considerable interest in high intensity zones as predictors of painful lumbar intervertebral discs. Aprill and Bogduk who considered it to represent inflammation in the annulus secondary to trapped nuclear material and reported a sensitivity of 82% and specificity of 89%. Schellhas et al and Lam et al found high intensity zone to be a reliable predictor. Ricketson et al11 did not find significant correlation between high intensity zone and concordant disc pain. Ito et al described a sensitivity of 52% and specificity of 90% for concordant disc pain. The sensitivity and specificity of high intensity zone in our study is close to those reported by Saifuddin et al and Smith et al. Vertebral end plate has been identified as a possible source of low back pain. Braithwaite et al and Ito et al reported low sensitivity and high specificity for end plate changes which closely corresponds to that in our study.

In 25% of the patients, the findings of MRI and discography did not correlate. This is due to low specificity of MRI in detecting the painful level. There were 3 discs, which were grade 1 on MRI and produced concordant pain and 12 grade 3 or grade 4 discs, which were not painful. We agree with Simmons et al that, in current practice, discography has a place as it can complement the interpretation of MRI findings and help to determine management of chronic low back pain. There were 18 discs, which produced discordant pain in our study. A disc with discordant pain is still painful disc though not necessarily causing the patient’s major symptoms.

In our study, the group of patients who had previous surgery had similar discography and MRI correlation. We find MRI is also useful in evaluation of this group of patients with previous spinal surgery. Improvements in biomaterials and scanning techniques allow good quality MRI scans in patients with previous instrumentation and artifacts do not hamper assessment of disc morphology.

The use of discography has a limitation that pain on discography is subjective and dependent on interaction between patients and surgeons/radiologists, and has to be interpreted as such. We studied a selected group of patients considered for spinal surgery. We believe that discography is a preoperative investigation and should be performed only when the diagnosis of discogenic pain is highly probable as determined by systematic clinical examination, nonoperative treatment has failed, and a decision is made to proceed to surgery. Our study does not address the implications for management of patients. This cohort of patients is being followed prospectively and their long-term outcome will answer whether these findings have implications for patient’s quality of life outcome after surgical treatment. We also believe that in future techniques, which can stimulate disc in noninvasive manner while performing MRI, will benefit assessment of discogenic pain. This will improve the specificity of MRI. Yrjama et al have reported such techniques but these have not been reported from other centers.