Comparison of magnetic resonance imaging and conventional neuroradiographies in lumbar herniated nucleus pulposus

To determine whether magnetic resonance imaging (MRI) would replace conventional neuroradiographies in lumbar herniated nucleus pulposus, preoperative MRI findings were compared with surgery-confirmed pathophysiology of 63 intervertebral disks in 58 patients. Conventional neuroradiographies consisted of myelography, CT myelography, discography, and CT discography. Pathophysiology of 63 herniated disks fall into normal (n=7), bulging (n=9), protrusion (n=14), extrusion (n=17), and free migrated (n=16). Diagnostic accuracy of MRI was evaluated in terms of the presence or absence of herniation, height, location, and morphology of herniated disk. In diagnosing herniation, MRI had a sensitivity of 96%, a specificity of 25%, and an accuracy of 78%. In determining the height of herniated disk, the diagnostic rate of MRI was 77%, being lower than both CT myelography (94%) and CT discography (89%). MRI had a concordance rate of 55% for bilateral location in transaxial view, compared with 76% for CT myelography and 74% for CT discography. MRI failed to differentiate extrusion from protrusion. In conclusion, MRI was the most suitable for screening the presence or absence of herniation; however, it was inferior to other neuroradiographies in the diagnosis of morphology of herniated disk, as well as the determination of height and surgical intervention site of herniated disk. (N.K.)