PAINFUL DISC LESION: CAN MODERN BIPLANAR MAGNETIC RESONANCE IMAGING REPLACE DISCOGRAPHY?

Abstract

This prospective study was carried out to correlate findings of magnetic resonance imaging (MRI) and discography. Fifty-five consecutive patients with degenerative disc disease not responding to non-operative treatment were included in the study. There were 19 men and 36 women and the mean age was 45 years. Discography was carried out on 131 disc levels. The discograms were classified using modified Adams’s classification and pain recorded into three grades. MRI scans were graded using a new classification system based on parasagittal and axial images by two independent observers blinded to discography findings. There was good intraobserver (kappa 0. 74) and interobserver (kappa 0. 70) agreement for the classification system. There was a significant correlation in the morphology of discs as determined by discograms and MRI classification (p< 0. 001).

Each disc was graded on MRI scan as painful or painless on basis of defined criteria. Concordant discography pain was considered as the gold standard. The sensitivity and specificity of MRI in predicting symptomatic disc using defined criteria was 94% and 77%. The sensitivities and specificity of high intensity zones was 27% and 87% and for end plate changes was 32% and 98% respectively. In 14 patients (25%) the findings of MRI and discography did not correlate.

In conclusion though MRI is an excellent investigation for assessing disc morphology it should be interpreted along with discography findings before planning fusion surgery. The proposed MRI classification is a useful aid in predicting painful degenerative disc. The utility of high intensity zones and end plate changes is limited due to low sensitivity.