IntroductionVertebral end-plate (Modic) changes were first described independently by de Roos et al. [9] and Modic et al. [15] as being a feature associated with degenerative intervertebral disc disease. Type I changes consist of reduced signal intensity (SI) in the vertebral end-plates on T1-and increased SI on T2-weighted sequences (Fig. 1). They are associated with fissuring of the cartilaginous end-plate and increased vascularity within the subchondral bone marrow on histological examination. Type II changes consist of increased SI on T1-and either increased SI or isointensity on T2-weighted sequences (Fig. 2). In such cases, biopsy reveals fatty replacement of the marrow [15], which is thought to be the result of marrow ischaemia [9]. Type III changes consist of reduced SI on both T1-and T2-weighted sequences due to subchondral sclerosis ( Fig. 3). Type I changes commonly progress to Type II changes and rarely revert back to normal, whereas Type II changes appear not to change with time [15]. Modic changes are identified in 20-50% of patients, with the incidence increasing with age [9,15]. However, it is not known why some degenerative discs are associated with Modic changes while others are not.Crock [6,7] proposed the concept of "internal disc disruption", suggesting that repeated trauma to the intervertebral disc could result in the production of inflammatory substances within the nucleus pulposus. Diffusion of such toxic chemicals through the vertebral end-plate could then result in a local inflammatory reaction resulting in back pain.Inflammation in the subchondral bone adjacent to the end-plate would result in reduced SI on T1-weighted MRI sequences and increased SI on T2-weighted MRI sequences, equivalent to a Type 1 Modic change. The possiAbstract The vertebral end-plate has been identified as a possible source of discogenic low back pain. MRI demonstrates end-plate (Modic) changes in 20-50% of patients with low back pain. The aim of this study was to investigate the association between Modic changes on MRI and discogenic back pain on lumbar discography. The MRI studies and discograms of 58 patients with a clinical diagnosis of discogenic back pain were reviewed and the presence of a Modic change was correlated with pain reproduction at 152 disc levels. Twenty-three discs with adjacent Modic changes were injected, 21 of which were associated with pain reproduction. However, pain was also reproduced at 69 levels where no Modic change was seen. The sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV) for a Modic change as a marker of a painful disc were 23.3%, 96.8%, 91.3% and 46.5% respectively. Modic changes, therefore, appear to be a relatively specific but insensitive sign of a painful lumbar disc in patients with discogenic low back pain.