In the investigation of patients with internal disc disruption, a discography has been the best method to decide whether intervertebral disc is normal or abnormal. Magnetic resonance imaging (MRI) of the spine is a new technique which offers an alternative method of obtaining this information. Some authors have reported MRI was more accurate than discography, but other reports have indicated discography was more useful test in symptomatic lumbar disc disease. The authors performed retrospective study to evaluate the diagnostic value of the image findings of MRI and discography in the internal disc disruption. Image findings of MRI and discography were analyzed comparing to pattern of provocative pain during discography. The concordant pain at discography was considered as a gold standard for evaluation of the diagnostic value of these modalities. The image findings of MRI and discography for 60 discs of 21 patients were graded and compared with the pain response at discography. The validity and predictive value (PV) of these modalities were calculated and the correlation between these image findings was evaluated. The sensitivities of discography and MRI were 83.3% and 90.9% respectively. The specificities of these findings were 47.1% and 46.7%. The positive PV of discography and MRI were 52.6% and 55.6%, and the negative PV were 80% and 87.5% respectively. There was no statistical significance in these discrepancies. The determinations of symptomatic disc level based on image findings are unreliable. The MRI can be considered as a screening test for internal disc disruption since it is non-invasive and sensitive diagnostic tool. The fusion level should be decided by the pain pattern at discography.