Which disc as visualized by magnetic resonance imaging is actually a source of pain? A correlation between magnetic resonance imaging and discography.

Sixty-three discs in 25 consecutive patients with nonradicular discogenic pain were studied with magnetic resonance imaging and lumbar discography. Sagittal T2 magnetic resonance imaging sequences were blindly classified according to the nuclear signal intensity and to the status of the posterior anulus. A positive discogram had to include abnormal morphology and a provocative pain response of significant intensity and familiar quality. Specific magnetic resonance imaging patterns with a high probability for positive and negative discograms are identified. Also identified are magnetic resonance imaging patterns with intermediate probability, ie, ones that should not be presumed either asymptomatic or symptomatic. In many cases, magnetic resonance imaging could not reliably predict or replace discography.