We evaluated magnetic resonance imaging studies of thirty patients before and after a contained herniation of a lumbar disc was treated with a percutaneous lumbar discectomy. The imaging studies were evaluated to determine whether the preoperative appearance of the herniated disc was predictive of the outcome of percutaneous discectomy and also to determine a possible mechanism of action of the procedure in the relief of symptoms. The index operation was successful in seventeen (57 per cent) of the thirty patients. The preoperative imaging studies showed no differences in the appearance of the discs that went on to have a successful result and those that went on to have an unsuccessful result. Imaging studies made four to six weeks after the operation showed no measurable changes in the morphology of the disc. Imaging studies made a mean of fourteen months after the operation showed no changes in the morphology of the disc in twenty-four (80 per cent) of the patients, irrespective of the clinical outcome. Only three of the seventeen patients who had a successful result had a reduction of more than two millimeters in the size of the herniated segment, and two of the thirteen patients who had an unsuccessful result had an increase of more than one millimeter in the size of the herniated segment. We found that preoperative imaging studies cannot predict the clinical outcome of percutaneous lumbar discectomy.(ABSTRACT TRUNCATED AT 250 WORDS)