The sensitivity, specificity, and accuracy of computed tomography (CT)-myelography, magnetic resonance imaging (MRI), and myelography in making the diagnosis of herniated nucleus pulposus (HNP) and spinal stenosis were compared in a retrospective study involving 59 surgical procedures in 57 patients who had all three tests performed preoperatively. One hundred nineteen levels were surgically explored for evidence of HNP and spinal stenosis. The results of each test were correlated with what was found at each surgical level explored. Overall, myelo-CT was the most accurate test for diagnosing HNP (76.4%) as well as the most sensitive (77.8%), whereas myelography was the most specific (89.2%). In making the diagnosis of spinal stenosis, myelo-CT and MRI were equally accurate (85.3%) and sensitive (87.2%), whereas myelography was the most specific (88.9%). In a special subset of patients who had revision surgery, the accuracy rate in diagnosing spinal stenosis or HNP was highest with MRI (84.9%), as was the sensitivity (69.2%) and specificity (95%). According to the results obtained from this series of patients, myelo-CT seems to be the most sensitive and accurate test in diagnosing HNP and spinal stenosis, whereas myelography is the most specific, although no statistical significance was noted in this study. However, because MRI did compare favorably with myelo-CT in most instances, particularly in revision surgery, it may be the procedure of choice due to its noninvasiveness and relative lack of side effects.