STUDY DESIGN: Evaluation of magnetic resonance images (MRIs) with surgical reference standard. OBJECTIVES: To determine whether the addition of contrast-enhanced MRI scans increases diagnostic efficacy in the evaluation of recurrent disc herniation. SUMMARY OF BACKGROUND DATA: Many centers now routinely use gadolinium-enhanced examinations in the evaluation of recurrent disc herniation. Others, noting the additional expense of contrast injection, advocate a more limited role for contrast injection and emphasize the importance of T2-weighted axial sequences. METHODS: The study included 165 consecutive patients who were referred to the authors' outpatient imaging center and had a history of previous lumbar discectomy and recurrent back and/or leg pain. The scanning protocol included sagittal and axial T1-weighted spin-echo pre- and postcontrast injection images and sagittal and axial T2-weighted fast spin-echo images. Twenty-eight patients (32 vertebral levels) had subsequent surgical exploration of a disc margin that had previously undergone discectomy. The surgical findings formed the reference standard. Three spine radiologists interpreted the MRI examinations without knowledge of the surgical results. They first interpreted the unenhanced studies, indicated whether they felt contrast injection would be helpful in further evaluation, and then (regardless of this determination) read the postcontrast study. RESULTS: On pre- and post-contrast examinations Reader 1 had a sensitivity of 95% (20/21), a specificity of 100% (10/10), and an accuracy of 97% (30/31). Reader 2 had a sensitivity of 95% (20/21), a specificity of 90% (9/10), and an accuracy of 94% (29/31). Reader 3 had a sensitivity of 90% (19/21), a specificity of 100% (10/10), and an accuracy of 94% on the precontrast examinations. His postcontrast performance demonstrated a sensitivity of 86% (18/21), a specificity of 100% (10/10), and an accuracy of 90% (28/31). In the nine interpretations wherein the readers thought that a contrast-enhanced examination might provide useful additional information, they did not change their interpretations in three cases, improved their interpretations in two, and made their interpretations worse in four on the basis of addition of the enhanced images. CONCLUSIONS: Routine use of contrast-enhanced examinations in patients who have had prior lumbar surgery probably adds little diagnostic value and may be confusing.
STUDY DESIGN: Evaluation of magnetic resonance images (MRIs) with surgical reference standard. OBJECTIVES: To determine whether the addition of contrast-enhanced MRI scans increases diagnostic efficacy in the evaluation of recurrent disc herniation. SUMMARY OF BACKGROUND DATA: Many centers now routinely use gadolinium-enhanced examinations in the evaluation of recurrent disc herniation. Others, noting the additional expense of contrast injection, advocate a more limited role for contrast injection and emphasize the importance of T2-weighted axial sequences. METHODS: The study included 165 consecutive patients who were referred to the authors' outpatient imaging center and had a history of previous lumbar discectomy and recurrent back and/or leg pain. The scanning protocol included sagittal and axial T1-weighted spin-echo pre- and postcontrast injection images and sagittal and axial T2-weighted fast spin-echo images. Twenty-eight patients (32 vertebral levels) had subsequent surgical exploration of a disc margin that had previously undergone discectomy. The surgical findings formed the reference standard. Three spine radiologists interpreted the MRI examinations without knowledge of the surgical results. They first interpreted the unenhanced studies, indicated whether they felt contrast injection would be helpful in further evaluation, and then (regardless of this determination) read the postcontrast study. RESULTS: On pre- and post-contrast examinations Reader 1 had a sensitivity of 95% (20/21), a specificity of 100% (10/10), and an accuracy of 97% (30/31). Reader 2 had a sensitivity of 95% (20/21), a specificity of 90% (9/10), and an accuracy of 94% (29/31). Reader 3 had a sensitivity of 90% (19/21), a specificity of 100% (10/10), and an accuracy of 94% on the precontrast examinations. His postcontrast performance demonstrated a sensitivity of 86% (18/21), a specificity of 100% (10/10), and an accuracy of 90% (28/31). In the nine interpretations wherein the readers thought that a contrast-enhanced examination might provide useful additional information, they did not change their interpretations in three cases, improved their interpretations in two, and made their interpretations worse in four on the basis of addition of the enhanced images. CONCLUSIONS: Routine use of contrast-enhanced examinations in patients who have had prior lumbar surgery probably adds little diagnostic value and may be confusing.