Improved Magnetic Resonance Myelography of the Cervical Spine Using Image Fusion and Volumetry

Abstract

Discrepancies between symptoms and imaging findings may be significant in spinal diseases. The relative advantages and disadvantages of postmyelographic CT (pmCT) and magnetic resonance myelography (MRM) in measuring cervical cerebrospinal fluid dimensions and volumes are poorly understood. This study investigated the usefulness of an improved MRM approach and compared its findings of dural areas and volumes of normal and pathologic cervical spine levels with those obtained by pmCT. Sixty patients with suspected cervical spinal stenosis underwent myelography, pmCT, and an improved MRM approach using a dedicated heavily T2-weighted magnetic resonance (MR) sequence in combination with image fusion. Cross-sectional areas and volumes of the dural sac from C2-3 to C7-T1 were compared for significance of differences between mean values. Pearson correlations between pmCT and MRM values were analyzed. Then 99% confidence intervals for normal levels and levels with obvious stenosis were calculated. For normal cervical levels we found no significant differences for both dural areas and volumes between pmCT and MRM. For cervical spine levels C3-4 to C6-7 with pathologic alterations, we found significant larger areas and volumes on MRM when compared with pmCT but not for the levels C2-3 and C7-T1. Both dural area and volume of normal and pathologic levels showed strong correlation between pmCT and MRM. We found no overlap between 99% confidence intervals for normal levels and stenotic levels. The optimized MRM approach (i.e., a dedicated MRM sequence in combination with image fusion) and pmCT demonstrated excellent quantitative volumetric agreement. Image fusion with conventional T2 may be used to aid in diagnosis.