Disc space-targeted angled axial MR images of the lumbar spine: a potential source of diagnostic error

Abstract

To compare the sensitivities of contiguous axial (CA) images and disc space-targeted angled axial (DSTAA) images of the lumbar spine for: (1) the detection of spondylolysis defects, and (2) the identification of disc material which has migrated away from the parent disc. Prospective imaging of the lumbar spine was performed over a 22-month period in 103 patients. Imaging protocols included spin-echo T1- and fast spin-echo (FSE) T2-weighted images in the sagittal and axial planes. For each patient, axial images were obtained both contiguously throughout the lumbar spine and as angled images, targeted at the region of the disc space only. Two separate data sets were compiled: one that included contiguous axial images (CA data set) and another that included targeted angled images through the disc spaces only (DSTAA data set). Identical sagittal images were included with both sets. A single radiologist independently interpreted the two data sets for a given patient following an intervening time lapse. The radiologist was blinded to the initial interpretation. Results from the two independent interpretations were then compared. Spondylolysis defects were identified at 15 different levels in 14 patients (14%) using the contiguous axial imaging protocol compared with 8 (7%) identified using the disc space targeted angled axial imaging protocol (P = 0.12). Migrated disc material was identified in 12 patients (12%) using the CA protocol compared with 3 patients (3%) identified with the DSTAA protocol (P = 0.016). MR imaging of the lumbar spine using contiguous axial data obtained through both the level of the disc and vertebral body demonstrates migrated disc material and spondylolysis defects better than did disc space-targeted angled data.