Purpose: Anatomical landmarks and their relation to the lumbar vertebrae are well described in subjects with normal spine anatomy, but not for subjects with lumbosacral transitional vertebra (LSTV), in whom correct numbering of the vertebrae is challenging and can lead to wrong-level treatment. The aim of this study was to quantify the value of different anatomical landmarks for correct identification of the lumbar vertebra level in subjects with LSTV.
Methods: After IRB approval, 71 subjects (57 ± 17 years) with and 62 without LSTV (57 ± 17 years), all with imaging studies that allowed correct numbering of the lumbar vertebrae by counting down from C2 (n = 118) or T1 (n = 15) were included. Commonly used anatomical landmarks (ribs, aortic bifurcation (AB), right renal artery (RRA) and iliac crest height) were documented to determine the ability to correctly number the lumbar vertebrae. Further, a tangent to the top of the iliac crests was drawn on coronal MRI images by two blinded, independent readers and named the 'iliac crest tangent sign'. The sensitivity, specificity and the interreader agreement were calculated.
Results: While the level of the AB and the RRA were found to be unreliable in correct numbering of the lumbar vertebrae in LSTV subjects, the iliac crest tangent sign had a sensitivity and specificity of 81 % and 64-88 %, respectively, with an interreader agreement of k = 0.75.
Conclusion: While anatomical landmarks are not always reliable, the 'iliac crest tangent sign' can be used without advanced knowledge in MRI to most accurately number the vertebrae in subjects with LSTV, if only a lumbar spine MRI is available.