Imaging diagnosis of lumbar foraminal stenosis in the fifth lumbar nerve root: reliability and reproducibility of T1-weighted three-dimensional lumbar MRI.

Background: Various magnetic resonance imaging (MRI) techniques have been reported in detection of lumbar foraminal stenosis (LFS), especially for T2-weighted three-dimensional MRI (3D-MRI) describing the shape of nerve roots. The detection of LFS in the fifth lumbar nerve root (L5 root), however, is still less reliable compared to other lumbar nerve roots. Then we have been using T1-weighted 3D-MRI aiming to depict the shape of, and also pathology affecting the L5 root. The aim of this study is to evaluate our T1-weighted 3D-MRI in diagnosing LFS of the L5 root.

Methods: This retrospective study included 24 patients with intracanal stenosis (ICS) at L4-5, and 30 patients with LFS at L5-S causing unilateral L5 root lesion. The pre-operative T1-weighted 3D-MRI aiming bilateral L5 nerve roots of each patient were blinded and reviewed twice by five spine surgeons, independently. The image evaluation was performed in two conditions: (I) the symptomatic side was judged in 30 patients of LFS patients, and (II) the symptomatic side or the absence of LFS was judged in images of all the 54 patients including LFS and ICS patients. The correct-answer-rate, sensitivity and specificity of the imaging study were calculated. Also, the intra- and interobserver agreement of the imaging study by five spine surgeons were evaluated by the kappa (κ) statistics.

Results: For conditions (I) and (II) above, the mean correct-answer-rate was 92.3% and 69.8%, respectively. The sensitivity and specificity of the imaging study was 72.6% and 66.3%, respectively. The average of intraobserver κ-value of five examiners was 0.874 and 0.708, and the average of interobserver κ-value was 0.837 and 0.578, respectively.

Conclusions: As well as previously reported T2-weighted 3D-MRI, our T1-weighted 3D-MRI was found to be reliable in diagnosing LFS of the L5 root.