Can the supine extended position with legs straightened and a lumbar pillow complement the psoas-relaxed position for the diagnosis of lumbar spinal stenosis on MRI? A prospective cross-sectional study

When the diagnosis of lumbar spinal stenosis is uncertain in the conventional psoas-relaxed position (PRP) MRI, upright MRI is a not widely available alternative. The aim of this study was to investigate whether a modified supine MRI position with legs straight and an adjustable pillow under the lumbar spine (EXT) could mimic the effect of standing. We hypothesized that EXT would mimic upright MRI in stenosis severity. Forty patients (58 % male, mean age 57 years) underwent a 3T (65 %) or 1.5T MRI. For the levels L2-S1, central stenosis (Lee: 0-3), dural cross sectional area (DCSA: mm2); lateral recess stenosis (Bartinsky: 0-3), foraminal stenosis (Lee: 0-3) and lordosis angle (LA: L1-S1°) were measured in PRP and EXT by three radiologists. Cohen's kappa and Cronbach's alpha for intra/interrater reliability and Pearson's correlation coefficient between PRP and EXT were measured. P < 0.05 was used for statistical significance. At L4/L5 EXT showed increased narrowing in 40 % (central), 37.5 % (lateral), and 17.5 % (foraminal). DCSA decreased (P < 0.0001) and LA increased (P < 0.0001), with strong correlations and high reliability (P < 0.0001). Lateral recess stenosis increased significantly in EXT at L4/L5 (P = 0.014; K = 0.646). No significant difference was observed in foraminal stenosis (K = 0.488). EXT increased lumbar lordosis to the normal standing range of approximately 50°, mimicking an upright MRI. A more realistic degree of central stenosis at all levels and of lateral recess stenosis at L4/L5 was shown. Foraminal stenosis did not increase.