75 The Conus Medullaris– Vertebral Column Ratio: A New Way to Identify Tethered Cord on MRI
Andrea Bischoff1, Alberto Peña1, Jill Ketzer1, Kristen Campbell2, Brent O'Neill3, Nicholas Stence4, David Mirsky4
1International Center for Colorectal and Urogenital Care, Department of Pediatric Surgery, Children's Hospital Colorado, Aurora, USA. 2Children's Hospital Colorado, Aurora, USA. 3Department of Neurosurgery, Children's Hospital Colorado, Aurora, USA. 4Department of Radiology, Children's Hospital Colorado, Aurora, USA


Purpose: Approximately 25% of patients with anorectal malformation have tethered cord. The traditional way of determining conus medullaris level on MRI relies on either using lumbar spine images and “counting up” or acquiring a whole-spine sequence and “counting down” from C1. Either way, the task may be challenging in patients with vertebral numeric variation or transitional, lumbarized, or sacralized vertebrae. Following the sacral ratio model to predict bowel control in patients with anorectal malformation, our purpose was to use reliable anatomic landmarks (foramen magnum, conus termination, and the upper limit of the iliac crest), to differentiate patients with normal and low lying conus.  
Methods: A single institution database search identified 2 groups of patients: 256 with normal and 85 with abnormal spinal MRI, who underwent tethered cord release. The conus medullaris - vertebral column ratio was calculated in both groups. The ratio was obtained by dividing the distance between the conus level and the iliac crest by the distance between the foramen magnum and the conus level (figure). IRB approval was obtained (# 16-2330).
Results: The mean ratio was 0.1843 (sd 0.033) for the non-tethered group and 0.1179 (sd 0.093) for the tethered one, with a p-value <0.0001. By ROC curve analysis, the ratio proved to be a good discriminator between normal and abnormal patients, with Area under the Curve (AUC) equal to 0.749, meaning that at random there is a 75% chance that the tethered cord patient will have a lower ratio than the non-tethered cord patient.  
Conclusion: “The Conus Medullaris – Vertebral Column Ratio” is a good predictor of low lying conus level on MRI and offer an easy alternative to avoid the need for vertebral counting. The next step will consist in evaluating the clinical symptoms to select a cut-off in which the patient should be referred for neurosurgical evaluation.