Atlas-based Volumetric Analysis of Blunt Spinal Cord Contusion Injury

• Published in: Neurosurgery Vol. 67; no. Supplement_1
• Main Authors: Burke, John F, Mummaneni, Nikhil, Ferguson, Adam, Huie, J. Russell, Thomas, Leigh H, Pascual, Lisa, DiGiorgio, Anthony M, Hemmerle, Debra P, Singh, Vineeta, Torres, Abel, Kyritsis, Nikolaos, Weinstein, Philip R, Whetstone, William, Dhall, Sanjay S, Fernandez, Xuan Duong, Beattie, Michael, Bresnahan, Jacqueline, Talbott, Jason
• Format: Journal Article
• Language: English
• Published: Philadelphia Wolters Kluwer Health, Inc 01-12-2020
• Subjects: Neurosurgery; Spinal cord; Spine; Surgical techniques; Volumetric analysis
• ISSN: 0148-396X; 1524-4040
• DOI: 10.1093/neuros/nyaa447_416

INTRODUCTION Measures of intrinsic spinal cord (SC) signal abnormality on T2-weighted (T2w) MRI have been shown to correlate with injury severity and outcome after acute spinal cord injury (SCI). However, current MR evaluation of intrinsic cord signal abnormality is limited to gross morphologic feature description. Recent advances in SC analysis have led to the development of a robust anatomic atlas incorporated into an open-source platform referred to as the Spinal Cord Toolbox (SCT) that allow for quantitiative volumetric injury analysis. METHODS This IRB-approved prospective cohort study involved the image analysis of 72 blunt cervical SCI patients enrolled in the TRACK-SCI clinical research protocol. Axial T2w MRI data obtained within 24 hours of injury (TR = 3800ms, TE = 102ms, section thickness = 3.3-4.4mm) were processed using the SCT. Briefly, (i) SC was automatically segmented using Deepseg tool in the SCT and manually corrected by a neuro-radiologist, (ii) the PAM50 template was then registered to the T2w data, (iii) the template white matter (WM) and gray matter (GM) maps were used to identify anatomical subdomains in the subject space, and (iv) a lesion mask was manually created and used to calculate the probabilistic volume of lesion in total and in anatomic subdomains of the spinal cord using the extract-metric function in the SCT. Lesion volume data were used as predictor variables for correlation with outcome measures in including ASIA grade and motor scores. RESULTS Volumetric MRI measures of T2-hyperintense lesion accurately predict chronic total motor scores and ASIA grade. Similarly, MRI measures of injury volume signficantly correlated with motor scores in the chronic stage (P < .05). Ventral horn lesion volume was identified as an important predictor of acute and chronic total motor scores. CONCLUSION Atlas-based volumetric assessment of T2-abnormality in the spinal cord from MRI performed acutely after injury can predict neurologic outcome. The volume of total injury and of injured spinal cord motor regions on T2w MRI is significantly and independently associated with neurologic outcome at chronic time-points after injury.