D B McCoy1,2, S M Dupont1, C Gros3, J Cohen-Adad3, R J Huie4,2, A Ferguson4,2, X Duong-Fernandez4,2, L H Thomas4,2, V Singh5, J Narvid1, L Pascual6, N Kyritsis4,2, M S Beattie4,2, J C Bresnahan4,2, S Dhall4,2, W Whetstone4,2, J F Talbott7,4. 1. From the Departments of Radiology and Biomedical Imaging (D.B.M., S.M.D., J.N., J.F.T.). 2. Brain and Spinal Injury Center (D.B.M., R.J.H., A.F., X.D.-F., L.H.T., N.K., M.S.B., J.C.B., S.D., W.W.). 3. NeuroPoly Lab (C.G., J.C.-A.), Institute of Biomedical Engineering, Polytechnique Montreal, Montreal, Quebec, Canada. 4. Neurological Surgery (R.J.H., A.F., X.D.-F., L.H.T., N.K., M.S.B., J.C.B., S.D., W.W., J.F.T.). 5. Departments of Neurology (V.S.). 6. Orthopedic Surgery (L.P.), Zuckerberg San Francisco General Hospital, University of California, San Francisco, San Francisco, California. 7. From the Departments of Radiology and Biomedical Imaging (D.B.M., S.M.D., J.N., J.F.T.) jason.talbott@ucsf.edu.
Abstract
BACKGROUND AND PURPOSE: Our aim was to use 2D convolutional neural networks for automatic segmentation of the spinal cord and traumatic contusion injury from axial T2-weighted MR imaging in a cohort of patients with acute spinal cord injury. MATERIALS AND METHODS: Forty-seven patients who underwent 3T MR imaging within 24 hours of spinal cord injury were included. We developed an image-analysis pipeline integrating 2D convolutional neural networks for whole spinal cord and intramedullary spinal cord lesion segmentation. Linear mixed modeling was used to compare test segmentation results between our spinal cord injury convolutional neural network (Brain and Spinal Cord Injury Center segmentation) and current state-of-the-art methods. Volumes of segmented lesions were then used in a linear regression analysis to determine associations with motor scores. RESULTS: Compared with manual labeling, the average test set Dice coefficient for the Brain and Spinal Cord Injury Center segmentation model was 0.93 for spinal cord segmentation versus 0.80 for PropSeg and 0.90 for DeepSeg (both components of the Spinal Cord Toolbox). Linear mixed modeling showed a significant difference between Brain and Spinal Cord Injury Center segmentation compared with PropSeg (P < .001) and DeepSeg (P < .05). Brain and Spinal Cord Injury Center segmentation showed significantly better adaptability to damaged areas compared with PropSeg (P < .001) and DeepSeg (P < .02). The contusion injury volumes based on automated segmentation were significantly associated with motor scores at admission (P = .002) and discharge (P = .009). CONCLUSIONS: Brain and Spinal Cord Injury Center segmentation of the spinal cord compares favorably with available segmentation tools in a population with acute spinal cord injury. Volumes of injury derived from automated lesion segmentation with Brain and Spinal Cord Injury Center segmentation correlate with measures of motor impairment in the acute phase. Targeted convolutional neural network training in acute spinal cord injury enhances algorithm performance for this patient population and provides clinically relevant metrics of cord injury.
BACKGROUND AND PURPOSE: Our aim was to use 2D convolutional neural networks for automatic segmentation of the spinal cord and traumatic contusion injury from axial T2-weighted MR imaging in a cohort of patients with acute spinal cord injury. MATERIALS AND METHODS: Forty-seven patients who underwent 3T MR imaging within 24 hours of spinal cord injury were included. We developed an image-analysis pipeline integrating 2D convolutional neural networks for whole spinal cord and intramedullary spinal cord lesion segmentation. Linear mixed modeling was used to compare test segmentation results between our spinal cord injury convolutional neural network (Brain and Spinal Cord Injury Center segmentation) and current state-of-the-art methods. Volumes of segmented lesions were then used in a linear regression analysis to determine associations with motor scores. RESULTS: Compared with manual labeling, the average test set Dice coefficient for the Brain and Spinal Cord Injury Center segmentation model was 0.93 for spinal cord segmentation versus 0.80 for PropSeg and 0.90 for DeepSeg (both components of the Spinal Cord Toolbox). Linear mixed modeling showed a significant difference between Brain and Spinal Cord Injury Center segmentation compared with PropSeg (P < .001) and DeepSeg (P < .05). Brain and Spinal Cord Injury Center segmentation showed significantly better adaptability to damaged areas compared with PropSeg (P < .001) and DeepSeg (P < .02). The contusion injury volumes based on automated segmentation were significantly associated with motor scores at admission (P = .002) and discharge (P = .009). CONCLUSIONS: Brain and Spinal Cord Injury Center segmentation of the spinal cord compares favorably with available segmentation tools in a population with acute spinal cord injury. Volumes of injury derived from automated lesion segmentation with Brain and Spinal Cord Injury Center segmentation correlate with measures of motor impairment in the acute phase. Targeted convolutional neural network training in acute spinal cord injury enhances algorithm performance for this patient population and provides clinically relevant metrics of cord injury.
Authors: Christopher S Ahuja; Jefferson R Wilson; Satoshi Nori; Mark R N Kotter; Claudia Druschel; Armin Curt; Michael G Fehlings Journal: Nat Rev Dis Primers Date: 2017-04-27 Impact factor: 52.329
Authors: J Haefeli; M C Mabray; W D Whetstone; S S Dhall; J Z Pan; P Upadhyayula; G T Manley; J C Bresnahan; M S Beattie; A R Ferguson; J F Talbott Journal: AJNR Am J Neuroradiol Date: 2016-12-22 Impact factor: 3.825
Authors: Allan R Martin; Izabela Aleksanderek; Julien Cohen-Adad; Zenovia Tarmohamed; Lindsay Tetreault; Nathaniel Smith; David W Cadotte; Adrian Crawley; Howard Ginsberg; David J Mikulis; Michael G Fehlings Journal: Neuroimage Clin Date: 2015-12-04 Impact factor: 4.881
Authors: Mark Jenkinson; Christian F Beckmann; Timothy E J Behrens; Mark W Woolrich; Stephen M Smith Journal: Neuroimage Date: 2011-09-16 Impact factor: 6.556
Authors: Nikos Kyritsis; Abel Torres-Espín; Patrick G Schupp; J Russell Huie; Austin Chou; Xuan Duong-Fernandez; Leigh H Thomas; Rachel E Tsolinas; Debra D Hemmerle; Lisa U Pascual; Vineeta Singh; Jonathan Z Pan; Jason F Talbott; William D Whetstone; John F Burke; Anthony M DiGiorgio; Philip R Weinstein; Geoffrey T Manley; Sanjay S Dhall; Adam R Ferguson; Michael C Oldham; Jacqueline C Bresnahan; Michael S Beattie Journal: J Exp Med Date: 2021-03-01 Impact factor: 14.307
Authors: Austin Chou; Abel Torres-Espin; Nikos Kyritsis; J Russell Huie; Sarah Khatry; Jeremy Funk; Jennifer Hay; Andrew Lofgreen; Rajiv Shah; Chandler McCann; Lisa U Pascual; Edilberto Amorim; Philip R Weinstein; Geoffrey T Manley; Sanjay S Dhall; Jonathan Z Pan; Jacqueline C Bresnahan; Michael S Beattie; William D Whetstone; Adam R Ferguson Journal: PLoS One Date: 2022-04-07 Impact factor: 3.240
Authors: Jiamin Zhou; Pablo F Damasceno; Ravi Chachad; Justin R Cheung; Alexander Ballatori; Jeffrey C Lotz; Ann A Lazar; Thomas M Link; Aaron J Fields; Roland Krug Journal: Front Endocrinol (Lausanne) Date: 2020-09-02 Impact factor: 6.055