From the Departments of Radiology and Biomedical Imaging (D.B.M., S.M.D., J.N., J.F.T.).
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.).
AJNR Am J Neuroradiol. 2019 Apr;40(4):737-744. doi: 10.3174/ajnr.A6020. Epub 2019 Mar 28.
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.
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.
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 ( < .05). Brain and Spinal Cord Injury Center segmentation showed significantly better adaptability to damaged areas compared with PropSeg ( < .001) and DeepSeg ( < .02). The contusion injury volumes based on automated segmentation were significantly associated with motor scores at admission ( = .002) and discharge ( = .009).
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.
我们的目的是使用 2D 卷积神经网络，自动从急性脊髓损伤患者的轴位 T2 加权磁共振成像中分割脊髓和外伤性挫伤损伤。
纳入了 47 例在脊髓损伤后 24 小时内行 3T MRI 检查的患者。我们开发了一种图像分析管道，该管道集成了用于整个脊髓和脊髓内病变分割的 2D 卷积神经网络。线性混合模型用于比较我们的脊髓损伤卷积神经网络（脑与脊髓损伤中心分割）与当前最先进方法之间的测试分割结果。然后，使用线性回归分析来确定分割病变体积与运动评分之间的关联。
与手动标记相比，Brain and Spinal Cord Injury Center 分割模型的平均测试集 Dice 系数在脊髓分割方面为 0.93，在 PropSeg 方面为 0.80，在 DeepSeg 方面为 0.90（均为 Spinal Cord Toolbox 的组成部分）。线性混合模型显示，Brain and Spinal Cord Injury Center 分割与 PropSeg（ <.001）和 DeepSeg（ <.05）之间存在显著差异。与 PropSeg（ <.001）和 DeepSeg（ <.02）相比，Brain and Spinal Cord Injury Center 分割在适应损伤区域方面表现出更好的适应性。基于自动分割的挫伤损伤体积与入院时（ =.002）和出院时（ =.009）的运动评分显著相关。
在急性脊髓损伤患者中，Brain and Spinal Cord Injury Center 脊髓分割与现有的分割工具相比具有优势。基于 Brain and Spinal Cord Injury Center 分割的损伤体积与急性阶段运动障碍的测量值相关。在急性脊髓损伤中进行有针对性的卷积神经网络训练可提高该患者人群的算法性能，并提供脊髓损伤的临床相关指标。
