Polzer Constanze, Yilmaz Eren, Meyer Carsten, Jang Hyungseok, Jansen Olav, Lorenz Cristian, Bürger Christian, Glüer Claus-Christian, Sedaghat Sam
Department of Radiology and Neuroradiology, University Hospital Schleswig-Holstein, Campus Kiel, Kiel, Germany.
Section Biomedical Imaging, Department of Radiology and Neuroradiology, University Hospital Schleswig-Holstein, Campus Kiel, Kiel, Germany; Department of Computer Science, Ostfalia University of Applied Sciences, Wolfenbüttel, Germany.
Eur J Radiol. 2024 Apr;173:111364. doi: 10.1016/j.ejrad.2024.111364. Epub 2024 Feb 13.
We developed and tested a neural network for automated detection and stability analysis of vertebral body fractures on computed tomography (CT).
257 patients who underwent CT were included in this Institutional Review Board (IRB) approved study. 463 fractured and 1883 non-fractured vertebral bodies were included, with 190 fractures unstable. Two readers identified vertebral body fractures and assessed their stability. A combination of a Hierarchical Convolutional Neural Network (hNet) and a fracture Classification Network (fNet) was used to build a neural network for the automated detection and stability analysis of vertebral body fractures on CT. Two final test settings were chosen: one with vertebral body levels C1/2 included and one where they were excluded.
The mean age of the patients was 68 ± 14 years. 140 patients were female. The network showed a slightly higher diagnostic performance when excluding C1/2. Accordingly, the network was able to distinguish fractured and non-fractured vertebral bodies with a sensitivity of 75.8 % and a specificity of 80.3 %. Additionally, the network determined the stability of the vertebral bodies with a sensitivity of 88.4 % and a specificity of 80.3 %. The AUC was 87 % and 91 % for fracture detection and stability analysis, respectively. The sensitivity of our network in indicating the presence of at least one fracture / one unstable fracture within the whole spine achieved values of 78.7 % and 97.2 %, respectively, when excluding C1/2.
The developed neural network can automatically detect vertebral body fractures and evaluate their stability concurrently with a high diagnostic performance.
我们开发并测试了一种用于在计算机断层扫描（CT）上自动检测和分析椎体骨折稳定性的神经网络。
本研究经机构审查委员会（IRB）批准，纳入了257例行CT检查的患者。共纳入463个骨折椎体和1883个未骨折椎体，其中190处骨折为不稳定骨折。两名阅片者识别椎体骨折并评估其稳定性。使用分层卷积神经网络（hNet）和骨折分类网络（fNet）相结合的方法构建了一个用于在CT上自动检测和分析椎体骨折稳定性的神经网络。选择了两种最终测试设置：一种包含C1/2椎体水平，另一种排除C1/2椎体水平。
患者的平均年龄为68±14岁。140例为女性。排除C1/2椎体水平时，该网络的诊断性能略高。相应地，该网络能够区分骨折和未骨折的椎体，灵敏度为75.8%，特异度为80.3%。此外，该网络判断椎体稳定性的灵敏度为88.4%，特异度为80.3%。骨折检测和稳定性分析的AUC分别为87%和91%。排除C1/2椎体水平时，我们的网络在指示整个脊柱中至少存在一处骨折/一处不稳定骨折方面的灵敏度分别达到78.7%和97.2%。
所开发的神经网络能够自动检测椎体骨折并同时评估其稳定性，具有较高的诊断性能。
