Sekuboyina Anjany, Rempfler Markus, Valentinitsch Alexander, Menze Bjoern H, Kirschke Jan S
Department of Informatics (A.S., B.H.M.) and Department of Neuroradiology, School of Medicine (A.S., A.V., J.S.K.), Technical University of Munich; Department of Diagnostic and Interventional Neuroradiology, Klinikum Rechts der Isar, Ismaninger Str 22, 81675 Munich, Germany (A.S.); and Friedrich Miescher Institute for Biomedical Engineering, Basel, Switzerland (M.R.).
Radiol Artif Intell. 2020 Mar 25;2(2):e190074. doi: 10.1148/ryai.2020190074. eCollection 2020 Mar.
To use and test a labeling algorithm that operates on two-dimensional reformations, rather than three-dimensional data to locate and identify vertebrae.
The authors improved the Btrfly Net, a fully convolutional network architecture described by Sekuboyina et al, which works on sagittal and coronal maximum intensity projections (MIPs) and augmented it with two additional components: spine localization and adversarial a priori learning. Furthermore, two variants of adversarial training schemes that incorporated the anatomic a priori knowledge into the Btrfly Net were explored. The superiority of the proposed approach for labeling vertebrae on three datasets was investigated: a public benchmarking dataset of 302 CT scans and two in-house datasets with a total of 238 CT scans. The Wilcoxon signed rank test was employed to compute the statistical significance of the improvement in performance observed with various architectural components in the authors' approach.
On the public dataset, the authors' approach using the described Btrfly Net with energy-based prior encoding (Btrfly) network performed as well as current state-of-the-art methods, achieving a statistically significant ( < .001) vertebrae identification rate of 88.5% ± 0.2 (standard deviation) and localization distances of less than 7 mm. On the in-house datasets that had a higher interscan data variability, an identification rate of 85.1% ± 1.2 was obtained.
An identification performance comparable to existing three-dimensional approaches was achieved when labeling vertebrae on two-dimensional MIPs. The performance was further improved using the proposed adversarial training regimen that effectively enforced local spine a priori knowledge during training. Spine localization increased the generalizability of our approach by homogenizing the content in the MIPs.© RSNA, 2020.
使用并测试一种在二维重组图像上运行而非三维数据上运行的标记算法，以定位和识别椎骨。
作者改进了由Sekuboyina等人描述的全卷积网络架构——蝴蝶网（Btrfly Net），该网络在矢状面和冠状面最大强度投影（MIP）上运行，并通过两个附加组件对其进行增强：脊柱定位和对抗性先验学习。此外，还探索了两种将解剖学先验知识纳入蝴蝶网的对抗训练方案变体。研究了所提出的方法在三个数据集上标记椎骨的优越性：一个包含302例CT扫描的公共基准数据集和两个共有238例CT扫描的内部数据集。采用Wilcoxon符号秩检验来计算作者方法中各种架构组件在性能提升方面的统计显著性。
在公共数据集上，作者使用所描述的带有基于能量的先验编码（Btrfly）网络的蝴蝶网方法与当前的先进方法表现相当，实现了具有统计学显著性（<0.001）的88.5%±0.2（标准差）的椎骨识别率和小于7毫米的定位距离。在具有更高扫描间数据变异性的内部数据集上，获得了85.1%±1.2的识别率。
在二维MIP上标记椎骨时，实现了与现有三维方法相当的识别性能。使用所提出的对抗训练方案进一步提高了性能，该方案在训练期间有效地强化了局部脊柱先验知识。脊柱定位通过使MIP中的内容均匀化提高了我们方法的通用性。©RSNA，2020年。
