Zhou Tianshu, Tan Tao, Pan Xiaoyan, Tang Hui, Li Jingsong
Engineering Research Center of EMR and Intelligent Expert System, Ministry of Education, Key Laboratory for Biomedical Engineering of Ministry of Education, College of Biomedical Engineering and Instrument Science, Zhejiang University, Hangzhou, China.
Department of Mathematics and Computer Science, Eindhoven University of Technology and Radiology, Netherlands Cancer Institute, Amsterdam, the Netherlands.
Quant Imaging Med Surg. 2021 Jan;11(1):67-83. doi: 10.21037/qims-20-286.
The objectives of this study were to develop a 3D convolutional deep learning framework (CarotidNet) for fully automatic segmentation of carotid bifurcations in computed tomography angiography (CTA) images and to facilitate the quantification of carotid stenosis and risk assessment of stroke.
Our pipeline was a two-stage cascade network that included a localization phase and a segmentation phase. The network framework was based on the 3D version of U-Net, but was refined in three ways: (I) by adding residual connections and a deep supervision strategy to cope with the vanishing problem in back-propagation; (II) by adopting dilated convolution in order to strengthen the capacity to capture contextual information; and (III) by establishing a hybrid objective function to address the extreme imbalance between foreground and background voxels.
We trained our networks on 15 cases and evaluated their performance based on 41 cases from the MICCAI Challenge 2009 dataset. A Dice similarity coefficient of 82.3% was achieved for the test cases.
We developed a carotid segmentation method based on U-Net that can segment tiny carotid bifurcation lumens from very large backgrounds with no manual intervention. This was the first attempt to use deep learning to achieve carotid bifurcation segmentation in 3D CTA images. Our results indicate that deep learning is a promising method for automatically extracting carotid bifurcation lumens.
本研究的目的是开发一种三维卷积深度学习框架（颈动脉网络），用于在计算机断层血管造影（CTA）图像中全自动分割颈动脉分叉，并促进颈动脉狭窄的量化和中风风险评估。
我们的流程是一个两阶段的级联网络，包括定位阶段和分割阶段。网络框架基于三维版的U-Net，但在三个方面进行了改进：（I）通过添加残差连接和深度监督策略来应对反向传播中的梯度消失问题；（II）采用空洞卷积以增强捕获上下文信息的能力；（III）通过建立混合目标函数来解决前景体素和背景体素之间的极端不平衡问题。
我们在15个病例上训练了我们的网络，并基于2009年医学图像计算方法与计算机辅助干预国际会议（MICCAI）挑战赛数据集中的41个病例评估了它们的性能。测试病例的骰子相似系数达到了82.3%。
我们开发了一种基于U-Net的颈动脉分割方法，该方法可以在无需人工干预的情况下，从非常大的背景中分割出微小的颈动脉分叉管腔。这是首次尝试使用深度学习在三维CTA图像中实现颈动脉分叉分割。我们的结果表明，深度学习是一种自动提取颈动脉分叉管腔的有前途的方法。
