Shan Wei, Duan Yunyun, Zheng Yu, Wu Zhenzhou, Chan Shang Wei, Wang Qun, Gao Peiyi, Liu Yaou, He Kunlun, Wang Yongjun
Department of Neurology, Beijing Tiantan Hospital, Capital Medical University, Beijing, China.
National Center for Clinical Medicine of Neurological Diseases, Beijing, China.
Front Med (Lausanne). 2021 Nov 25;8:681183. doi: 10.3389/fmed.2021.681183. eCollection 2021.
Reliable quantification of white matter hyperintensities (WHMs) resulting from cerebral small vessel diseases (CSVD) is essential for understanding their clinical impact. We aim to develop and clinically validate a deep learning system for automatic segmentation of CSVD-WMH from fluid-attenuated inversion recovery (FLAIR) imaging using large multicenter data. A FLAIR imaging dataset of 1,156 patients diagnosed with CSVD associated WMH (median age, 54 years; 653 males) obtained between September 2018 and September 2019 from Beijing Tiantan Hospital was retrospectively analyzed in this study. Locations of CSVD-WMH on the FLAIR scans were manually marked by two experienced neurologists. Using the manually labeled data of 996 patients (development set), a U-shaped novel 2D convolutional neural network (CNN) architecture was trained for automatic segmentation of CSVD-WMH. The segmentation performance of the network was evaluated with per pixel and lesion level dice scores using an independent internal test set ( = 160) and a multi-center external test set ( = 90, three medical centers). The clinical suitability of the segmentation results, classified as acceptable, acceptable with minor revision, acceptable with major revision, and not acceptable, was analyzed by three independent neuroradiologists. The inter-neuroradiologists agreement rate was assessed by the Kendall-W test. On the internal and external test sets, the proposed CNN architecture achieved per pixel and lesion level dice scores of 0.72 (external test set), and they were significantly better than the state-of-the-art deep learning architectures proposed for WMH segmentation. In the clinical evaluation, neuroradiologists observed the segmentation results for 95% of the patients were acceptable or acceptable with a minor revision. A deep learning system can be used for automated, objective, and clinically meaningful segmentation of CSVD-WMH with high accuracy.
对脑小血管疾病（CSVD）导致的脑白质高信号（WMH）进行可靠量化对于理解其临床影响至关重要。我们旨在开发并在临床上验证一种深度学习系统，该系统利用大型多中心数据从液体衰减反转恢复（FLAIR）成像中自动分割CSVD-WMH。本研究回顾性分析了2018年9月至2019年9月期间从北京天坛医院获取的1156例被诊断为与CSVD相关的WMH患者（中位年龄54岁；男性653例）的FLAIR成像数据集。两名经验丰富的神经科医生手动标记了FLAIR扫描上CSVD-WMH的位置。利用996例患者的手动标注数据（开发集），训练了一种U形新型二维卷积神经网络（CNN）架构用于CSVD-WMH的自动分割。使用独立的内部测试集（=160）和多中心外部测试集（=90，三个医学中心），通过每像素和病灶水平的骰子分数评估网络的分割性能。由三名独立的神经放射科医生分析分割结果的临床适用性，分为可接受、轻微修订后可接受、重大修订后可接受和不可接受。通过肯德尔-W检验评估神经放射科医生之间的一致性率。在内部和外部测试集上，所提出的CNN架构实现了每像素和病灶水平的骰子分数为0.72（外部测试集），并且显著优于为WMH分割提出的最先进的深度学习架构。在临床评估中，神经放射科医生观察到95%患者的分割结果是可接受的或轻微修订后可接受的。深度学习系统可用于对CSVD-WMH进行自动化、客观且具有临床意义的高精度分割。
