Division of Neurocritical Care, Department of Neurology, University of North Carolina at Chapel Hill, Chapel Hill, NC, USA.
Division of Brain Injury Outcomes, Johns Hopkins University, Baltimore, MD, USA.
Neuroinformatics. 2021 Jul;19(3):403-415. doi: 10.1007/s12021-020-09493-5. Epub 2020 Sep 27.
Intracranial hemorrhage (ICH) occurs when a blood vessel ruptures in the brain. This leads to significant morbidity and mortality, the likelihood of which is predicated on the size of the bleeding event. X-ray computed tomography (CT) scans allow clinicians and researchers to qualitatively and quantitatively diagnose hemorrhagic stroke, guide interventions and determine inclusion criteria of patients in clinical trials. There is no currently available open source, validated tool to quickly segment hemorrhage. Using an automated pipeline and 2D and 3D deep neural networks, we show that we can quickly and accurately estimate ICH volume with high agreement with time-consuming manual segmentation. The training and validation datasets include significant heterogeneity in terms of pathology, such as the presence of intraventricular (IVH) or subdural hemorrhages (SDH) as well as variable image acquisition parameters. We show that deep neural networks trained with an appropriate anatomic context in the network receptive field, can effectively perform ICH segmentation, but those without enough context will overestimate hemorrhage along the skull and around calcifications in the ventricular system. We trained with all data from a multi-center phase II study (n = 112) achieving a best mean and median Dice coefficient of 0.914 and 0.919, a volume correlation of 0.979 and an average volume difference of 1.7 ml and root mean squared error of 4.7 ml in 500 out-of-sample scans from the corresponding multi-center phase III study. 3D networks with appropriate anatomic context outperformed both 2D and random forest models. Our results suggest that deep neural network models, when carefully developed can be incorporated into the workflow of an ICH clinical trial series to quickly and accurately segment ICH, estimate total hemorrhage volume and minimize segmentation failures. The model, weights and scripts for deployment are located at https://github.com/msharrock/deepbleed . This is the first publicly available neural network model for segmentation of ICH, the only model evaluated with the presence of both IVH and SDH and the only model validated in the workflow of a series of clinical trials.
颅内出血 (ICH) 是指血管在大脑中破裂。这会导致严重的发病率和死亡率，其可能性取决于出血事件的大小。X 射线计算机断层扫描 (CT) 扫描使临床医生和研究人员能够定性和定量诊断出血性中风，指导干预措施，并确定临床试验患者的纳入标准。目前没有可用的开源、经过验证的工具可以快速分割出血。我们使用自动化流水线和 2D 和 3D 深度神经网络，展示了我们可以快速准确地估计 ICH 体积，并且与耗时的手动分割具有高度一致性。训练和验证数据集在病理学方面存在很大的异质性，例如存在脑室 (IVH) 或硬膜下血肿 (SDH) 以及可变的图像采集参数。我们表明，在网络感受野中具有适当解剖上下文的深度神经网络可以有效地执行 ICH 分割，但那些没有足够上下文的网络将高估颅骨周围和脑室系统中的钙化周围的出血。我们使用来自多中心 II 期研究 (n=112) 的所有数据进行训练，在来自相应多中心 III 期研究的 500 个样本外扫描中，实现了最佳平均和中位数 Dice 系数分别为 0.914 和 0.919，体积相关性为 0.979，平均体积差异为 1.7ml，均方根误差为 4.7ml。具有适当解剖上下文的 3D 网络优于 2D 和随机森林模型。我们的结果表明，当深度神经网络模型经过精心开发后，可以纳入 ICH 临床试验系列的工作流程中，以快速准确地分割 ICH，估计总出血体积并最大限度地减少分割失败。模型、权重和部署脚本位于 https://github.com/msharrock/deepbleed。这是第一个可用于分割 ICH 的公开可用的神经网络模型，是唯一一个评估存在 IVH 和 SDH 的模型，也是唯一一个在一系列临床试验工作流程中验证的模型。
