Department of Radiology and Biomedical Imaging, Yale University School of Medicine, New Haven, Connecticut, United States of America.
Charité Center for Diagnostic and Interventional Radiology, Charité-Universitätsmedizin Berlin, Berlin, Germany.
PLoS One. 2021 Dec 1;16(12):e0260630. doi: 10.1371/journal.pone.0260630. eCollection 2021.
Accurate liver segmentation is key for volumetry assessment to guide treatment decisions. Moreover, it is an important pre-processing step for cancer detection algorithms. Liver segmentation can be especially challenging in patients with cancer-related tissue changes and shape deformation. The aim of this study was to assess the ability of state-of-the-art deep learning 3D liver segmentation algorithms to generalize across all different Barcelona Clinic Liver Cancer (BCLC) liver cancer stages.
This retrospective study, included patients from an institutional database that had arterial-phase T1-weighted magnetic resonance images with corresponding manual liver segmentations. The data was split into 70/15/15% for training/validation/testing each proportionally equal across BCLC stages. Two 3D convolutional neural networks were trained using identical U-net-derived architectures with equal sized training datasets: one spanning all BCLC stages ("All-Stage-Net": AS-Net), and one limited to early and intermediate BCLC stages ("Early-Intermediate-Stage-Net": EIS-Net). Segmentation accuracy was evaluated by the Dice Similarity Coefficient (DSC) on a dataset spanning all BCLC stages and a Wilcoxon signed-rank test was used for pairwise comparisons.
219 subjects met the inclusion criteria (170 males, 49 females, 62.8±9.1 years) from all BCLC stages. Both networks were trained using 129 subjects: AS-Net training comprised 19, 74, 18, 8, and 10 BCLC 0, A, B, C, and D patients, respectively; EIS-Net training comprised 21, 86, and 22 BCLC 0, A, and B patients, respectively. DSCs (mean±SD) were 0.954±0.018 and 0.946±0.032 for AS-Net and EIS-Net (p<0.001), respectively. The AS-Net 0.956±0.014 significantly outperformed the EIS-Net 0.941±0.038 on advanced BCLC stages (p<0.001) and yielded similarly good segmentation performance on early and intermediate stages (AS-Net: 0.952±0.021; EIS-Net: 0.949±0.027; p = 0.107).
To ensure robust segmentation performance across cancer stages that is independent of liver shape deformation and tumor burden, it is critical to train deep learning models on heterogeneous imaging data spanning all BCLC stages.
准确的肝脏分割是进行体积评估以指导治疗决策的关键。此外，它还是癌症检测算法的重要预处理步骤。在患有与癌症相关的组织变化和形状变形的患者中，肝脏分割可能特别具有挑战性。本研究旨在评估最先进的深度学习 3D 肝脏分割算法在所有不同巴塞罗那临床肝癌（BCLC）肝癌分期中的泛化能力。
这项回顾性研究纳入了来自机构数据库的患者，这些患者具有动脉期 T1 加权磁共振成像和相应的手动肝脏分割。数据分为 70/15/15%用于训练/验证/测试，每个比例在 BCLC 分期中均相等。使用两种具有相同 U 型网络衍生结构的 3D 卷积神经网络进行训练，训练数据集大小相等：一种涵盖所有 BCLC 分期（“全分期网络”：AS-Net），另一种仅限于早期和中期 BCLC 分期（“早期-中期分期网络”：EIS-Net）。使用涵盖所有 BCLC 分期的数据集评估分割准确性，并使用 Wilcoxon 符号秩检验进行两两比较。
219 名符合纳入标准的患者（170 名男性，49 名女性，62.8±9.1 岁）来自所有 BCLC 分期。两个网络都使用 129 名患者进行训练：AS-Net 训练分别包含 19、74、18、8 和 10 名 BCLC 0、A、B、C 和 D 患者；EIS-Net 训练分别包含 21、86 和 22 名 BCLC 0、A 和 B 患者。AS-Net 和 EIS-Net 的 DSC（平均值±标准差）分别为 0.954±0.018 和 0.946±0.032（p<0.001）。AS-Net 的 0.956±0.014 在高级 BCLC 分期上明显优于 EIS-Net 的 0.941±0.038（p<0.001），并且在早期和中期阶段的分割性能也相似（AS-Net：0.952±0.021；EIS-Net：0.949±0.027；p=0.107）。
为了确保在独立于肝脏形状变形和肿瘤负担的情况下在癌症阶段具有稳健的分割性能，在涵盖所有 BCLC 阶段的异质成像数据上训练深度学习模型至关重要。
