From the Department of Radiology, the Institute of Medical Science, the University of Tokyo (K.Y., H.A., A.K.); Department of Radiology, Graduate School of Medicine, the University of Tokyo (O.A.); and Department of Radiology, Graduate School of Medical Sciences, International University of Health and Welfare, 4-3 Kozunomori, Narita, Chiba, Japan 286-8686 (S.K.).
Radiology. 2018 Apr;287(1):146-155. doi: 10.1148/radiol.2017171928. Epub 2017 Dec 14.
Purpose To investigate the performance of a deep convolutional neural network (DCNN) model in the staging of liver fibrosis using gadoxetic acid-enhanced hepatobiliary phase magnetic resonance (MR) imaging. Materials and Methods This retrospective study included patients for whom input data (hepatobiliary phase MR images, static magnetic field of the imaging unit, and hepatitis B and C virus testing results available, either positive or negative) and reference standard data (liver fibrosis stage evaluated from biopsy or surgical specimens obtained within 6 months of the MR examinations) were available were assigned to the training (534 patients) or the test (100 patients) group. For the training group (54, 53, 81, 113, and 233 patients with fibrosis stages F0, F1, F2, F3, and F4, respectively; mean patient age, 67.4 ± 9.7 years; 388 men and 146 women), MR images with three different section levels were augmented 90-fold (rotated, parallel-shifted, brightness-changed and contrast-changed images were generated; a total of 144 180 images). Supervised training was performed by using the DCNN model to minimize the difference between the output data (fibrosis score obtained through deep learning [F score]) and liver fibrosis stage. The performance of the DCNN model was evaluated in the test group (10, 10, 15, 20, and 45 patients with fibrosis stages F0, F1, F2, F3, and F4, respectively; mean patient age, 66.8 years ± 10.7; 71 male patients and 29 female patients) with receiver operating characteristic (ROC) analyses. Results The F score was correlated significantly with fibrosis stage (Spearman rank correlation coefficient: 0.63; P < .001). Fibrosis stages F4, F3, and F2 were diagnosed with areas under the ROC curve of 0.84, 0.84, and 0.85, respectively. Conclusion The DCNN model exhibited a high diagnostic performance in the staging of liver fibrosis. RSNA, 2017 Online supplemental material is available for this article.
目的 利用钆塞酸增强肝胆期磁共振成像（MR）探讨深度卷积神经网络（DCNN）模型在肝纤维化分期中的性能。
材料与方法 本回顾性研究纳入了输入数据（肝胆期 MR 图像、成像单元的静态磁场以及乙型肝炎和丙型肝炎病毒检测结果，无论阳性还是阴性）和参考标准数据（活检或手术标本评估的肝纤维化分期，MR 检查后 6 个月内获得）均可用的患者，这些患者被分配到训练组（534 例）或测试组（100 例）。对于训练组（分别为纤维化分期 F0、F1、F2、F3 和 F4 的 54、53、81、113 和 233 例患者；平均患者年龄 67.4±9.7 岁；388 名男性和 146 名女性），MR 图像在三个不同层面进行了 90 倍扩充（生成旋转、平行移位、亮度和对比度变化的图像；共 144180 张图像）。通过 DCNN 模型最小化输出数据（通过深度学习获得的纤维化评分[F 评分]）和肝纤维化分期之间的差异来进行有监督训练。在测试组（纤维化分期 F0、F1、F2、F3 和 F4 的 10、10、15、20 和 45 例患者；平均患者年龄 66.8 岁±10.7 岁；71 名男性和 29 名女性）中，通过受试者工作特征（ROC）分析评估 DCNN 模型的性能。
结果 F 评分与纤维化分期显著相关（Spearman 秩相关系数：0.63；P<0.001）。纤维化分期 F4、F3 和 F2 的 ROC 曲线下面积分别为 0.84、0.84 和 0.85。
结论 DCNN 模型在肝纤维化分期中具有较高的诊断性能。
RSNA，2017
在线补充材料可在本文中获得。
