Department of Radiology, Juntendo University School of Medicine.
Magn Reson Med Sci. 2023 Jan 1;22(1):57-66. doi: 10.2463/mrms.mp.2021-0068. Epub 2021 Dec 10.
Myelination-related MR signal changes in white matter are helpful for assessing normal development in infants and children. A rule-based myelination evaluation workflow regarding signal changes on T1-weighted images (T1WIs) and T2-weighted images (T2WIs) has been widely used in radiology. This study aimed to simulate a rule-based workflow using a stacked deep learning model and evaluate age estimation accuracy.
The age estimation system involved two stacked neural networks: a target network-to extract five myelination-related images from the whole brain, and an age estimation network from extracted T1- and T2WIs separately. A dataset was constructed from 119 children aged below 2 years with two MRI systems. A four-fold cross-validation method was adopted. The correlation coefficient (CC), mean absolute error (MAE), and root mean squared error (RMSE) of the corrected chronological age of full-term birth, as well as the mean difference and the upper and lower limits of 95% agreement, were measured. Generalization performance was assessed using datasets acquired from different MR images. Age estimation was performed in Sturge-Weber syndrome (SWS) cases.
There was a strong correlation between estimated age and corrected chronological age (MAE: 0.98 months; RMSE: 1.27 months; and CC: 0.99). The mean difference and standard deviation (SD) were -0.15 and 1.26, respectively, and the upper and lower limits of 95% agreement were 2.33 and -2.63 months. Regarding generalization performance, the performance values on the external dataset were MAE of 1.85 months, RMSE of 2.59 months, and CC of 0.93. Among 13 SWS cases, 7 exceeded the limits of 95% agreement, and a proportional bias of age estimation based on myelination acceleration was exhibited below 12 months of age (P = 0.03).
Stacked deep learning models automated the rule-based workflow in radiology and achieved highly accurate age estimation in infants and children up to 2 years of age.
脑白质髓鞘形成相关的 MR 信号变化有助于评估婴儿和儿童的正常发育。基于规则的髓鞘评估工作流程已广泛应用于影像学中，该流程涉及 T1 加权图像（T1WI）和 T2 加权图像（T2WI）上的信号变化。本研究旨在使用堆叠深度学习模型模拟基于规则的工作流程，并评估年龄估计的准确性。
年龄估计系统涉及两个堆叠神经网络：目标网络——从全脑提取五个髓鞘形成相关图像，以及分别从提取的 T1 和 T2WI 中提取年龄估计网络。使用来自两个 MRI 系统的 119 名年龄在 2 岁以下的儿童构建了一个数据集。采用四折交叉验证方法。测量全期出生时校正后的实际年龄的相关系数（CC）、平均绝对误差（MAE）和均方根误差（RMSE），以及平均差值和 95%一致性的上下限。使用来自不同 MRI 图像的数据集评估泛化性能。对斯特奇-韦伯综合征（SWS）病例进行年龄估计。
估计年龄与校正后的实际年龄具有很强的相关性（MAE：0.98 个月；RMSE：1.27 个月；CC：0.99）。平均差值和标准差（SD）分别为-0.15 和 1.26，95%一致性的上下限分别为 2.33 和-2.63 个月。关于泛化性能，外部数据集的性能值为 MAE 为 1.85 个月，RMSE 为 2.59 个月，CC 为 0.93。在 13 例 SWS 病例中，有 7 例超出了 95%一致性的上限，并且在 12 个月以下的年龄表现出基于髓鞘加速的年龄估计的比例偏差（P = 0.03）。
堆叠深度学习模型自动化了放射科基于规则的工作流程，并在 2 岁以下的婴儿和儿童中实现了高度准确的年龄估计。
