Wang Zhuo, Yu Zhezhou, Wang Yao, Zhang Huimao, Luo Yishan, Shi Lin, Wang Yan, Guo Chunjie
Key Laboratory of Symbol Computation & Knowledge Engineering, Ministry of Education, College of Computer Science & Technology, Jilin University, Changchun, China.
Department of Radiology, the First Hospital of Jilin University, Changchun, China.
Front Physiol. 2020 Dec 23;11:612928. doi: 10.3389/fphys.2020.612928. eCollection 2020.
Magnetic resonance imaging (MRI) has a wide range of applications in medical imaging. Recently, studies based on deep learning algorithms have demonstrated powerful processing capabilities for medical imaging data. Previous studies have mostly focused on common diseases that usually have large scales of datasets and centralized the lesions in the brain. In this paper, we used deep learning models to process MRI images to differentiate the rare neuromyelitis optical spectrum disorder (NMOSD) from multiple sclerosis (MS) automatically, which are characterized by scattered and overlapping lesions.
We proposed a novel model structure to capture 3D MRI images' essential information and converted them into lower dimensions. To empirically prove the efficiency of our model, firstly, we used a conventional 3-dimensional (3D) model to classify the T2-weighted fluid-attenuated inversion recovery (T2-FLAIR) images and proved that the traditional 3D convolutional neural network (CNN) models lack the learning capacity to distinguish between NMOSD and MS. Then, we compressed the 3D T2-FLAIR images by a two-view compression block to apply two different depths (18 and 34 layers) of 2D models for disease diagnosis and also applied transfer learning by pre-training our model on ImageNet dataset.
We found that our models possess superior performance when our models were pre-trained on ImageNet dataset, in which the models' average accuracies of 34 layers model and 18 layers model were 0.75 and 0.725, sensitivities were 0.707 and 0.708, and specificities were 0.759 and 0.719, respectively. Meanwhile, the traditional 3D CNN models lacked the learning capacity to distinguish between NMOSD and MS.
The novel CNN model we proposed could automatically differentiate the rare NMOSD from MS, especially, our model showed better performance than traditional3D CNN models. It indicated that our 3D compressed CNN models are applicable in handling diseases with small-scale datasets and possess overlapping and scattered lesions.
磁共振成像（MRI）在医学成像中有着广泛的应用。近年来，基于深度学习算法的研究已证明对医学成像数据具有强大的处理能力。以往的研究大多集中在通常具有大规模数据集且病变集中在脑部的常见疾病上。在本文中，我们使用深度学习模型处理MRI图像，以自动区分罕见的视神经脊髓炎谱系障碍（NMOSD）和多发性硬化症（MS），这两种疾病的特征是病变分散且相互重叠。
我们提出了一种新颖的模型结构来捕捉3D MRI图像的基本信息并将其转换为更低维度。为了通过实验证明我们模型的效率，首先，我们使用传统的三维（3D）模型对T2加权液体衰减反转恢复（T2-FLAIR）图像进行分类，并证明传统的3D卷积神经网络（CNN）模型缺乏区分NMOSD和MS的学习能力。然后，我们通过双视图压缩块对3D T2-FLAIR图像进行压缩，以应用两种不同深度（18层和34层）的2D模型进行疾病诊断，并且还通过在ImageNet数据集上对我们的模型进行预训练来应用迁移学习。
我们发现，当我们的模型在ImageNet数据集上进行预训练时，它们具有卓越的性能，其中34层模型和18层模型的平均准确率分别为0.75和0.725，敏感度分别为0.707和0.708，特异性分别为0.759和0.719。同时，传统的3D CNN模型缺乏区分NMOSD和MS的学习能力。
我们提出的新型CNN模型可以自动区分罕见的NMOSD和MS，特别是，我们的模型表现出比传统3D CNN模型更好的性能。这表明我们的3D压缩CNN模型适用于处理具有小规模数据集且病变相互重叠和分散的疾病。
