Deep multi-instance transfer learning for pneumothorax classification in chest X-ray images
Deep multi-instance transfer learning for pneumothorax classification in chest X-ray images
复制标题
DOI：
10.1002/mp.15328
复制
发表时间：
2021-12-07
期刊：
影响因子：
3.8
通讯作者：
Qian, Dahong
中科院分区：
文献类型：
--
作者：
Tian, Yuchi;Wang, Jiawei;Qian, Dahong
目的气胸是一种危及生命的急症，需要立即治疗。在临床实践中，胸部正位X光图像通常用于气胸的检测。然而，人工审查X线片耗时、劳动强度大，并且高度依赖放射科医生的经验，这可能会导致误诊。在这里，我们的目的是开发一种可靠的自动分类方法，以帮助放射科医生快速准确地诊断额位胸片中的气胸。方法提出了一种基于残差神经网络(ResNet)的气胸识别两阶段深度学习策略：先局部特征学习(LFL)，然后进行全局多示例学习(GMIL)。为了学习判别特征，去除了图像中的大部分非病变区域。有两个数据集用于大规模验证：一个私人数据集(27 955张正视胸部X光图像)和一个公共数据集(美国国立卫生研究院胸片X光图像14；正视X光图像112 120张)。用准确度、精确度、召回率、特异度、F1评分、受试者操作特征(ROC)和ROC曲线下面积(AUC)来评价识别模型的性能。对数据集进行五次交叉验证，然后计算上述指标的均值和标准差，以评估模型的整体性能。结果实验结果表明，该学习策略在NIH数据集上的准确率、AUC、准确率、召回率、特异度和F1-Score分别为94.4%+/-0.7%、97.3%+/-0.5%、94.2%+/-0.3%、94.6%+/-1.5%、94.2%+/-0.4%和94.4%+/-0.7%。结论实验结果表明，该系统是一种有效的气胸识别辅助工具。
Purpose Pneumothorax is a life-threatening emergency that requires immediate treatment. Frontal-view chest X-ray images are typically used for pneumothorax detection in clinical practice. However, manual review of radiographs is time-consuming, labor-intensive, and highly dependent on the experience of radiologists, which may lead to misdiagnosis. Here, we aim to develop a reliable automatic classification method to assist radiologists in rapidly and accurately diagnosing pneumothorax in frontal chest radiographs. Methods A novel residual neural network (ResNet)-based two-stage deep-learning strategy is proposed for pneumothorax identification: local feature learning (LFL) followed by global multi-instance learning (GMIL). Most of the nonlesion regions in the images are removed for learning discriminative features. Two datasets are used for large-scale validation: a private dataset (27 955 frontal-view chest X-ray images) and a public dataset (the National Institutes of Health [NIH] ChestX-ray14; 112 120 frontal-view X-ray images). The model performance of the identification was evaluated using the accuracy, precision, recall, specificity, F1-score, receiver operating characteristic (ROC), and area under ROC curve (AUC). Fivefold cross-validation is conducted on the datasets, and then the mean and standard deviation of the above-mentioned metrics are calculated to assess the overall performance of the model. Results The experimental results demonstrate that the proposed learning strategy can achieve state-of-the-art performance on the NIH dataset with an accuracy, AUC, precision, recall, specificity, and F1-score of 94.4% +/- 0.7%, 97.3% +/- 0.5%, 94.2% +/- 0.3%, 94.6% +/- 1.5%, 94.2% +/- 0.4%, and 94.4% +/- 0.7%, respectively. Conclusions The experimental results demonstrate that our proposed CAD system is an efficient assistive tool in the identification of pneumothorax.