Breast Cancer Diagnosis in Digital Breast Tomosynthesis: Effects of Training Sample Size on Multi-Stage Transfer Learning Using Deep Neural Nets
Breast Cancer Diagnosis in Digital Breast Tomosynthesis: Effects of Training Sample Size on Multi-Stage Transfer Learning Using Deep Neural Nets
复制标题
DOI：
10.1109/tmi.2018.2870343
复制
发表时间：
2019-03-01
影响因子：
10.6
通讯作者：
Cha, Kenny H.
中科院分区：
文献类型：
--
作者：
Samala, Ravi K.;Chan, Heang-Ping;Cha, Kenny H.
在本文中，我们开发了一种深度卷积神经网络（CNN），用于数字乳腺断层合成摄影（DBT）中恶性和良性肿块的分类，使用多阶段转移学习方法，利用来自类似辅助域的数据进行中期微调。收集了来自DBT、数字化屏片乳腺X线摄影和数字化乳腺X线摄影的乳腺成像数据，共计4039个独特的感兴趣区域（1797个恶性区域和2242个良性区域）。使用交叉验证，我们通过改变卷积层被冻结的级别，从六个传输网络中选择了最佳传输网络。在单阶段迁移学习方法中，来自在ImageNet数据上训练的CNN的知识直接使用DBT数据进行微调。在多阶段迁移学习方法中，从ImageNet中学习到的知识首先使用乳房X光检查数据进行微调，然后使用DBT数据进行微调。通过冻结大多数CNN结构与仅冻结第一卷积层，比较了两个转移网络的第二阶段转移学习。我们通过将训练数据从1%变化到100%的可用集，研究了各种迁移学习和微调方案的分类性能对训练样本大小的依赖性。使用受试者工作特征曲线下面积（AUC）作为性能指标。单阶段迁移学习测试集上基于视图的AUC为0.85 +/- 0.05，多阶段学习显著提高（p < 0.05）至0.91 +/- 0.03。本文证明，当目标域的训练样本量有限时，使用来自类似辅助域的数据进行额外的迁移学习阶段是有利的。
In this paper, we developed a deep convolutional neural network (CNN) for the classification of malignant and benign masses in digital breast tomosynthesis (DBT) using amulti-stage transfer learning approach that utilized data from similar auxiliary domains for intermediate-stage fine-tuning. Breast imaging data from DBT, digitized screen-film mammography, and digital mammography totaling 4039 unique regions of interest (1797 malignant and 2242 benign) were collected. Using cross validation, we selected the best transfer network from six transfer networks by varying the level up to which the convolutional layers were frozen. In a single-stage transfer learning approach, knowledge from CNN trained on the ImageNet data was fine-tuned directly with the DBT data. In a multi-stage transfer learning approach, knowledge learned from ImageNet was first fine-tuned with the mammography data and then fine-tuned with the DBT data. Two transfer networks were compared for the second-stage transfer learning by freezing most of the CNN structures versus freezing only the first convolutional layer. We studied the dependence of the classification performance on training sample size for various transfer learning and fine-tuning schemes by varying the training data from 1% to 100% of the available sets. The area under the receiver operating characteristic curve (AUC) was used as a performance measure. The view based AUC on the test set for single-stage transfer learning was 0.85 +/- 0.05 and improved significantly (p < 0.05) to 0.91 +/- 0.03 for multi-stage learning. This paper demonstrated that, when the training sample size from the target domain is limited, an additional stage of transfer learning using data from a similar auxiliary domain is advantageous.