Department of Diagnostic and Interventional Radiology, Graduate School of Medicine, Osaka City University, 1-4-3 Asahi-machi, Abeno-ku, Osaka, 545-8585, Japan.
Department of Breast and Endocrine Surgery, Graduate School of Medicine, Osaka City University, 1-4-3 Asahi-machi, Abeno-ku, Osaka, 545-8585, Japan.
Jpn J Radiol. 2021 Apr;39(4):333-340. doi: 10.1007/s11604-020-01070-9. Epub 2020 Nov 16.
To demonstrate how artificial intelligence (AI) can expand radiologists' capacity, we visualized the features of invasive ductal carcinomas (IDCs) that our algorithm, developed and validated for basic pathological classification on mammograms, had focused on.
IDC datasets were built using mammograms from patients diagnosed with IDCs from January 2006 to December 2017. The developing dataset was used to train and validate a VGG-16 deep learning (DL) network. The true positives (TPs) and accuracy of the algorithm were externally evaluated using the test dataset. A visualization technique was applied to the algorithm to determine which malignant findings on mammograms were revealed.
The datasets were split into a developing dataset (988 images) and a test dataset (131 images). The proposed algorithm diagnosed 62 TPs with an accuracy of 0.61-0.70. The visualization of features on the mammograms revealed that the tubule forming, solid, and scirrhous types of IDCs exhibited visible features on the surroundings, corners of the masses, and architectural distortions, respectively.
We successfully showed that features isolated by a DL-based algorithm trained to classify IDCs were indeed those known to be associated with each pathology. Thus, using AI can expand the capacity of radiologists through the discovery of previously unknown findings.
为了展示人工智能（AI）如何扩展放射科医生的能力，我们对我们的算法所关注的浸润性导管癌（IDC）的特征进行了可视化，该算法是针对乳腺 X 光片中的基本病理分类而开发和验证的。
使用 2006 年 1 月至 2017 年 12 月期间被诊断为 IDC 的患者的乳腺 X 光片构建 IDC 数据集。开发数据集用于训练和验证 VGG-16 深度学习（DL）网络。使用测试数据集对外评估算法的真阳性（TP）和准确性。应用可视化技术来确定乳腺 X 光片中揭示的恶性发现。
数据集分为开发数据集（988 张图像）和测试数据集（131 张图像）。所提出的算法诊断出 62 个 TP，准确率为 0.61-0.70。对乳腺 X 光片上特征的可视化显示，管状形成、实体和硬癌型 IDC 在肿块的周围、角落和结构扭曲处分别显示出可见的特征。
我们成功地表明，通过训练来对 IDC 进行分类的基于 DL 的算法所分离的特征确实是与每种病理学相关的特征。因此，通过使用 AI 可以通过发现以前未知的发现来扩展放射科医生的能力。
