The potential and promise of deep learning systems to provide an independent assessment and relieve radiologists’ burden in screening mammography have been recognized in several studies. However, the low cancer prevalence, the need to process high-resolution images, and the need to combine information from multiple views and scales still pose technical challenges. Multi-view architectures that combine information from the four mammographic views to produce an exam-level classification score are a promising approach to the automated processing of screening mammography. However, training such architectures from exam-level labels, without relying on pixel-level supervision, requires very large datasets and may result in suboptimal accuracy. Emerging architectures such as Visual Transformers (ViT) and graph-based architectures can potentially integrate ipsi-lateral and contra-lateral breast views better than traditional convolutional neural networks, thanks to their stronger ability of modeling long-range dependencies. In this paper, we extensively evaluate novel transformer-based and graph-based architectures against state-of-the-art multi-view convolutional neural networks, trained in a weakly-supervised setting on a middle-scale dataset, both in terms of performance and interpretability. Extensive experiments on the CSAW dataset suggest that, while transformer-based architecture outperform other architectures, different inductive biases lead to complementary strengths and weaknesses, as each architecture is sensitive to different signs and mammographic features. Hence, an ensemble of different architectures should be preferred over a winner-takes-all approach to achieve more accurate and robust results. Overall, the findings highlight the potential of a wide range of multi-view architectures for breast cancer classification, even in datasets of relatively modest size, although the detection of small lesions remains challenging without pixel-wise supervision or ad-hoc networks.

中文翻译：

---

使用多视图深度学习技术进行乳房X线摄影分类：研究基于图形和变压器的架构


深度学习系统提供独立评估并减轻放射科医生筛查乳房X光检查负担的潜力和前景已在多项研究中得到认可。然而，癌症患病率低、处理高分辨率图像的需要以及结合多个视图和尺度的信息的需要仍然构成技术挑战。多视图架构结合了来自四个乳房 X 线摄影视图的信息以产生检查级分类分数，是一种有前途的筛查乳房 X 线摄影自动化处理方法。然而，从考试级标签训练此类架构而不依赖于像素级监督，需要非常大的数据集，并且可能会导致准确性不佳。诸如视觉变换器（ViT）和基于图的架构等新兴架构可以比传统的卷积神经网络更好地整合同侧和对侧乳房视图，这要归功于它们更强的远程依赖性建模能力。在本文中，我们针对最先进的多视图卷积神经网络广泛评估了新颖的基于变压器和基于图的架构，这些网络在中等规模数据集的弱监督设置中进行了训练，无论是在性能方面和可解释性。 CSAW 数据集上的大量实验表明，虽然基于变压器的架构优于其他架构，但不同的归纳偏差会导致优势和劣势互补，因为每种架构对不同的体征和乳房 X 线摄影特征敏感。因此，应该优先选择不同架构的集成，而不是赢家通吃的方法，以获得更准确和稳健的结果。 总体而言，这些发现凸显了多种多视图架构用于乳腺癌分类的潜力，即使在规模相对较小的数据集中也是如此，尽管在没有像素级监督或临时网络的情况下，小病变的检测仍然具有挑战性。