Johns Hopkins University Applied Physics Laboratory, Laurel, MD, USA; Malone Center for Engineering in Healthcare, Baltimore, MD, USA.
Johns Hopkins University Applied Physics Laboratory, Laurel, MD, USA.
Comput Biol Med. 2019 Feb;105:46-53. doi: 10.1016/j.compbiomed.2018.12.006. Epub 2018 Dec 11.
We address the challenge of finding anomalies in ultrasound images via deep learning, specifically applying this to screening for myopathies and finding rare presentations of myopathic disease. Among myopathic diseases, this study focuses on the use case of myositis given the spectrum of muscle involvement seen in these inflammatory muscle diseases, as well as the potential for treatment. For this study, we have developed a fully annotated dataset (called "Myositis3K") which includes 3586 images of eighty-nine individuals (35 control and 54 with myositis) acquired with informed consent. We approach this challenge as one of performing unsupervised novelty detection (ND), and use tools leveraging deep embeddings combined with several novelty scoring methods. We evaluated these various ND algorithms and compared their performance against human clinician performance, against other methods including supervised binary classification approaches, and against unsupervised novelty detection approaches using generative methods. Our best performing approach resulted in a (ROC) AUC (and 95% CI error margin) of 0.7192 (0.0164), which is a promising baseline for developing future clinical tools for unsupervised prescreening of myopathies.
我们通过深度学习解决了在超声图像中发现异常的挑战，特别是将其应用于肌病筛查，并发现肌病的罕见表现。在肌病中，本研究侧重于肌炎的应用案例，因为这些炎症性肌肉疾病中可见到肌肉受累的范围，以及可能的治疗方法。为此，我们开发了一个完全注释的数据集（称为“肌炎 3K”），其中包括 89 名个体的 3586 张图像（35 名对照和 54 名肌炎），这些图像是在知情同意的情况下采集的。我们将这一挑战视为执行无监督新颖性检测（ND）之一，并使用结合了几种新颖性评分方法的深度学习嵌入工具来解决这一挑战。我们评估了这些不同的 ND 算法，并将其性能与人类临床医生的性能进行了比较，与包括监督二进制分类方法在内的其他方法进行了比较，还与使用生成方法的无监督新颖性检测方法进行了比较。我们表现最好的方法的（ROC）AUC（和 95%置信区间误差幅度）为 0.7192（0.0164），这为开发用于肌病无监督预筛选的未来临床工具提供了一个有前途的基线。
