Xiao Mengsu, Zhao Chenyang, Zhu Qingli, Zhang Jing, Liu He, Li Jianchu, Jiang Yuxin
Department of Ultrasound, Chinese Academy of Medical Sciences and Peking Union Medical College Hospital, Beijing 100730, China.
J Thorac Dis. 2019 Dec;11(12):5023-5031. doi: 10.21037/jtd.2019.12.10.
Deep learning-based computer-aided diagnosis (CAD) is an important method in aiding diagnosis for radiologists. We investigated the accuracy of a deep learning-based CAD in classifying breast lesions with different histological types.
A total of 448 breast lesions were detected on ultrasound (US) and classified by an experienced radiologist, a resident and deep learning-based CAD respectively. The pathological results of the lesions were chosen as the golden standard. The diagnostic performances of the three raters in different pathological types were analyzed.
For the overall diagnostic performance, deep learning-based CAD presented a significantly higher specificity (76.96%) compared with the two radiologists. The area under ROC of CAD was almost equal with the experienced radiologist (0.81 0.81), while significantly higher than the resident (0.81 0.70, P<0.0001). In the benign lesions, deep learning-based CAD had a higher accuracy than both the two radiologists, which correctly classified as benign lesions in 119/135 of fibroadenomas (88.1%), 25/35 of adenosis (71.4%), 14/27 of intraductal papillary tumors (51.9%), 5/10 of inflammation (50%), and 4/8 of sclerosing adenosis (50%). But only the differences between CAD and the two radiologists in fibroadenomas had statistical significance (P=0.0011 and P=0.0313), and the differences between CAD and the resident in adenosis had statistical significance (P=0.012). In the malignant lesions, 151/168 of invasive ductal carcinomas (89.9%), 21/29 of ductal carcinoma in situ (DCIS) (72.4%) and 6/7 of invasive lobular carcinomas (85.7%) were diagnosed as malignancies by deep learning-based CAD, with no significant differences between CAD and the two radiologists.
In the diagnosis of these common types of breast lesions, deep learning-based CAD had a satisfying performance. Deep learning-based CAD had a better performance in the breast benign lesions, especially in fibroadenomas and adenosis. Therefore, deep learning-based CAD is a promising supplemental tool to US to increase the specificity and avoid unnecessary benign biopsies.
基于深度学习的计算机辅助诊断（CAD）是辅助放射科医生进行诊断的重要方法。我们研究了基于深度学习的CAD对不同组织学类型乳腺病变进行分类的准确性。
共检测出448例乳腺病变的超声（US）图像，并分别由一名经验丰富的放射科医生、一名住院医生和基于深度学习的CAD进行分类。病变的病理结果作为金标准。分析了三位评估者在不同病理类型中的诊断性能。
在总体诊断性能方面，基于深度学习的CAD与两位放射科医生相比，特异性显著更高（76.96%）。CAD的ROC曲线下面积与经验丰富的放射科医生几乎相等（0.81对0.81），但显著高于住院医生（0.81对0.70，P<0.0001）。在良性病变中，基于深度学习的CAD的准确性高于两位放射科医生，在119/135例纤维腺瘤（88.1%）、25/35例腺病（71.4%）、14/27例导管内乳头状瘤（51.9%）、5/10例炎症（50%）和4/8例硬化性腺病（50%）中正确分类为良性病变。但只有CAD与两位放射科医生在纤维腺瘤中的差异具有统计学意义（P=0.0011和P=0.0313），CAD与住院医生在腺病中的差异具有统计学意义（P=0.012）。在恶性病变中，基于深度学习的CAD将151/168例浸润性导管癌（89.9%）、21/29例导管原位癌（DCIS）（72.4%）和6/7例浸润性小叶癌（85.7%）诊断为恶性肿瘤，CAD与两位放射科医生之间无显著差异。
在这些常见类型乳腺病变的诊断中，基于深度学习的CAD表现令人满意。基于深度学习的CAD在乳腺良性病变中表现更好，尤其是在纤维腺瘤和腺病中。因此，基于深度学习的CAD是一种有前景的US辅助工具，可提高特异性并避免不必要的良性活检。
