Wang Shu-Hui, Han Xin-Jun, Du Jing, Wang Zhen-Chang, Yuan Chunwang, Chen Yinan, Zhu Yajing, Dou Xin, Xu Xiao-Wei, Xu Hui, Yang Zheng-Han
Department of Radiology, Beijing Friendship Hospital, Capital Medical University, No. 95 Yongan Road, Xicheng District, Beijing, 100050, People's Republic of China.
Department of Radiology, Weihai Municipal Hospital, Cheeloo College of Medicine, Shandong University, Weihai, Shandong Province, People's Republic of China.
Insights Imaging. 2021 Nov 24;12(1):173. doi: 10.1186/s13244-021-01117-z.
The imaging features of focal liver lesions (FLLs) are diverse and complex. Diagnosing FLLs with imaging alone remains challenging. We developed and validated an interpretable deep learning model for the classification of seven categories of FLLs on multisequence MRI and compared the differential diagnosis between the proposed model and radiologists.
In all, 557 lesions examined by multisequence MRI were utilised in this retrospective study and divided into training-validation (n = 444) and test (n = 113) datasets. The area under the receiver operating characteristic curve (AUC) was calculated to evaluate the performance of the model. The accuracy and confusion matrix of the model and individual radiologists were compared. Saliency maps were generated to highlight the activation region based on the model perspective.
The AUC of the two- and seven-way classifications of the model were 0.969 (95% CI 0.944-0.994) and from 0.919 (95% CI 0.857-0.980) to 0.999 (95% CI 0.996-1.000), respectively. The model accuracy (79.6%) of the seven-way classification was higher than that of the radiology residents (66.4%, p = 0.035) and general radiologists (73.5%, p = 0.346) but lower than that of the academic radiologists (85.4%, p = 0.291). Confusion matrices showed the sources of diagnostic errors for the model and individual radiologists for each disease. Saliency maps detected the activation regions associated with each predicted class.
This interpretable deep learning model showed high diagnostic performance in the differentiation of FLLs on multisequence MRI. The analysis principle contributing to the predictions can be explained via saliency maps.
肝脏局灶性病变（FLLs）的影像学特征多样且复杂。仅依靠影像学诊断FLLs仍具有挑战性。我们开发并验证了一种可解释的深度学习模型，用于在多序列磁共振成像（MRI）上对七类FLLs进行分类，并比较了该模型与放射科医生之间的鉴别诊断能力。
本回顾性研究共纳入了557个经多序列MRI检查的病变，并将其分为训练-验证集（n = 444）和测试集（n = 113）。计算受试者操作特征曲线（ROC）下的面积（AUC）以评估模型的性能。比较了模型和个体放射科医生的准确性及混淆矩阵。基于模型视角生成显著性图以突出激活区域。
该模型的二元和七元分类的AUC分别为0.969（95%可信区间0.944 - 0.994）和0.919（95%可信区间0.857 - 0.980）至0.999（95%可信区间0.996 - 1.000）。七元分类的模型准确率（79.6%）高于放射科住院医师（66.4%，p = 0.035）和普通放射科医生（73.5%，p = 0.346），但低于学术放射科医生（85.4%，p = 0.291）。混淆矩阵显示了模型和个体放射科医生对每种疾病的诊断错误来源。显著性图检测到了与每个预测类别相关的激活区域。
这种可解释的深度学习模型在多序列MRI上对FLLs的鉴别诊断中表现出较高的诊断性能。有助于预测的分析原理可通过显著性图进行解释。
