Yang Kaiqiang, Liu Jinsha, Tang Wen, Zhang Huiling, Zhang Rongguo, Gu Jun, Zhu Ruiping, Xiong Jingtong, Ru Xiaoshuang, Wu Jianlin
Department of Radiology, Zhongshan Hospital, Dalian University, Dalian, Liaoning Province, China; Infervision, Beijing, China.
Department of Radiology, Zhongshan Hospital, Dalian University, Dalian, Liaoning Province, China.
Eur J Radiol. 2020 Aug;129:109013. doi: 10.1016/j.ejrad.2020.109013. Epub 2020 May 23.
To accurately distinguish benign from malignant pulmonary nodules with CT based on partial structures of 3D U-Net integrated with Capsule Networks (CapNets) and provide a reference for the early diagnosis of lung cancer.
The dataset consisted of 1177 samples (benign/malignant: 414/763) from 997 patients provided by collaborating hospital. All nodules were biopsy or surgery proven, and pathologic results were regarded as the "golden standard". This study utilized partial U-Net to capture the low-level (edge, corner, etc.) information and CapNets to preserve high-level (semantic information) information of nodules. For CapNets, each capsule had a 4 × 4 matrix representing the pose and an activation probability representing the presence of an object. Furthermore, we chose accuracy (ACC), area under the curve (AUC), sensitivity (SE) and specificity (SP) to evaluate the generalization of the proposed architecture and compared its identification performance with 3D U-Net and experienced radiologists.
The AUC of our architecture (0.84) was superior to that (0.81) of the original 3D U-Net (p = 0.04, DeLong's test). Moreover, ACC (84.5 %) and SE (92.9 %) of our model were clearly higher than radiologists' ACC (81.0 %) and SE (84.3 %) at the optimal operating point. However, SP (70 %) of our model was slightly lower than radiologists' SP (75 %), which might be the result of class imbalance with limited benign samples involved for algorithm training.
Our architecture showed a high performance for identifying benign and malignant pulmonary nodules, indicating the improved model has a promising application in clinic.
基于集成胶囊网络（CapNets）的3D U-Net部分结构，利用CT准确区分肺结节的良恶性，为肺癌早期诊断提供参考。
数据集由合作医院提供的997例患者的1177个样本（良性/恶性：414/763）组成。所有结节均经活检或手术证实，病理结果被视为“金标准”。本研究利用部分U-Net捕捉结节的低级（边缘、角落等）信息，利用CapNets保留结节的高级（语义信息）信息。对于CapNets，每个胶囊有一个表示姿态的4×4矩阵和一个表示物体存在的激活概率。此外，我们选择准确率（ACC）、曲线下面积（AUC）、灵敏度（SE）和特异度（SP）来评估所提架构的泛化能力，并将其识别性能与3D U-Net和经验丰富的放射科医生进行比较。
我们架构的AUC（0.84）优于原始3D U-Net的AUC（0.81）（p = 0.04，德龙检验）。此外，在最佳工作点，我们模型的ACC（84.5%）和SE（92.9%）明显高于放射科医生的ACC（81.0%）和SE（84.3%）。然而，我们模型的SP（70%）略低于放射科医生的SP（75%），这可能是由于算法训练中良性样本有限导致的类别不平衡所致。
我们的架构在识别肺结节良恶性方面表现出高性能，表明改进后的模型在临床上具有广阔的应用前景。
