WCSE

CT Image Classification of Invasive Depth of Gastric Cancer based on 3D-DPN Structure

Abstract— With the role of accurate preoperative staging in improving the prognosis of gastric cancer patients is clear, how to effectively improve the preoperative depth of invasion has become one of the hot and difficult issues in the medical community. In recent years, using deep learning to classify computed tomography (CT) medical images effectively improves the accuracy and efficiency of classification. The traditional deep learning method processes the 3-dimension (3D) CT data into 2-dimension (2D) data, which is able to lose the spatial information of the data itself. We propose a 3D convolutional network to classify the depth of invasion of gastric cancer，which is effective to extract 3D volume feature of CT data. We also use self-supervised learning to overcome the shortage of medical image data. Combined with our contributions, we are able to achieve most 0.975 areas under the receiver operating characteristic curve (AUC). The overall results show better AUC performance than the state-of-the-art.

Index Terms— Invasive depth of gastric cancer, CT image classification, self-supervised learning

Wenzhi Bao
School of Data Science Engineering, East China Normal University, CHINA
Yinghui Jin
School of Data Science Engineering, East China Normal University, CHINA
Chenchao Huang
School of Data Science Engineering, East China Normal University, CHINA
Wei Peng
Information Technology Services, East China Normal University, CHINA

[Download]

Cite: Wenzhi Bao, Yinghui Jin, Chenchao Huang, Wei Peng, "CT Image Classification of Invasive Depth of Gastric Cancer based on 3D-DPN Structure ," 2021 The 11th International Workshop on Computer Science and Engineering (WCSE 2021), pp. 115-121, Shanghai, China, June 19-21, 2021.