AUTHOR=Fu Hao , Mi Weiming , Pan Boju , Guo Yucheng , Li Junjie , Xu Rongyan , Zheng Jie , Zou Chunli , Zhang Tao , Liang Zhiyong , Zou Junzhong , Zou Hao 

TITLE=Automatic Pancreatic Ductal Adenocarcinoma Detection in Whole Slide Images Using Deep Convolutional Neural Networks

JOURNAL=Frontiers in Oncology

VOLUME=Volume 11 - 2021

YEAR=2021

URL=https://www.frontiersin.org/journals/oncology/articles/10.3389/fonc.2021.665929

DOI=10.3389/fonc.2021.665929

ISSN=2234-943X

ABSTRACT=Pancreatic ductal adenocarcinoma(PDAC) is one of the deadliest cancer types worldwide, with the lowest 5-year survival rate among all kinds of cancers. Histopathology image analysis is considered a gold standard for PDAC recognition and diagnosis. However, the manual diagnosis used in current clinical practice is a tedious and time-consuming task and can lead to low diagnosis concordance. With the development of digital imaging and machine learning technology, several scholars have proposed PDAC detection approaches via feature extraction methods that rely on field-knowledge. Meanwhile, deep learning methods based on convolutional neural networks have been shown to make an accurate prediction by learning hidden features automatically. In this paper, a method for the classification and segmentation of pancreatic histopathological images using deep convolutional neural networks is proposed. All digital histopathological images used are authorized by Peking Union Medical College Hospital. Our automatic patch-level approach achieved 95.3% classification accuracy and the WSI-level achieved 100%. Additionally, we visualize the outcomes of classification and segmentation methods based on the same histopathological image to show which areas on the histopathological image are more important for pancreas ductal adenocarcinoma prediction, respectively. All these experimental results demonstrate that our proposed model can effectively diagnose PDAC using histopathological images, and show the powerful potential of practical application.