한빛사 논문
Yongju Lee # 1, Jeong Hwan Park # 2 3, Sohee Oh # 4, Kyoungseob Shin # 1, Jiyu Sun 4, Minsun Jung 2 5, Cheol Lee 2 6, Hyojin Kim 2 7, Jin-Haeng Chung 2 7, Kyung Chul Moon 8 9, Sunghoon Kwon 10 11 12 13 14 15
1Department of Electrical and Computer Engineering, Seoul National University, Seoul, Republic of Korea.
2Department of Pathology, Seoul National University College of Medicine, Seoul, Republic of Korea.
3Department of Pathology, SMG-SNU Boramae Medical Center, Seoul, Republic of Korea.
4Medical Research Collaborating Center, SMG-SNU Boramae Medical Center, Seoul, Republic of Korea.
5Department of Pathology, Severance Hospital, Yonsei University College of Medicine, Seoul, Republic of Korea.
6Department of Pathology, Seoul National University Hospital, Seoul, Republic of Korea.
7Department of Pathology and Translational Medicine, Seoul National University Bundang Hospital, Seongnam, Republic of Korea.
8Department of Pathology, Seoul National University College of Medicine, Seoul, Republic of Korea.
9Department of Pathology, Seoul National University Hospital, Seoul, Republic of Korea.
10Department of Electrical and Computer Engineering, Seoul National University, Seoul, Republic of Korea.
11Interdisciplinary Program in Bioengineering, Seoul National University, Seoul, Republic of Korea.
12Bio-MAX Institute, Seoul National University, Seoul, Republic of Korea.
13BK21+ Creative Research Engineer Development for IT, Seoul National University, Seoul, Republic of Korea.
14Biomedical Research Institute, Seoul National University, Seoul, Republic of Korea.
15Institutes of Entrepreneurial BioConvergence, Seoul National University, Seoul, Republic of Korea.
#Contributed equally.
These authors contributed equally: Yongju Lee, Jeong Hwan Park, Sohee Oh, Kyoungseob Shin.
Corresponding authors: Correspondence to Kyung Chul Moon or Sunghoon Kwon.
Abstract
Methods of computational pathology applied to the analysis of whole-slide images (WSIs) do not typically consider histopathological features from the tumour microenvironment. Here, we show that a graph deep neural network that considers such contextual features in gigapixel-sized WSIs in a semi-supervised manner can provide interpretable prognostic biomarkers. We designed a neural-network model that leverages attention techniques to learn features of the heterogeneous tumour microenvironment from memory-efficient representations of aggregates of highly correlated image patches. We trained the model with WSIs of kidney, breast, lung and uterine cancers and validated it by predicting the prognosis of 3,950 patients with these four different types of cancer. We also show that the model provides interpretable contextual features of clear cell renal cell carcinoma that allowed for the risk-based retrospective stratification of 1,333 patients. Deep graph neural networks that derive contextual histopathological features from WSIs may aid diagnostic and prognostic tasks.
논문정보
관련 링크
연구자 ID
관련분야 연구자보기
소속기관 논문보기
관련분야 논문보기