Accurate recognition of colorectal cancer with semi-supervised deep learning on pathological images

Gang Yu, Kai Sun, Chao Xu, Xing-Hua Shi, Chong Wu, Ting Xie, Run-Qi Meng, Xiang-He Meng, Kuan-Song Wang (), Hong-Mei Xiao () and Hong-Wen Deng ()
Additional contact information
Gang Yu: School of Basic Medical Science, Central South University
Kai Sun: School of Basic Medical Science, Central South University
Chao Xu: University of Oklahoma Health Sciences Center
Xing-Hua Shi: College of Science and Technology, Temple University
Chong Wu: Florida State University
Ting Xie: School of Basic Medical Science, Central South University
Run-Qi Meng: Electronic Information Science and Technology, School of Physics and Electronics, Central South University
Xiang-He Meng: Center for System Biology, Data Sciences and Reproductive Health, School of Basic Medical Science, Central South University
Kuan-Song Wang: Xiangya Hospital, School of Basic Medical Science, Central South University
Hong-Mei Xiao: Center for System Biology, Data Sciences and Reproductive Health, School of Basic Medical Science, Central South University
Hong-Wen Deng: Center for System Biology, Data Sciences and Reproductive Health, School of Basic Medical Science, Central South University

Nature Communications, 2021, vol. 12, issue 1, 1-13

Abstract: Abstract Machine-assisted pathological recognition has been focused on supervised learning (SL) that suffers from a significant annotation bottleneck. We propose a semi-supervised learning (SSL) method based on the mean teacher architecture using 13,111 whole slide images of colorectal cancer from 8803 subjects from 13 independent centers. SSL (~3150 labeled, ~40,950 unlabeled; ~6300 labeled, ~37,800 unlabeled patches) performs significantly better than the SL. No significant difference is found between SSL (~6300 labeled, ~37,800 unlabeled) and SL (~44,100 labeled) at patch-level diagnoses (area under the curve (AUC): 0.980 ± 0.014 vs. 0.987 ± 0.008, P value = 0.134) and patient-level diagnoses (AUC: 0.974 ± 0.013 vs. 0.980 ± 0.010, P value = 0.117), which is close to human pathologists (average AUC: 0.969). The evaluation on 15,000 lung and 294,912 lymph node images also confirm SSL can achieve similar performance as that of SL with massive annotations. SSL dramatically reduces the annotations, which has great potential to effectively build expert-level pathological artificial intelligence platforms in practice.

Date: 2021
References: View references in EconPapers View complete reference list from CitEc
Citations: View citations in EconPapers (4)

Downloads: (external link)
https://www.nature.com/articles/s41467-021-26643-8 Abstract (text/html)

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:nat:natcom:v:12:y:2021:i:1:d:10.1038_s41467-021-26643-8

Ordering information: This journal article can be ordered from
https://www.nature.com/ncomms/

DOI: 10.1038/s41467-021-26643-8

Access Statistics for this article

Nature Communications is currently edited by Nathalie Le Bot, Enda Bergin and Fiona Gillespie

More articles in Nature Communications from Nature
Bibliographic data for series maintained by Sonal Shukla () and Springer Nature Abstracting and Indexing ().