Development of deep learning-based narrow-band imaging endocytoscopic classification for predicting colorectal lesions from a retrospective study

Jie Wang, Mingqing Liu, Haiming Liao, Jiawei Fan, He Zhu, Tantan Ma, Dong Yang, Fengming Ni, Fan Zhang, Guohua Jin, Juan Yu, Jiahui He, Xiaokun Liang, Nan Zhang (), Hong Xu () and Zhicheng Zhang ()
Additional contact information
Jie Wang: The First Hospital of Jilin University
Mingqing Liu: The First Hospital of Jilin University
Haiming Liao: Boluo County Health Bureau
Jiawei Fan: The First Hospital of Jilin University
He Zhu: The First Hospital of Jilin University
Tantan Ma: The First Hospital of Jilin University
Dong Yang: The First Hospital of Jilin University
Fengming Ni: The First Hospital of Jilin University
Fan Zhang: The First Hospital of Jilin University
Guohua Jin: The First Hospital of Jilin University
Juan Yu: The First Affiliated Hospital of Shenzhen University
Jiahui He: University of Nottingham Ningbo China
Xiaokun Liang: Chinese Academy of Sciences
Nan Zhang: The First Hospital of Jilin University
Hong Xu: The First Hospital of Jilin University
Zhicheng Zhang: The First Hospital of Jilin University

Nature Communications, 2025, vol. 16, issue 1, 1-12

Abstract: Abstract Data-driven approaches have advanced colorectal lesion diagnosis in digestive endoscopy, yet their application in endocytoscopy (EC)—a high-magnification imaging technique—remains limited, with most studies relying on conventional machine learning methods like support vector machines. Inspired by the success of large-scale language models that leverage progressive pre-training, we develop a computer-aided diagnosis (CAD) model using narrow-band imaging endocytoscopy (EC-NBI) to classify colorectal lesions (non-neoplastic lesions, adenomas, and invasive cancers). Here, we show that our model, trained through a multi-stage pre-training strategy combined with supervised deep clustering, outperforms state-of-the-art supervised methods in a multi-center retrospective cohort. Notably, it surpasses endoscopists’ diagnostic accuracy in human-machine competitions and enhances their performance when used as an assistive tool. This EC-NBI CAD model significantly improves the accuracy and consistency of diagnosing colorectal lesions, laying a foundation for future early cancer screening, particularly for distinguishing superficial and deep submucosal invasive cancers, pending further expansive multi-center data.

Date: 2025
References: Add references at CitEc
Citations:

Downloads: (external link)
https://www.nature.com/articles/s41467-025-63812-5 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:16:y:2025:i:1:d:10.1038_s41467-025-63812-5

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

DOI: 10.1038/s41467-025-63812-5

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 ().