已发表论文

通过整合生物信息学和机器学习识别炎症性肠病和胆管癌的共有诊断生物标志物 LCN2 和 DUOX2

 

Authors Yuan S, Hu Z, Xu Z, Qian J, Wang X

Received 6 August 2025

Accepted for publication 6 November 2025

Published 14 November 2025 Volume 2025:18 Pages 15939—15958

DOI https://doi.org/10.2147/JIR.S558859

Checked for plagiarism Yes

Review by Single anonymous peer review

Peer reviewer comments 2

Editor who approved publication: Dr Fatih Türker

Shiqing Yuan,1,2,* Zhen Hu,2,* Zihan Xu,1,2 Jiarun Qian,1,2 Xiaoyun Wang2 

1School of Wuxi Medicine, Jiangnan University, Wuxi, Jiangsu, 214000, People’s Republic of China; 2Division of Gastroenterology, Jiangnan University Medical Center (Wuxi No. 2 People’s Hospital), Wuxi, Jiangsu, 214000, People’s Republic of China

*These authors contributed equally to this work

Correspondence: Xiaoyun Wang, Division of Gastroenterology, Jiangnan University Medical Center (Wuxi No. 2 People’s Hospital), Wuxi, Jiangsu, 214000, People’s Republic of China, Email xiaoyunwang68@aliyun.com

Background: Inflammatory bowel disease (IBD) is a common chronic inflammatory illness affecting the gastrointestinal tract. Cholangiocarcinoma (CCA) comprises a group of highly heterogeneous cancers originating from the biliary tract. Epidemiological evidence suggests a positive correlation between IBD and CCA. However, the pathological mechanisms are unestablished. This research aimed to identify common biomarkers and immune infiltration features in IBD and CCA to support early detection and development of targeted therapies.
Methods: Based on transcriptomic data from the GEO and TCGA databases, candidate genes were obtained by integrating differentially expressed genes (DEGs) with key modules derived from weighted gene co-expression network analysis (WGCNA). Hub genes were subsequently screened using four machine learning algorithms. A comprehensive validation of the hub genes (LCN2 and DUOX2) was performed through receiver operating characteristic (ROC) curve analysis, survival analysis, immune infiltration profiling (CIBERSORT), and drug sensitivity assessment.
Results: By overlapping DEGs with key WGCNA modules, thirteen candidate genes were obtained. Eight common hub genes (CCL11, CCL20, DUOX2, DUOXA2, LCN2, NOS2, PDZK1IP1, and TRIM40) were ultimately selected through integrated ML approaches. ROC analysis demonstrated that LCN2 and DUOX2 exhibited excellent diagnostic performance for both IBD and CCA, as reflected by high area under the curve values.
Conclusion: This research confirmed that LCN2 and DUOX2 are key shared genes in the comorbidity of IBD and CCA. It also highlighted the distinct immune infiltration features between the two diseases. These findings provide novel perspectives on the pathogenesis of IBD and CCA, thereby offering a basis for early screening of CCA in individuals with IBD.

Keywords: inflammatory bowel disease, cholangiocarcinoma, biomarkers, molecular mechanisms