Title : Early Progression Prediction in Korean Crohn’s Disease Using a Korean-Specific PrediXcan Model

Journal: International Journal of Molecular Sciences

Authors:  Tae-woo Kim^1,† , Soo Kyung Park^1,2,†, Jaeyoung Chun³, Suji Kim⁴, Chang Hwan Choi⁵, Sang-Bum Kang⁶, Ki Bae Bang ⁷, Tae Oh Kim⁸, Geom Seog Seo⁹, Jae Myung Cha¹⁰ , Yunho Jung¹¹, Hyun Gun Kim¹², Jong Pil Im¹³ , Kwang Sung Ahn¹⁴ , Chang Kyun Lee¹⁵, Hyo Jong Kim¹⁵, Sangsoo Kim^4,* and Dong Il Park^1,2,*

Affiliations

¹ Division of Gastroenterology, Department of Internal Medicine and Inflammatory Bowel Disease Center, Kangbuk Samsung Hospital, Sungkyunkwan University School of Medicine, Seoul 03181, Republic of Korea

² Medical Research Institute, Kangbuk Samsung Hospital, Sungkyunkwan University School of Medicine, Seoul 03181, Republic of Korea

³ Department of Internal Medicine, Gangnam Severance Hospital, Yonsei University College of Medicine,Seoul 06273, Republic of Korea

⁴ Department of Bioinformatics, Soongsil University, Seoul 06978, Republic of Korea

⁵ Department of Internal Medicine, College of Medicine, Chung-Ang University, Seoul 06973, Republic of Korea

⁶ Department of Internal Medicine, College of Medicine, Daejeon St. Mary’s Hospital, The Catholic University of Republic of Korea, Daejeon 34943, Republic of Korea

⁷ Department of Internal Medicine, Dankook University College of Medicine,Cheonan 31116, Republic of Korea

⁸ Department of Internal Medicine, Haeundae Paik Hospital, Inje University College of Medicine, Busan 48108, Republic of Korea

⁹ Department of Internal Medicine, Digestive Disease Research Institute, Wonkwang University School of Medicine, Iksan 54538, Republic of Korea

¹⁰ Department of Internal Medicine, Kyung Hee University Hospital at Gang Dong, Kyung Hee University College of Medicine, Seoul 05278, Republic of Korea

¹¹ Division of Gastroenterology, Soonchunhyang University Cheonan Hospital, Cheonan 31151, Republic of Korea

¹² Department of Internal Medicine, Soonchunhyang University College of Medicine, Seoul 04401, Republic of Korea

¹³ Department of Internal Medicine and Liver Research Institute, College of Medicine, Seoul National University, Seoul 03080, Republic of Korea

¹⁴ Functional Genome Institute, PDXen Biosystems Inc., Yongin 17111, Republic of Korea;

¹⁵ Department of Gastroenterology, Center for Crohn’s and Colitis, Kyung Hee University Hospital, Kyung Hee University College of Medicine, Seoul 02453, Republic of Korea

^* Correspondence

^† These authors contributed equally to this work.

Abstract:

Crohn’s disease (CD) is a chronic inflammatory disorder with potential progression to stricturing (B2) or penetrating (B3) phenotypes, leading to significant complications. Early identification of patients at risk for these complications is critical for personalized management. This study aimed to develop a predictive model using clinical data and a Korean-specific transcriptome-wide association study (TWAS) to forecast early progression in CD patients. A retrospective analysis of 430 Korean CD patients from 15 hospitals was conducted. Genotyping was performed using the Korea Biobank Array, and gene expression predictions were derived from a TWAS model based on terminal ileum data. Logistic regression models incorporating clinical and gene expression data predicted progression to B2 or B3 within 24 months of diagnosis. Among the cohort, 13.9% (60 patients) progressed to B2 and 16.9% (73 patients) to B3. The combined model achieved mean area under the curve (AUC) values of 0.788 for B2 and 0.785 for B3 progression. Key predictive genes for B2 included CCDC154, FAM189A2, and TAS2R19, while PUS7, CCDC146, and MLXIP were linked to B3 progression. This integrative model provides a robust approach for identifying high-risk CD patients, potentially enabling early, targeted interventions to reduce disease progression and associated complications.

Keywords: Crohn’s disease; machine learning; early progression; structuring; penetrating

DOI: https://doi.org/10.3390/ijms26072910

​Published: 23 March 2025

붙임: Early Progression Prediction in Korean Crohn’s Disease Using a Korean-Specific PrediXcan Model.pdf