Wang RR, Zhu M, Ren HC, Yu WL. Machine learning models for predicting acute kidney injury after pediatric living donor liver transplantation in biliary atresia. World J Gastrointest Surg 2026; 18(7): 120759 [DOI: 10.4240/wjgs.v18.i7.120759]
Corresponding Author of This Article
Wen-Li Yu, PhD, Department of Anesthesiology, Tianjin First Central Hospital, Nankai University, No. 24 Fukang Road, Tianjin 300192, China. yzxyuwenli@163.com
Research Domain of This Article
Surgery
Article-Type of This Article
research-article
Open-Access Policy of This Article
This article is an open-access article which was selected by an in-house editor and fully peer-reviewed by external reviewers. It is distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/4.0/
Baishideng Publishing Group Inc, 7041 Koll Center Parkway, Suite 160, Pleasanton, CA 94566, USA
Share the Article
Wang RR, Zhu M, Ren HC, Yu WL. Machine learning models for predicting acute kidney injury after pediatric living donor liver transplantation in biliary atresia. World J Gastrointest Surg 2026; 18(7): 120759 [DOI: 10.4240/wjgs.v18.i7.120759]
World J Gastrointest Surg. Jul 27, 2026; 18(7): 120759 Published online Jul 27, 2026. doi: 10.4240/wjgs.v18.i7.120759
Machine learning models for predicting acute kidney injury after pediatric living donor liver transplantation in biliary atresia
Rong-Rong Wang, Min Zhu, Heng-Chang Ren, Wen-Li Yu
Rong-Rong Wang, School of Medicine, Nankai University, Tianjin 300071, China
Min Zhu, Heng-Chang Ren, Wen-Li Yu, Department of Anesthesiology, Tianjin First Central Hospital, Nankai University, Tianjin 300192, China
Author contributions: Wang RR and Yu WL conceived the manuscript; Yu WL coordinated and supervised data collection; Zhu M carried out the statistical analysis; Ren HC wrote and prepared the tables and figures; Zhu M and Wang RR revised the manuscript with additional detail; Ren HC and Yu WL critically reviewed the manuscript for important intellectual content; Yu WL is the study’s guarantor. All authors have read and agreed to the published version of the manuscript.
Supported by Tianjin Key Clinical Specialty Construction Project, Tianjin Key Medical Discipline Construction Project, No. TJYXZDXK-3-022C; and Scientific Research Program of the Tianjin Municipal Education Commission, No. 2025ZD40.
Institutional review board statement: The study protocol was compliant with the principles of the Declaration of Helsinki and was approved by the Institutional Review Board and Ethics Committee of Tianjin First Central Hospital, No. KYAP2025-170.
Informed consent statement: The requirement for informed consent was waived by the Ethics Committee due to the retrospective nature of the study.
Conflict-of-interest statement: The authors declare that they have no competing interests.
Data sharing statement: No additional data are available.
Corresponding author: Wen-Li Yu, PhD, Department of Anesthesiology, Tianjin First Central Hospital, Nankai University, No. 24 Fukang Road, Tianjin 300192, China. yzxyuwenli@163.com
Received: March 13, 2026 Revised: April 14, 2026 Accepted: May 8, 2026 Published online: July 27, 2026 Processing time: 142 Days and 18.2 Hours
Abstract
BACKGROUND
Living donor liver transplantation (LDLT) is an important treatment method for end-stage pediatric liver diseases, e.g., biliary atresia (BA). Acute kidney injury (AKI) after transplantation is a common and serious complication in clinical practice that significantly influences patient mortality and survival rate.
AIM
To construct a clinical prediction model for AKI after pediatric LDLT based on machine learning (ML).
METHODS
This study included 340 children with BA who underwent LDLT at our center between December 2022 and December 2024. Complete clinical data were collected, including baseline characteristics, preoperative assessments, intraoperative variables, and postoperative recovery indicators. Least absolute shrinkage and selection operator regression was used for feature selection, and nine ML models were developed for model training and evaluation. After training on the training set, the predictive performance of each model was tested and compared. Finally, the best-performing model was interpreted and visualized using the SHapley Additive exPanations (SHAP) algorithm.
RESULTS
Excluding postoperative creatinine (Cr) levels, this study identified a total of six potential predictors associated with AKI after LDLT. The random forest model showed comprehensive and optimal predictive performance after 10-fold cross-validation, with an area under the curve of 0.875 (95% confidence interval: 0.805-0.944). In addition, the importance of predictors for AKI occurrence was ranked by SHAP analysis, and preoperative Cr level was identified as the most important predictor.
CONCLUSION
This study employed ML algorithms to construct a predictive model for early AKI following pediatric liver transplantation. The developed model is expected to assist doctors in performing timely treatment interventions, thereby reducing the occurrence of post-transplant complications and improve the survival time and quality of life in children undergoing liver transplantation.
Core Tip: This study included 340 children with biliary atresia who underwent liver transplantation. Seven key predictors of acute kidney injury (AKI) were screened out by least absolute shrinkage and selection operator algorithm, including pre-operative/post-operative creatinine (Cr), blood calcium and lactic acid levels during the anhepatic phase, gender, the amount of fresh frozen plasma infused during the operation, and post-operative aspartate aminotransferase level. Nine machine learning methods, including XGBoost, were used to construct the postoperative AKI prediction model based on other features after excluding postoperative Cr, and their prediction performance was compared. This study aims to assist clinicians in early intervention and improve the prognosis of children.
