120017
论文已发表
注册即可获取德孚的最新动态
IF 收录期刊
- 3.6 Breast Cancer (Dove Med Press)
- 4.3 Clin Epidemiol
- 2.6 Cancer Manag Res
- 3.2 Infect Drug Resist
- 4.1 Clin Interv Aging
- 6.1 Drug Des Dev Ther
- 4.1 Int J Chronic Obstr
- 8.7 Int J Nanomed
- 2.5 Int J Women's Health
- 3.2 Neuropsych Dis Treat
- 2.4 OncoTargets Ther
- 2.6 Patient Prefer Adher
- 2.6 Ther Clin Risk Manag
- 3.1 J Pain Res
- 3.5 Diabet Metab Synd Ob
- 4.5 Psychol Res Behav Ma
- 3.4 Nat Sci Sleep
- 2.4 Pharmgenomics Pers Med
- 2.6 Risk Manag Healthc Policy
- 4.6 J Inflamm Res
- 2.3 Int J Gen Med
- 3.9 J Hepatocell Carcinoma
- 3.3 J Asthma Allergy
- 2.5 Clin Cosmet Investig Dermatol
- 3.0 J Multidiscip Healthc
已发表论文
用于预测上消化道出血大量输血的可解释机器学习模型的开发与验证
Authors Shang S , Lv H , Ma G, Wei M, Wang K , Lu Y, Tang B
Received 6 November 2025
Accepted for publication 16 March 2026
Published 23 March 2026 Volume 2026:19 574069
DOI https://doi.org/10.2147/IJGM.S574069
Checked for plagiarism Yes
Review by Single anonymous peer review
Peer reviewer comments 2
Editor who approved publication: Dr Redoy Ranjan
Zixi Lin,1,* Hailiang Zhao,2,* Yilong Hu3
1Department of Blood Transfusion, Affiliated Hospital of Xuzhou Medical University, Xuzhou, Jiangsu Province, People’s Republic of China; 2Department of Gastroenterology, Affiliated Hospital of Youjiang Medical University for Nationalities; Guangxi Medical and Health Key Cultivation Discipline Construction Project; Guangxi Clinical Medical Research Center for Hepatobiliary Disease, Baise, Guangxi Zhuang Autonomous Region, People’s Republic of China; 3Department of General Surgery of the International Medical Center; The Fourth Affiliated Hospital of Soochow University; Suzhou Dushu Lake Hospital, Suzhou, Jiangsu Province, People’s Republic of China
*These authors contributed equally to this work
Correspondence: Yilong Hu, Department of General Surgery of the International Medical Center; The Fourth Affiliated Hospital of Soochow University; Suzhou Dushu Lake Hospital, No. 9 Chongwen Road, Suzhou Industrial District, Suzhou, Jiangsu Province, People’s Republic of China, Email yiminhyl@163.com
Background: Upper gastrointestinal bleeding (UGIB) is a medical emergency with high mortality, especially when massive transfusion (MT) is required. Traditional scores like Glasgow-Blatchford provide moderate accuracy but overlook complex variable interactions. We developed and validated an explainable machine learning (ML) model for MT prediction in UGIB, improving precision and interpretability.
Methods: In this retrospective study, 700 UGIB patients from The Affiliated Hospital of Xuzhou Medical University (2021– 2025) were divided into training (n=490) and testing (n=210) cohorts. An external validation cohort (n=300) was sourced from The Fourth Affiliated Hospital of Soochow University. From 18 clinical variables, 8 key features were selected using Boruta and LASSO regression. Seven ML algorithms were compared to identify the optimal model, which was then evaluated for discrimination, calibration, and clinical utility. SHapley Additive exPlanations (SHAP) provided interpretability.
Results: The Random Forest (RF) model achieved superior performance with an AUC of 0.862 (95% CI 0.785– 0.939) in training, 0.823 (95% CI 0.768– 0.879) in testing, and 0.807 (95% CI 0.748– 0.866) in external validation. Calibration plots showed strong agreement between predicted and observed probabilities. Decision curve analysis indicated higher net benefit than “treat all” or “treat none” strategies. SHAP analysis ranked impaired mental status, liver cirrhosis, and international normalized ratio (INR) as top predictors, aligning with clinical intuition.
Conclusion: The developed machine learning model demonstrated promising performance in identifying UGIB patients at high risk of massive transfusion. While the model shows potential to assist clinicians in optimizing blood management strategies, further prospective validation is required to confirm its clinical utility in diverse settings.
Keywords: upper gastrointestinal bleeding, massive transfusion, machine learning, SHAP, random forest
1Department of Blood Transfusion, Affiliated Hospital of Xuzhou Medical University, Xuzhou, Jiangsu Province, People’s Republic of China; 2Department of Gastroenterology, Affiliated Hospital of Youjiang Medical University for Nationalities; Guangxi Medical and Health Key Cultivation Discipline Construction Project; Guangxi Clinical Medical Research Center for Hepatobiliary Disease, Baise, Guangxi Zhuang Autonomous Region, People’s Republic of China; 3Department of General Surgery of the International Medical Center; The Fourth Affiliated Hospital of Soochow University; Suzhou Dushu Lake Hospital, Suzhou, Jiangsu Province, People’s Republic of China
*These authors contributed equally to this work
Correspondence: Yilong Hu, Department of General Surgery of the International Medical Center; The Fourth Affiliated Hospital of Soochow University; Suzhou Dushu Lake Hospital, No. 9 Chongwen Road, Suzhou Industrial District, Suzhou, Jiangsu Province, People’s Republic of China, Email yiminhyl@163.com
Background: Upper gastrointestinal bleeding (UGIB) is a medical emergency with high mortality, especially when massive transfusion (MT) is required. Traditional scores like Glasgow-Blatchford provide moderate accuracy but overlook complex variable interactions. We developed and validated an explainable machine learning (ML) model for MT prediction in UGIB, improving precision and interpretability.
Methods: In this retrospective study, 700 UGIB patients from The Affiliated Hospital of Xuzhou Medical University (2021– 2025) were divided into training (n=490) and testing (n=210) cohorts. An external validation cohort (n=300) was sourced from The Fourth Affiliated Hospital of Soochow University. From 18 clinical variables, 8 key features were selected using Boruta and LASSO regression. Seven ML algorithms were compared to identify the optimal model, which was then evaluated for discrimination, calibration, and clinical utility. SHapley Additive exPlanations (SHAP) provided interpretability.
Results: The Random Forest (RF) model achieved superior performance with an AUC of 0.862 (95% CI 0.785– 0.939) in training, 0.823 (95% CI 0.768– 0.879) in testing, and 0.807 (95% CI 0.748– 0.866) in external validation. Calibration plots showed strong agreement between predicted and observed probabilities. Decision curve analysis indicated higher net benefit than “treat all” or “treat none” strategies. SHAP analysis ranked impaired mental status, liver cirrhosis, and international normalized ratio (INR) as top predictors, aligning with clinical intuition.
Conclusion: The developed machine learning model demonstrated promising performance in identifying UGIB patients at high risk of massive transfusion. While the model shows potential to assist clinicians in optimizing blood management strategies, further prospective validation is required to confirm its clinical utility in diverse settings.
Keywords: upper gastrointestinal bleeding, massive transfusion, machine learning, SHAP, random forest