Research advances in the application of artificial intelligence in transfusion medicine
10.13303/j.cjbt.issn.1004-549x.2025.11.005
- VernacularTitle:人工智能在输血医学中的应用研究进展
- Author:
Xinxin YANG
1
;
Shilan XU
1
;
Bing HAN
1
;
Lixin WANG
1
;
Fu CHENG
1
;
Dongmei YANG
1
;
Bin TAN
1
;
Li QIN
1
;
Chunxia CHEN
1
Author Information
1. Department of Transfusion Medicine, West China Hospital, Sichuan University, Chengdu 610041, China
- Publication Type:Journal Article
- Keywords:
artificial intelligence;
machine learning;
transfusion/reaction;
blood banks;
prediction model
- From:
Chinese Journal of Blood Transfusion
2025;38(11):1502-1513
- CountryChina
- Language:Chinese
-
Abstract:
Objective: To review the current development of artificial intelligence (AI) technology in the field of transfusion medicine. Methods: A systematic search was conducted in the Clarivate Web of Science Database from inception to December 2024 for literature related to AI and transfusion. A total of 4 775 publications were identified. Based on inclusion and exclusion criteria, 133 original studies were ultimately included and analyzed using a narrative synthesis approach. Results: Research on AI in transfusion has surged since 2020 (accounting for 77% of all publications), with China ranking second globally in publication volume. Among the included studies, 69.2% focused on predicting individual transfusion needs, followed by inventory management (8.3%), diagnosis and prediction of adverse transfusion reactions (6.0%), factors influencing transfusion outcomes (5.3%), blood group identification (5.3%), blood quality testing (4.5%), and precise blood volume measurement (1.5%). Additionally, 4.5% of the studies were published in journals with an impact factor greater than 10; 19.5% developed software or applications; 31.5% were multi-center studies; 48.1% utilized decision tree methods, while 31.5% employed neural network approaches; and 14.2% conducted external validation of the algorithms. Conclusion: AI demonstrates significant potential in transfusion risk prediction, decision support, and blood management. However, challenges remain, including limited model generalizability, insufficient algorithm interpretability, and barriers to clinical translation. The deep integration of AI with transfusion medicine will accelerate the advent of precision transfusion era, maximizing blood resource utilization, reducing waste, and ensuring transfusion safety.
