Objective To systematically analyze the current application status of artificial intelligence (AI) in risk assessment and management of venous thromboembolism (VTE), evaluate the predictive performance of AI models and identify key risk factors, thereby providing evidence-based references for optimizing clinical VTE prevention and treatment strategies. Methods A scoping review framework was used. We searched for literature in both Chinese (CNKI, Wanfang, CBM) and English databases (PubMed, Web of Science, Embase, CINAHL, and The Cochrane Library) to find studies on AI applications in VTE risk assessment, covering the time from when the databases started until 10 March 2025. By creating research questions, reviewing the literature, gathering data, and summarizing the results, we organized various AI models, assessed how accurately they predicted outcomes, and looked at important risk factors. Results This review included a total of 23 studies. AI models showed better accuracy in predicting VTE risk, with AUC values between 0.740 and 0.990, greatly surpassing traditional scoring tools. Key risk factors identified included patient-related factors, disease-related factors, treatment-related factors, laboratory indicators, and catheter-related factors. Conclusion AI technology shows remarkable advantages in VTE risk assessment by integrating multi-source data to achieve dynamic and personalized prediction. Future research should aim to conduct studies across multiple centers to confirm how useful these models are in real-life situations and also look into combining real-time monitoring data with AI to enhance the accuracy of preventing and treating VTE, which will help lower the number of cases and improve patient results.
Gu et al. (Wed,) studied this question.