Dalil Davood, Esmaeili Sina, Safaee Ehsan, Asgari Sajad, Kejani Nooshin
Faculty of Medicine, Shahed University, Tehran, Iran.
Student Research Committee, Faculty of Medicine, Shahed University, Tehran, Iran.
Arthroplast Today. 2025 Mar 29;33:101672. doi: 10.1016/j.artd.2025.101672. eCollection 2025 Jun.
Venous thromboembolism (VTE), including deep vein thrombosis and pulmonary embolism, is a common and serious complication following lower extremity arthroplasty, such as total hip and knee arthroplasty. Due to the increasing number of these surgeries, accurately predicting VTE risk is crucial. Traditional clinical prediction models often fall short due to their complexity and limited accuracy.
This Preferred Reporting Items for Systematic Review and Meta-Analyses-guided systematic review summarized the application of artificial intelligence (AI) and machine learning models in predicting VTE after total joint arthroplasty. Databases including PubMed, Scopus, Web of Science, and Embase were searched for relevant studies published up to January 2024. Eligible studies focused on the predictive accuracy of AI algorithms for VTE post arthroplasty and were assessed for quality using the Newcastle-Ottawa Scale.
A total of 7 retrospective cohort studies, encompassing 579,454 patients, met the inclusion criteria. These studies primarily employed the extreme gradient boosting model, which generally demonstrated strong predictive performance with area under the curve values ranging from 0.71 to 0.982. Models like random forest and support vector machines also performed well. However, only 1 study included external validation, critical for assessing generalizability.
AI and machine learning models, particularly extreme gradient boosting, exhibit significant potential in predicting VTE after lower extremity arthroplasty, outperforming traditional clinical prediction tools. Yet, the need for external validation and high-quality, generalizable datasets remains critical before these models can be widely implemented in clinical practice. The study underscores the role of AI in preoperative planning to enhance patient outcomes in orthopaedic surgery.
静脉血栓栓塞症（VTE），包括深静脉血栓形成和肺栓塞，是下肢关节置换术（如全髋关节和膝关节置换术）后常见且严重的并发症。由于此类手术数量不断增加，准确预测VTE风险至关重要。传统的临床预测模型往往因其复杂性和有限的准确性而不足。
本项循证医学报告系统评价和Meta分析优先报告条目指导的系统评价总结了人工智能（AI）和机器学习模型在预测全关节置换术后VTE中的应用。检索了包括PubMed、Scopus、Web of Science和Embase在内的数据库，以查找截至2024年1月发表的相关研究。纳入的研究聚焦于AI算法对关节置换术后VTE的预测准确性，并使用纽卡斯尔-渥太华量表评估质量。
共有7项回顾性队列研究符合纳入标准，涉及579,454例患者。这些研究主要采用极端梯度提升模型，其通常表现出强大的预测性能，曲线下面积值范围为0.71至0.982。随机森林和支持向量机等模型也表现良好。然而，只有1项研究纳入了外部验证，这对评估模型的通用性至关重要。
AI和机器学习模型，尤其是极端梯度提升模型，在预测下肢关节置换术后VTE方面具有显著潜力，优于传统临床预测工具。然而，在这些模型能够在临床实践中广泛应用之前，外部验证以及高质量、可推广的数据集的需求仍然至关重要。该研究强调了AI在术前规划中对改善骨科手术患者预后的作用。
