Abstract: Objective: To systematically evaluate the prediction models for venous embolism in gynecologic cancer patients. Methods: PubMed, Web of Science, Embase, Cochrane Library, CNKI, Wanfang, and VIP databases were searched to collect studies on predictive models of venous embolism in gynecological cancer patients with a search time from the inception of the database to December 2024. Literatures was independently screened by two investigators, a standardized form was developed to extract information using the Data Extraction Checklist for the Evaluation of Predictive Modeling Systems (CHARMS), and the Risk of Bias and Suitability Assessment Tool for Predictive Modeling Studies (PROBAST) was used to evaluate the risk of bias and suitability for the included literatures. Results: A total of 14 papers were included and 17 risk prediction models were constructed, the AUC ranged from 0.563 to 0.929; containing 3 to 7 predictors. The meta-analysis results indicated that the pooled AUC value of the predictive model was 0.83 [95%CI (0.78, 0.88)]. Age[OR=1.12, 95%CI (1.05, 1.19)], body mass index[OR=3.20, 95%CI (1.45, 7.10)], plasma D-dimer level[OR=3.26, 95%CI (1.88,5.63)], the revised surgical pathological stage of the International Federation of Gynecology and Obstetrics[OR=1.90, 95%CI(1.01, 3.55)], and operation time[OR=1.48, 95%CI (1.05, 2.08)] were predictive factors for venous embolism in gynecological tumor patients(P<0.001). Conclusion: The overall predictive performance of the risk prediction models for venous embolism in gynecological cancer patients is relatively good, but the model quality needs to be improved. There is room for optimization in aspects such as data sources, model construction, and validation analysis.

Key words: Gynecological cancer, Venous embolism, Risk prediction model, Systematic review

摘要： 目的：系统评价妇科肿瘤患者静脉栓塞风险预测模型。方法：检索 PubMed、Web of Science、Embase、Cochrane Library、中国知 网、万方和维普数据库中关于妇科肿瘤患者静脉栓塞风险预测模型的研究，检索时限为建库至2024年12月。由2名研究者独立筛选文献，使用预测模型系统评价数据提取清单提取资料，采用预测模型研究偏倚风险与适用性评估工具评价纳入文献的偏倚风险和适用性。结果：共纳入14篇文献构建17个风险预测模型，曲线下面积为 0.563~0.929，包含3~7个预测因子。Meta分析结果显示，预测模型的合并曲线下面积为 0.83[95%CI（0.78，0.88）]。年龄[OR=1.12，95%CI（1.05，1.19）]、身体质量指数[OR=3.20，95%CI（1.45，7.10）]、血浆 D-二聚体水平[OR=3.26，95%CI（1.88，5.63）]、国际妇产科联盟分期修订的手术病理分期[OR =1.90，95%CI（1.01，3.55）]、手术时间[OR=1.48，95%CI（1.05，2.08）]是妇科肿瘤患者静脉栓塞的预测因子（P＜0.001）。结论：妇科肿瘤患者静脉栓塞风险预测模型整体预测性能较好，但模型质量有待提升，在数据来源、模型构建和验证分析等方面均有待优化。

关键词: 妇科肿瘤, 静脉栓塞, 风险预测模型, 系统评价