关键词: 妇科肿瘤, 静脉栓塞, 风险预测模型, 系统评价

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