■本研究的目的是探讨导管溶栓(CDT)治疗后血栓形成综合征(PTS)对急性下肢深静脉血栓形成(DVT)的危险因素。
■我们回顾性选择了171例接受CDT治疗的急性下肢DVT患者,收集患者的临床数据,根据治疗后1年的随访结果进行分组,将PTS患者纳入并发组,未发生PTS的患者纳入非并发组.应用单因素分析和Logistic回归分析急性下肢DVT置管溶栓治疗后发生PTS的危险因素。我们应用R4.2.3软件构建了三个混合机器学习模型,包括一个列线图,决策树,和以独立影响因素为预测变量的随机森林。
■急性下肢DVT中CDT后PTS的发生率为36.84%。BMI>24.33kg/m2,发病时间>7d,混合性DVT,静脉曲张病史,应激治疗时间>6.5个月,和过滤器类别是CDT治疗急性下肢DVT后PTS的独立危险因素。随机森林模型预测的AUC值高于列线图模型(Z=-2.337,P=0.019)和决策树模型(Z=-2.995,P=0.003)。
■CDT治疗急性下肢DVT后PTS的发生与多种因素密切相关,建立的随机森林模型对PTS并发PTS的预测效果最好。
UNASSIGNED: The purpose of the research is to explore post-thrombotic syndrome (PTS) after catheter-directed thrombolysis (CDT) treatment for acute lower extremity deep vein thrombosis (DVT) risk factors.
UNASSIGNED: We retrospectively selected 171 patients with acute lower extremity DVT undergoing CDT treatment, collected clinical data of the patients, grouped them according to the follow-up results of 1 year after treatment, and included patients with PTS into the concurrent group and patients who did not develop PTS assigned to the unconcurrent group. Univariate analysis and Logistic regression were applied to analyze the risk factors of PTS after catheterization and thrombolytic therapy for acute lower extremity DVT. We applied R4.2.3 software to build three hybrid machine-learning models, including a nomogram, decision tree, and random forest with independent influencing factors as predictive variables.
UNASSIGNED: The incidence of PTS after CDT in acute lower extremity DVT was 36.84 %. BMI >24.33 kg/m2, disease time >7 d, mixed DVT, varicose vein history, stress treatment time>6.5 months, and filter category were independent risk factors for PTS after CDT treatment for acute lower extremity DVT. The AUC value predicted by the random forest model was higher than that of the nomogram model (Z = -2.337, P = 0.019) and the decision tree model (Z = -2.995, P = 0.003).
UNASSIGNED: The occurrence of PTS after CDT treatment of acute lower extremity DVT is closely related to many factors, and the established random forest model had the best effect in predicting PTS complicated with PTS.