【摘要】 目的 探讨急性前壁ST段抬高型心肌梗死(STEMI)患者左心室血栓形成的预测因素并构建决策树预测模型。方法 回顾性分析沧州市中心医院自2014年10月至2024年10月收治的411例急性前壁STEMI患者的临床资料。将并发左心室血栓的137例患者纳入血栓组；将根据年龄(年龄±2岁)和性别匹配的无左心室血栓的274例患者纳入非血栓组。比较两组患者的一般资料。根据7∶3比例将患者分为训练集(n=286)和验证集(n=125),训练集用于模型构建，验证集用于模型验证。采用logistic回归分析筛选急性前壁STEMI患者左心室血栓形成的预测因素并构建决策树预测模型。采用受试者工作特征曲线计算曲线下面积(AUC)评估训练集和验证集决策树模型的预测价值。结果 两组患者肌酸激酶同工酶、肌钙蛋白、左心室射血分数(LVEF)水平及室壁瘤、早期罪犯血管再通失败概率比较，差异有统计学意义(P<0.05)。LVEF、室壁瘤、早期罪犯血管再通失败为急性前壁STEMI患者左心室血栓形成的独立预测因素(P<0.05)。基于训练集数据，以急性前壁STEMI左心室血栓为因变量，LVEF、室壁瘤、早期罪犯血管再通失败为自变量构建决策树模型。决策树模型共形成4条决策分类：(1)LVEF≥47%时，患者左心室血栓形成的风险为2.2%;(2)LVEF<47%、早期罪犯血管再通失败时，患者左心室血栓形成的风险为68.1%;(3)LVEF<47%、早期罪犯血管再通成功、无室壁瘤时，患者左心室血栓形成的风险为26.0%;(4)LVEF<47%、早期罪犯血管再通成功、有室壁瘤时，患者左心室血栓形成的风险为85.7%。训练集中，决策树模型预测急性前壁STEMI患者左心室血栓形成的AUC为0.841(95%可信区间0.798～0.883),敏感度为71.6%,特异度为84.3%;验证集中，决策树模型预测急性前壁STEMI患者左心室血栓形成的AUC为0.804(95%可信区间0.732～0.877),敏感度为57.1%,特异度为86.7%。结论 LVEF、室壁瘤、早期罪犯血管再通失败为急性前壁STEMI患者左心室血栓形成的独立预测因素，以此构建的决策树模型可有效预测急性前壁STEMI患者左心室血栓形成的风险。更多还原

【Abstract】 Objective To explore the predictive factors of left ventricular thrombosis in patients with acute anterior wall ST elevation myocardial infarction(STEMI) and construct a decision tree prediction model.Methods A retrospective analysis was conducted on the clinical data of 411 patients with acute anterior wall STEMI admitted to Cangzhou Central Hospital from October 2014 to October 2024.A total of 137 patients with concurrent left ventricular thrombosis were included in the thrombosis group.A total of 274 patients without left ventricular thrombosis who were matched according to age and gender were included in the non-thrombosis group.Compared the general information of the two groups of patients.The patients were divided into the training set(n=286) and the validation set(n=125) according to a 7∶3 ratio.The training set was used for model construction and the validation set was used for model validation.Logistic regression analysis was used to screen the predictive factors of left ventricular thrombosis in patients with acute anterior wall STEMI and to construct a decision tree prediction model.The area under the curve(AUC) was calculated using the receiver operator characteristic curve to evaluate the predictive value of the decision tree models in the training set and the validation set.Results There were statistically significant differences in the levels of creatine kinase isoenzyme, troponin, left ventricular ejection fraction(LVEF),and the probabilities of ventricular aneurysm and early culprit vessel recanalization failure between the two groups of patients(P<0.05).LVEF,ventricular aneurysm, and failure of early culprit vessel recanalization were independent predictors of left ventricular thrombosis in patients with acute anterior wall STEMI(P<0.05).Based on the training set data, a decision tree model was constructed with acute anterior wall STEMI left ventricular thrombosis as the dependent variable, while LVEF,ventricular aneurysm and failure of early culprit vascular recanalization as independent variables.The decision tree model formed a total of 4 decision classifications:(1)When LVEF≥47%,the risk of left ventricular thrombosis in patients was 2.2%.(2)When LVEF<47% and the recanalization of the early culprit vessel failed, the risk of left ventricular thrombosis in patients was 68.1%.(3)When LVEF<47%,the early culprit vessel recanalization was successful, and there was no ventricular aneurysm, the risk of left ventricular thrombosis in patients was 26.0%.(4)When LVEF<47%,the early culprit vessel recanalization was successful, and there was a ventricular aneurysm, the risk of left ventricular thrombosis in patients was 85.7%.In the training set, the AUC of the decision tree model for predicting left ventricular thrombosis in patients with acute anterior wall STEMI was 0.841(95% confidence interval 0.798-0.883),with a sensitivity of 71.6% and a specificity of 84.3%.In the validation set, the AUC of the decision tree model for predicting left ventricular thrombosis in patients with acute anterior wall STEMI was 0.804(95% confidence interval 0.732-0.877),with a sensitivity of 57.1% and a specificity of 86.7%.Conclusion LVEF,ventricular aneurysm, early failure of recanalization of the offender vessel are independent predictors of left ventricular thrombosis in patients with acute anterior wall STEMI,and the decision tree model constructed based on these can effectively predict the risk of left ventricular thrombosis in patients with acute anterior wall STEMI.更多还原