左心耳CT形态及功能改变预测心房颤动患者血栓形成  

Predicting thrombosis formation in patients with atrial fibrillation based on CT morphological and functional changes of the left atrial appendage

作　　者：陈梦婷 余太慧[2] 赵君雅 周旭辉[1] 魏帝 CHEN Mengting;YU Taihui;ZHAO Junya;ZHOU Xuhui;WEI Di(Department of Radiology,The Eighth Affiliated Hospital,Sun Yat-Sen University,Guangdong 518033,China)

机构地区：[1]中山大学附属第八医院放射科,广东深圳518033 [2]中山大学孙逸仙纪念医院放射科,广东广州510120 [3]广州迈普再生医学科技股份有限公司,广东广州510663

出　　处：《影像诊断与介入放射学》2025年第1期36-43,共8页Diagnostic Imaging & Interventional Radiology

摘　　要：目的探讨心脏CT评估左心耳(LAA)形态和功能特征预测非瓣膜性心房颤动(NVAF)患者左心房血栓(LAAT)和自发超声显影(SEC)的风险。方法本回顾性研究共纳入129例NVAF患者,连续纳入104例患者作为训练集,后续纳入25例患者作为验证集。采用二元Logistic回归分析与LAAT/SEC相关的独立风险因素并建立LAAT/SEC综合预测模型。评价模型的预测效能并绘制综合预测模型的列线图,完成模型的校准度和临床实用性评估。结果在训练集中,32/104(30.8%)例患者存在LAAT/SEC;在验证集中,5/25(20.0%)例患者存在LAAT/SEC。二元Logistic回归分析筛选出预测LAAT/SEC的4个独立风险因素,包括心房颤动类型、LAA危险度分型(LAA L/H分型)、LAA最大和最小体积(LAA Vmax、LAA Vmin)。基于上述参数建立的LAAT/SEC综合预测模型在训练集中的AUC值为0.898(95%CI:0.832~0.965),准确度为86%,DeLong检验显示其AUC值显著高于其他单一模型(P<0.05)。在验证集中,该模型的AUC值为0.850(95%CI:0.674~1.000),准确度88%。H-L检验显示该模型具有良好的校准度,DCA表明该模型有较好的临床获益。结论基于4个独立风险因素建立的综合模型有助于提高NVAF患者LAAT/SEC风险预测的准确性。Objective To explore the geometrical and functional features of left arial appendage(LAA)assessed by cardiac CT in predicting the risk of LAA thrombosis(LAAT)and spontaneous echocardiographic contrast(SEC)in non-valvular atrial fibrillation(NVAF)patients.Methods This retrospective study enrolled 129 NVAF patients,with 104 patients enrolled designated as the training set and 25 patients enrolled as the validation set.Binary logistic regression was used to identify risk factors independently associated with LAAT/SEC and establish an integrative prediction model.The model’s predictive performance was evaluated.A nomogram of the prediction model was developed.The model’s calibration was assessed and its clinical utility was evaluated.Results In the training cohort,32/104(30.8%)patients exhibited LAAT/SEC,and in the validation cohort,5/20(20.0%)patients presented LAAT/SEC.Binary logistic regression identified four independent risk factors,including atrial fibrillation type,LAA low risk morphology/high risk morphology type,LAA maximal volume and LAA minimal volume.The integrative prediction model based on these factors demonstrated an AUC of 0.898(95%CI:0.832-0.965)with 86%accuracy in the training set,significantly outperforming single-factor models(all P<0.05).In the validation set,the model achieved an AUC of 0.850(95%CI:0.674-1.000)and 88%accuracy.The H-L test indicated good calibration,and decision curve analysis demonstrated favorable clinical utility.Conclusion Integrative prediction model based on four independent risk factors enhances the accuracy of LAAT/SEC risk prediction in patients with NVAF.

关 键 词：心脏计算机断层扫描 左心房 血栓 自发超声显影 非瓣膜性心房颤动 

分 类 号：R814.42[医药卫生—影像医学与核医学] R541.75[医药卫生—放射医学]