摘要
目的探讨基于冠状动脉CT血管成像(CCTA)的心肌影像组学模型鉴别肥厚型心肌病(HCM)与高血压性心脏病(HHD)的价值。方法回顾性纳入2022年1月至2024年12月郑州大学第一附属医院临床诊断为HCM(109例)和HHD(62例)的患者171例。按7:3比例随机分为训练组(120例)和测试组(51例)。于CCTA舒张期图像勾画左心室心肌容积感兴趣区(VOI)并提取影像组学特征。采用最小绝对收缩和选择算子(LASSO)等方法进行特征选择后,构建Logistic回归模型并计算影像组学评分(Rad-score)。基于临床及超声心动图指标构建传统诊断模型,通过受试者工作特征曲线曲线下面积(AUC)和决策曲线分析(DCA)比较两者的效能,并分析Rad-score与左心室功能参数[室间隔厚度(IVSd)、左心室后壁厚度(LVPWd)、左心室射血分数(LVEF)、左心室舒张末期内径(LVEDD)]的相关性。结果最终筛选出5个影像组学特征构建模型。在训练组(AUC:0.86 vs.0.87,P>0.05)和测试组(AUC:0.75 vs.0.76,P>0.05)中,影像组学模型与传统模型的诊断效能差异均无统计学意义。Rad-score与左心室功能参数存在一定的相关性(P<0.05)。结论基于CCTA的影像组学模型鉴别HCM与HHD的效能与传统模型相当,可为常规冠状动脉评估提供基于心肌组织特征的新增诊断价值。
Objective To explore the value of a myocardial radiomics model based on coronary computed tomography angiography(CCTA)in distinguishing hypertrophic cardiomyopathy(HCM)from hypertensive heart disease(HHD).Methods A total of 171 patients clinically diagnosed with HCM(109 cases)and HHD(62 cases)at the First Affiliated Hospital of Zhengzhou University from January 2022 to December 2024 were retrospectively enrolled.They were randomly divided into a training group(120 cases)and a testing group(51 cases)at a ratio of 7:3.The volume of interest(VOI)of the left ventricular myocardial volume was delineated on CCTA end-diastolic images,and radiomics features were extracted.After feature selection using methods such as the least absolute shrinkage and selection operator(LASSO),a Logistic regression model was constructed and the radiomics score(Rad-score)was calculated.A traditional diagnostic model was built based on clinical and echocardiographic indicators.The diagnostic efficacy was compared using the area under the receiver operating characteristic curve(AUC)and decision curve analysis(DCA),and the correlation between Radscore and left ventricular function parameters[interventricular septal thickness(IVSd),left ventricular posterior wall thickness(LVPWd),left ventricular ejection fraction(LVEF),left ventricular end-diastolic diameter(LVEDD)]was analyzed.Results Finally,5 radiomics features were selected to construct the model.There were no statistically significant differences in diagnostic efficacy between the radiomics model and the traditional model in the training group(AUC:0.86 vs.0.87,P>0.05)and the testing group(AUC:0.75 vs.0.76,P>0.05).Rad-score had certain correlations with left ventricular function parameters(P<0.05).Conclusion The CCTA-based radiomics model has comparable efficacy to the traditional model in distinguishing HCM from HHD,and can provide additional diagnostic value based on myocardial tissue characteristics for routine coronary artery assessment.
作者
张笑蕾
刘佳玮
姚广
张永高
ZHANG Xiaolei;LIU Jiawei;YAO Guang(Department of Radiology,the First Affiliated Hospital of Zhengzhou University,Zhengzhou,Henan Province 450000,P.R.China)
出处
《临床放射学杂志》
北大核心
2026年第1期163-168,共6页
Journal of Clinical Radiology
基金
河南省重点研发专项基金资助项目(基金编号:231111313100)。
