^(99m)Tc-MIBI心肌灌注显像联合临床特征的诊断模型可有效预测冠心病  

^(99m)Tc-MIBI myocardial imaging combined with clinical features can effectively predict coronary artery disease

作　　者：陈媛 李东 罗晓琴 凌苑娜 欧阳伟[2] CHEN Yuan;LI Dong;LUO Xiaoqin;LING Yuanna;OUYANG Wei(Department of Radiology,The Affiliated Hospital of Southwest Medical University,Luzhou 646000,China;Department of Nuclear Medicine,Zhujiang Hospital of Southern Medical University,Guangzhou 510280,China)

机构地区：[1]西南医科大学附属医院放射科,四川泸州646000 [2]南方医科大学珠江医院核医学科,广东广州510280

出　　处：《分子影像学杂志》2025年第2期145-151,共7页Journal of Molecular Imaging

基　　金：国家自然科学基金(82071955)。

摘　　要：目的构建基于^(99m)Tc-甲氧基异丁基异腈(MIBI)门控心肌灌注显像(GMPI)及临床特征的冠状动脉粥样硬化性心脏病(冠心病)诊断预测模型并进行内部验证,用于诊断预测冠心病的患病风险。方法回顾性收集2023年1月~2023年11月于南方医科大学珠江医院行^(99m)Tc-MIBI SPECT/CT门控心肌静息灌注显像检查的116例疑诊为冠心病患者的GMPI参数和临床特征,其中男性77例、女性39例,年龄23~93(62.66±12.22)岁。通过Stepwise回归及多因素Logistic回归分析筛选冠心病的预测因子并构建诊断预测模型,以列线图形式展示;通过曲线下面积(AUC)评估模型的预测效能。采用k折交叉验证法进行内部验证,通过校准曲线、决策曲线分析和临床影响曲线评估模型的校准度和临床应用价值。结果Stepwise回归显示,GMPI参数中的左室舒张末期容积、峰值充盈率、直方图偏度和直方图峰态为冠心病的有效诊断预测因子,联合临床特征(性别、吸烟史、心肌肌钙蛋白、高血压)后构建预测模型,其AUC为0.731(95%CI:0.636~0.825),特异度为0.735,敏感度为0.642。k折交叉验证法的平均AUC为0.699。校准曲线显示冠心病诊断预测模型的校准度良好,决策曲线分析和临床影响曲线表明该模型具有较高的临床应用价值。结论基于^(99m)Tc-MIBI GMPI参数及临床特征(性别、吸烟史、心肌肌钙蛋白和高血压)诊断预测模型能较好地对冠心病患者进行评估,为开发更个性化的冠心病诊断策略提供可能性。Objective To develop a diagnostic prediction model for coronary artery disease(CAD)based on^(99m)Tcmethoxyisobutylisonitrile(MIBI)gated myocardial perfusion imaging(GMPI)and clinical features,and to perform internal validation to assess its utility in predicting the risk of CAD.Methods A retrospective analysis was conducted to collect GMPI parameters and clinical characteristics of 116 patients suspected of having CAD who underwent^(99m)Tc-MIBI SPECT/CT gated myocardial resting perfusion imaging at Zhujiang Hospital of Southern Medical University from January 2023 to November 2023.Among the patients,77 were male and 39 were female,with an age range of 23-93(62.66±12.22)years old.Predictive factors for CAD were identified using stepwise regression and multivariate logistic regression analysis,and a diagnostic prediction model was constructed and presented in the form of a nomogram.The predictive performance of the model was evaluated by calculating the area under the ROC curve(AUC).Internal validation was performed using k-fold crossvalidation.The calibration and clinical utility of the model were assessed through calibration curves,decision curve analysis(DCA),and clinical impact curves.Results Stepwise regression analysis identified left ventricular end-diastolic volume,peak filling rate,histogram skewness,and histogram kurtosis among the GMPI parameters as effective diagnostic predictors of CAD.Incorporating clinical characteristics(gender,smoking history,cardiac troponin,hypertension),a predictive model was constructed with an AUC of 0.731(95%CI:0.636-0.825),specificity of 0.735,and sensitivity of 0.642.The average AUC from kfold cross-validation was 0.699.Calibration curves demonstrated good calibration of the CAD diagnostic prediction model,while decision curve analysis and clinical impact curves indicated its high clinical utility.Conclusion The diagnostic prediction model based on^(99m)Tc-MIBI GMPI parameters and clinical characteristics(gender,smoker,cardiac troponin,hypertension)demonstrates good perfor

关 键 词：心肌灌注显像 冠心病 预测模型 列线图 

分 类 号：R541.4[医药卫生—心血管疾病] R817.4[医药卫生—内科学]