摘要
目的评价应用国产自主研发的基于CT血管造影(CTA)的血流储备分数(FFRCT)软件对心肌缺血的诊断价值。方法回顾性纳入2009年7月至2018年4月在北京大学第一医院行冠状动脉CTA及血流储备分数(FFR)检查的患者。使用FFRCT软件计算出FFRCT值,以有创的FFR值为金标准,分别从患者水平和血管水平计算出FFRCT对心肌缺血诊断的准确性、敏感度、特异度、阳性预测值和阴性预测值。分别在患者水平和血管水平统计FFRCT的ROC曲线下面积(AUC)。在血管水平,采用Pearson’s相关分析绘制FFRCT值与FFR值的相关散点图,并绘制Bland-Altman图检验FFRCT与FFR值的一致性。结果纳入患者71例,共计80支血管。患者水平FFRCT的准确性为81.7%,敏感度为71.4%,特异度为91.7%,阳性预测值为89.3%,阴性预测值为76.7%;血管水平FFRCT的准确性为81.3%,敏感度为67.6%,特异度为93.0%,阳性预测值为89.3%,阴性预测值为76.9%。在患者水平,FFRCT用于诊断心肌缺血的ROC AUC为0.852(95%CI 0.758~0.946,P<0.001),最佳诊断界值为0.805;在血管水平,FFRCT用于诊断心肌缺血的ROC AUC为0.848(95%CI 0.758~0.938,P<0.001),最佳诊断界值为0.805。Pearson’s相关分析显示,在血管水平,FFRCT值与FFR值呈正相关(r=0.542,95%CI 0.366~0.681,P<0.001)。FFRCT值与FFR值的BlandAltman图显示两者具有良好的一致性。结论以FFR值为金标准,应用国产FFRCT软件诊断心肌缺血有较高的准确率,具有一定的临床应用前景。
Objective To evaluate the diagnostic value of FFRCT software for myocardial ischemia. Methods Patients undergoing coronary CTA and FFR examination from July 2009 to April 2018 in Peking University First Hospital were enrolled retrospectively. FFRCT value was calculated by FFRCT software, and the diagnostic accuracy, sensitivity, specificity, positive predictive value and negative predictive value of FFRCT for myocardial ischemia were calculated per-patient level and per-vessel level with invasive FFR value as gold standard. The area under the ROC curve(AUC) of FFRCT was calculated per-patient level and per-vessel level respectively. At per-vessel level, Pearson’s correlation analysis was used and Bland-Altman plot was drawn to check the correlation and consistency of FFRCT and FFR value. Results A total 80 vessels in 71 patients were included in the analysis. The accuracy, sensitivity, specifi city, positive predictive value and negative predictive value of FFRCT perpatient level were 81.7%, 71.4%, 91.7%, 89.3% and 76.7%, respectively. The accuracy, sensitivity, specificity, positive predictive value and negative predictive value of FFRCT per-vessel level were 81.3%, 67.6%, 93.0%, 89.3% and 76.9%, respectively. At per-patient level, the area under the ROC curve of FFRCT for diagnosing myocardial ischemia was 0.852(95%CI 0.758 -0.946,P<0.001), and the optimal diagnostic threshold was 0.805. At per-vessel level, the area under the ROC curve of FFRCT for diagnosing myocardial ischemia was 0.848(95%CI 0.758 -0.938, P<0.001), and the optimal diagnostic threshold was 0.805. Pearson correlation analysis showed that per-vessel level, the FFRCT value was positively correlated with the FFR value(r=0.542, 95%CI 0.366-0.681, P<0.001). BlandAltman diagrams of FFRCT values and FFR values showed good consistency. Conclusions Using FFR as the gold standard, domestic FFRCT software can diagnose myocardial ischemia with high performance, and has a certain clinical application prospect.
作者
龚艳君
易铁慈
杨帆
邱建星
范芳芳
王梓鸣
王文烨
霍云龙
李建平
霍勇
GONG Yan-jun;YI Tie-ci;YANG Fan;QIU Jian-xing;FAN Fang-fang;WANG Zi-ming;WANG Wen-ye;HUO Yun-long;LI Jian-ping;HUO Yong(Department of Cardiology,Peking University First Hospital,Beijing 100034,China)
出处
《中国介入心脏病学杂志》
2019年第12期673-678,共6页
Chinese Journal of Interventional Cardiology
基金
首都卫生发展科研专项（2016-2-4073）
