摘要
目的以冠状动脉造影(CAG)为金标准,探讨人工智能(AI)冠状动脉CT血管成像(CCTA-AI)联合基于冠状动脉CT的血流储备分数(FFR-CT)诊断冠状动脉狭窄病变的价值。方法选取先后行冠状动脉CT血管成像(CCTA)及冠状动脉造影(CAG)检查的90例患者的资料,两项检查时间间隔≤2周。利用AI算法获得CCTA病变直径的狭窄程度,并采用科亚医疗深脉软件计算FFR-CT数值。以CAG为金标准,血管狭窄≥70%为重度狭窄,绘制CCTA-AI、FFR-CT以及联合两种AI软件的受试者操作特征曲线,获得曲线下面积(AUC),并计算CCTA-AI联合FFR-CT的敏感度、特异度。结果将同时获得CCTA-AI及FFR-CT结果的90例患者的266支血管纳入分析。CCTA-AI、FFR-CT及两者联合诊断血管狭窄的AUC分别为0.817、0.850、0.883。CCTA-AI联合FFR-CT的诊断敏感度、特异度分别为82.09%和81.08%。结论CCTA-AI联合FFR-CT对冠状动脉狭窄的诊断效能得到提高。
Objective To investigate the effect of artificial intelligence based coronary computed tomography angiography(CCTA-AI)combined with the fractional flow reserve based on coronary CT(FFR-CT)on the diagnosis of coronary stenosis,with reference to the coronary angiography(CAG)results.Methods Ninety patients were retrospectively enrolled in this study,who underwent both coronary computed tomography angiography(CCTA)and the CAG within 2 weeks.The stenosis degree of pathological vessels was obtained by CCTA-AI algorithm,and the FFR-CT value was calculated by Keya Medical software based on CCTA images.Stenosis degree of vascular≥70%was considered to be obvious vascular stenosis.Receiver operating characteristic(ROC)curve analysis was performed to evaluate the diagnosis performance with area under the receiver operating characteristic curve(AUC).Sensitivity and specificity were recorded.Results Finally,a total of 266 vessels from 90 patients with CCTA-AI and FFR-CT value were included.The AUC values were 0.85,0.817 and 0.883 for CCTA-AI,FFR-CT and the combination of CCTA-AI and FFR-CT.The sensitivity,specificity to identify coronary stenosis were 82.09%,81.08%,for CCTA-AI combined with FFR-CT.Conclusion The diagnostic efficiency of CCTA-AI combined with FFR-CT for coronary artery stenosis has been improved.
作者
孙一
刘博
孙秀彬
李桂杰
崔丁也
马振申
SUN Yi;LIU Bo;SUN Xiubin;LI Guijie;CUI Dingye;MA Zhenshen(Department of Radiology,The First Affiliated Hospital of Shandong First Medical University,Jinan 250014,China;Department of Biostatistics,School of Public Health,Cheeloo College of Medicine,ShandongUniversity,Jinan 250012,China)
出处
《医学影像学杂志》
2022年第12期2081-2085,共5页
Journal of Medical Imaging
关键词
冠状动脉粥样硬化性心脏病
人工智能
冠状动脉血管成像
冠状动脉造影
血流储备分数
体层摄影术
X线计算机
Coronary atherosclerotic heart disease
Artificial intelligence
Coronary computed tomography angiography
Coronary angiography
Fractional flow reserve
Tomography,X-ray computed
