【摘要】 目的：探究基于人工智能的冠状动脉CT血管造影（coronary artery CT angiography,CCTA）定量评估与冠状动脉狭窄之间的关系。方法：回顾性选取2020年9月—2021年12月在连云港市第一人民医院接受CCTA检查并在一个月内接受数字减影-血管造影（Digital Subtraction Angiography,DSA）检查的患者198例，收集患者基本临床资料及CCTA数据。分别将人工智能分析CCTA及DSA检查测得的最高狭窄程度及单支血管狭窄程度进行对比，绘制冠状动脉钙化积分、CT-FFR和冠状动脉狭窄程度的受试者工作特征（receiver operating characteristic,ROC）曲线，并计算出曲线下面积。结果：CCTA与DSA检查测得的最高狭窄程度具有良好的一致性，差异有统计学意义（r=0.616,P<0.05）;CCTA与DSA检查测得的左回旋支、左前降支、右冠状动脉最高狭窄程度具有良好的一致性，差异有统计学意义（r值分别为0.907、0.729和0.925,P值均<0.05）。单一钙化积分、单一CT-FFR、二者联合值诊断冠状动脉狭窄程度的RDC曲线下面积分别为0.908 (95%CI为0.867～0.950,P<0.05)、0.959 (95%CI为0.920～0.998,P<0.05)、0.973 (95%CI为0.944～1.000,P<0.05)。结论：基于人工智能的冠状动脉定量评估有助于提高对冠状动脉狭窄程度判断的准确性，减少诊断时间。更多还原

【Abstract】 OBJECTIVE To verify the relationship between the quantitative assessment of coronary artery CT angiography(CCTA)based on artificial intelligence and coronary artery stenosis. METHODS A total of 198 patients who underwent coronary CCTA examination in Lianyungang First People’ s Hospital from September 2020 to December 2021 and coronary digital subtraction angiography(DSA) examination within one month were selected,and the basic clinical data and CCTA data of the patients were collected. The highest stenosis measured by the patient’ s artificial intelligence analysis CCTA was compared with the highest stenosis measured by DSA. The degree of stenosis of a single vessel measured by the patient’s artificial intelligence analysis of CCTA was compared with the degree of stenosis of a single vessel measured by DSA. The receiver operating characteristic(ROC)curves of the patient’s coronary artery calcification score,CT fractional flow reserve(CT-FFR) and coronary artery stenosis were plotted. RESULTS The highest degree of stenosis measured by CCTA was in good agreement with the highest degree of stenosis measured by DSA(r=0.616, P<0.05). There was a good agreement between the highest stenosis of each coronary artery measured by CCTA and the highest stenosis of each coronary artery measured by DSA(r=0.907,0.729,0.925,P<0.055). The area under the curve for the diagnosis of coronary stenosis with a single calcification score was 0.908(95%CI:0.867～0.950,P<0.05),the area under the curve for the diagnosis of coronary stenosis with a single CT-FFR was 0.959(95%CI:0.920～0.998,P<0.05), and the area under the curve for the diagnosis of the two was 0.973(95% CI: 0.944～1.000, P<0.05).CONCLUSION Quantitative assessment of coronary arteries based on artificial intelligence is helpful to improve the accuracy of the judgment of coronary artery stenosis.更多还原