基于示踪动力学的冠状动脉CT血流储备分数诊断冠状动脉缺血病变的初步研究

Preliminary evaluation of coronary fractional flow reserve derived from CT based on tracer-kinetics principle

目的评价基于示踪动力学原理的冠状动脉CT血流储备分数(CT-FFR)测量方法的应用可行性及诊断效能。方法回顾性纳入自2015年1月至2019年3月在首都医科大学附属北京安贞医院就诊并顺序完成冠状动脉CT血管成像(CCTA)、冠状动脉造影及有创FFR测量的患者共计130例(159支冠状动脉血管),对CCTA图像进行病变狭窄程度主观评价,并采用基于示踪动力学的CT-FFR测量方法进行无创FFR测量。以有创FFR结果为参照,采用Bland-Altman方法评价CT-FFR的诊断一致性。评价基于示踪动力学的CT-FFR测量结果及CCTA管腔狭窄程度对冠状动脉缺血性病变诊断的灵敏度、特异度、阳性预测值、阴性预测值及受试者操作特征(ROC)曲线下面积(AUC),并依据患者心率分为>65次/min和≤65次/min两组,评价基于示踪动力学的CT-FFR测量方法在不同心率组的诊断效能。采用χ2检验及DeLong检验比较不同评价方法及心率间的诊断效能差异。采用logistic回归分析,评价心率、图像层厚、图像强化程度及噪声等因素对诊断准确度的影响。结果Bland-Altman分析显示两方法的测量平均差值为-0.01(-0.11~0.10)。与有创FFR结果对照,基于示踪动力学的CT-FFR测量方法对冠状动脉缺血病变的诊断敏感度92.4%,特异度82.1%,阳性预测值87.6%,阴性预测值88.7%,ROC曲线下面积(AUC)值为0.94,与根据CCTA图像评价的管腔狭窄程度的诊断效能相比,两者间差异具有统计学意义(P<0.05)。CT-FFR测量方法在不同心率组间的诊断效能差异无统计学意义。回归分析显示心率、图像层厚、图像强化程度及噪声等因素对该方法的诊断准确度无显著影响。结论基于示踪动力学的CT-FFR测量方法可快速完成基于图像质量满足临床诊断需要的CCTA图像的无创FFR值测量,在诊断冠状动脉缺血病变中具有很好的诊断效能,在较快心率下仍能保持良好的诊断准确性,且其诊断准确度未受到心率、图像层厚、图像强化程度及噪声等因素的显著影响。

Objective To evaluate the feasibility and diagnostic efficacy of the coronary fractional flow reserve derived from CT(CT-FFR)measurement method based on tracer pharmacokinetic principle.Methods A total of 130 patients(159 coronary artery vessels)who were admitted to Beijing Anzhen Hospital from January 2015 to March 2019 were included in this study retrospectively.All patients had completed coronary CT angiography(CCTA)and invasive coronary angiography with invasive FFR.Subjective assessment of stenosis degree was performed on CCTA images and non-invasive FFR measurement was performed by using a tracer-kinetics based on CT-FFR measurement method.The Bland-Altman method was used to evaluate the diagnostic consistency of the two methods.Compared with the invasive FFR results,the sensitivity,specificity,positive predictive value,negative predictive value,and area under the ROC curve(AUC)of tracer-kinetics based on CT-FFR results for the diagnosis of coronary ischemic lesions were evaluated.All cases were divided into two heart rate groups:>65 bpm and≤65 bpm.The diagnostic efficacy of tracer-kinetics based on CT-FFR in different heart rate groups was evaluated.χ2 test and DeLong test were used to compare diagnostic performance in different evaluation methods and heart rate groups.Logistic regression analysis was used to evaluate the impact of factors such as heart rate,image thickness,image enhancement,and noise on the accuracy of diagnosis.Results Bland-Altman analysis showed that the average difference between the two methods was-0.01.(-0.11-0.10).Compared with invasive FFR results,the tracer-kinetics based on CT-FFR method had a diagnostic sensitivity of 92.4%,specificity of 82.1%,positive predictive value of 87.6%,negative predictive value of 88.7%,and the area under ROC curve(AUC)value was 0.94.Compared with the diagnostic efficacy of luminal stenosis evaluated based on CCTA images,the difference was significantly statistical(P<0.05).The diagnostic performance of CT-FFR had no statistically significant difference between the two heart rate groups.Factors such as heart rate,image thickness,image enhancement,and noise had no significant effect on the diagnostic accuracy of the tracer-kinetics based on CT-FFR method.Conclusions The tracer-kinetics based on CT-FFR method may quickly complete the non-invasive FFR measurement on CCTA images with image quality that meets the needs of clinical diagnosis.It has a good diagnostic performance in the diagnosis of coronary ischemic lesions even for those cases with a faster heart rate.The diagnostic accuracy of tracer-kinetics based on CT-FFR method is not significantly affected by factors such as heart rate,image thickness,image enhancement,and image noise.