Pizzuto Francesco, Voci Paolo, Bartolomucci Francesco, Puddu Paolo Emilio, Strippoli Giovanni, Broglia Laura, Rossi Plinio
Department of Cardiology, Tor Vergata University, Rome, Italy.
Am J Cardiol. 2009 Sep 1;104(5):657-64. doi: 10.1016/j.amjcard.2009.04.044. Epub 2009 Jun 24.
Multidetector computed tomography (MDCT) detects coronary artery disease. However, an overestimation of coronary artery stenosis and artifacts can prevent accurate identification of significant coronary narrowing. The combination of MDCT with coronary flow reserve (CFR), the hyperemic/baseline peak flow velocity ratio, measured by transthoracic Doppler echocardiography might be helpful. We studied 144 consecutive patients with CFR and quantitative coronary angiography, obtained using both MDCT and invasive coronary angiography (reference method). It was hypothesized that the CFR might provide an incremental value to MDCT in detecting significant (> or =70%) left anterior descending (LAD) coronary artery stenosis. A CFR cutoff of <2 was used to discriminate significant stenosis. CFR was feasible in 141 (98%) of 144 patients, and MDCT was feasible in 131 (91%) of 144 patients (p <0.02). In a univariate model, the prediction of significant LAD stenosis was slightly, but significantly (p <0.0001), better with CFR (sensitivity 90%, specificity 96%, positive predictive value 84%, negative predictive value 97%, and diagnostic accuracy 94%, chi-square = 97.5) than with MDCT (sensitivity 80%, specificity 93%, positive predictive value 71%, negative predictive value 95%, diagnostic accuracy 90%, chi-square = 63.2). When the findings from transthoracic Doppler echocardiography and MDCT agreed, the diagnostic accuracy increased (96%; chi-square = 86.1, p <0.0001). In a multivariate prediction of significant LAD stenosis using a logistic neural network, CFR overshadowed MDCT, and the area under the receiver operating curve was 0.99. Of the 13 patients missed by MDCT, the diagnostic accuracy of transthoracic Doppler echocardiography to predict significant LAD stenosis was 100%. Thus, CFR could improve the diagnostic accuracy of MDCT to detect significant LAD stenosis.
多排螺旋计算机断层扫描(MDCT)可检测冠状动脉疾病。然而,冠状动脉狭窄的高估和伪影可能会妨碍对显著冠状动脉狭窄的准确识别。将MDCT与通过经胸多普勒超声心动图测量的冠状动脉血流储备(CFR,即充血期/基线期峰值流速比值)相结合可能会有所帮助。我们对144例连续患者进行了CFR检查和定量冠状动脉造影,同时采用了MDCT和有创冠状动脉造影(参考方法)。研究假设CFR在检测显著(≥70%)的左前降支(LAD)冠状动脉狭窄方面可能为MDCT提供额外价值。采用CFR截断值<2来鉴别显著狭窄。144例患者中有141例(98%)可行CFR检查,144例患者中有131例(91%)可行MDCT检查(p<0.02)。在单变量模型中,CFR对显著LAD狭窄的预测略好,但具有显著差异(p<0.0001),其敏感性为90%,特异性为96%,阳性预测值为84%,阴性预测值为97%,诊断准确性为94%,卡方值=97.5,优于MDCT(敏感性80%,特异性93%,阳性预测值71%,阴性预测值95%,诊断准确性90%,卡方值=63.2)。当经胸多普勒超声心动图和MDCT的结果一致时,诊断准确性提高(96%;卡方值=86.1,p<0.0001)。在使用逻辑神经网络对显著LAD狭窄进行多变量预测时,CFR比MDCT更具优势,受试者工作特征曲线下面积为0.99。在MDCT漏诊的13例患者中,经胸多普勒超声心动图预测显著LAD狭窄的诊断准确性为100%。因此,CFR可提高MDCT检测显著LAD狭窄的诊断准确性。