Hambÿe A S, Vervaet A, Lieber S, Ranquin R
Department of Nuclear Medicine, Middelheim General Hospital, Antwerp, Belgium.
J Nucl Cardiol. 1996 Nov-Dec;3(6 Pt 1):464-74. doi: 10.1016/s1071-3581(96)90056-2.
The diagnostic value and incremental contribution of different noninvasive tests to the identification of coronary artery disease in 128 patients from a general population with intermediate pretest likelihood (48.0%) were determined by ordered logistic regression analysis and receiver-operating characteristic (ROC) curves.
Patients referred for suspicion of coronary heart disease were submitted to bicycle exercise testing under clinical and electrocardiographic control. AT peak exercise, first-pass radionuclide angiography was performed after injection of 99mTc-labeled sestamibi, followed by single-photon emission computed tomographic (SPECT) acquisition. A comparative rest study was obtained within 1 week, and qualitative and quantitative analysis was applied to assess the presence and extent of disease. With coronary angiography and 50% stenosis used as a standard, the discriminative accuracy of each test was calculated. The accuracies to diagnose coronary heart disease were 71.3% +/- 4.7% for the bicycle test, 66.7% +/- 5.3% for radionuclide angiography, and 81.6% +/- 3.9% for the SPECT data. By ROC curves, the optimal criteria for positivity were determined for the visual and quantitative analysis for both presence and extent of coronary artery disease. Results of visual and quantitative SPECT were compared in terms of area under the ROC curves. The diagnostic performances showed no significant difference, ranging from 74.3% to 81.6%. The first-pass radionuclide angiographic and SPECT data were added progressively to the stress testing to evaluate their incremental diagnostic contribution. Only the addition of SPECT results significantly increased the accuracy to 85.6% +/- 3.3% (p < 0.0001).
Exercise electrocardiography and first-pass radionuclide angiography showed comparable accuracy to detect coronary artery disease. However, the combination of exercise testing and visual SPECT analytic data sufficed to ensure diagnostic accuracy, without significant benefit from the addition of other tests or the application of quantification.
通过有序逻辑回归分析和受试者工作特征（ROC）曲线，确定了不同无创检查对128例具有中等预检可能性（48.0%）的普通人群中冠心病识别的诊断价值和增量贡献。
因疑似冠心病而转诊的患者在临床和心电图监测下进行自行车运动试验。运动高峰时，注射99mTc标记的 sestamibi后进行首次通过放射性核素血管造影，随后进行单光子发射计算机断层扫描（SPECT）采集。在1周内进行对照静息研究，并应用定性和定量分析来评估疾病的存在和范围。以冠状动脉造影和50%狭窄为标准，计算每项检查的鉴别准确性。自行车试验诊断冠心病的准确性为71.3%±4.7%，放射性核素血管造影为66.7%±5.3%，SPECT数据为81.6%±3.9%。通过ROC曲线，确定了冠状动脉疾病存在和范围的视觉和定量分析的最佳阳性标准。比较了视觉和定量SPECT在ROC曲线下面积方面的结果。诊断性能无显著差异，范围为74.3%至81.6%。将首次通过放射性核素血管造影和SPECT数据逐步添加到负荷试验中，以评估其增量诊断贡献。仅添加SPECT结果可显著提高准确性至85.6%±3.3%（p<0.0001）。
运动心电图和首次通过放射性核素血管造影在检测冠状动脉疾病方面显示出相当的准确性。然而，运动试验和视觉SPECT分析数据的组合足以确保诊断准确性，添加其他检查或应用定量分析并无显著益处。
