Diagnostic accuracy of stress myocardial perfusion MRI and late gadolinium-enhanced MRI for detecting flow-limiting coronary artery disease: a multicenter study

Diagnostic accuracy of stress myocardial perfusion MRI and late gadolinium-enhanced MRI for detecting flow-limiting coronary artery disease: a multicenter study
复制标题

DOI:
10.1007/s00330-008-1097-4
复制
发表时间:
2008-12-01
期刊:
影响因子:
5.9
通讯作者:
Takeda, Kan
Takeda, Kan
中科院分区:
医学2区
文献类型:
--
作者:
Kitagawa, Kakuya;Sakuma, Hajime;Takeda, Kan

文献摘要

被引文献

相似文献

本研究的目的是确定负荷和静息灌注磁共振成像(MRI)和晚期钆增强(LGE)MRI识别阻塞性冠状动脉疾病(CAD)患者的诊断性能。共有50例疑似CAD患者接受了负荷-静息灌注MRI,随后使用1.5 T系统进行了LGE MRI。负荷-静息灌注MRI导致观察者1的受试者工作特征曲线下面积(AUC)为0.92,观察者2为0.84,观察者1的灵敏度和特异性分别为89%(32/36)和79%(11/14),观察者2的灵敏度和特异性分别为83%(30/36)和71%(10/14),显示中等程度的观察者间一致性(Cohen's kappa=0.49)。虽然负荷-静息灌注和LGE MRI的组合并没有提高预测限流性阻塞性CAD的准确性(观察者1的AUC为0.81,观察者2的AUC为0.80),但两名观察者的敏感性均增加至92%,观察者间基本一致(kappa=0.70)。负荷-静息心肌灌注MRI是识别阻塞性CAD患者的准确诊断测试。
The aim of this study was to determine the diagnostic performance of stress and rest perfusion magnetic resonance imaging (MRI) and late gadolinium-enhanced (LGE) MRI for identifying patients with obstructive coronary artery disease (CAD). A total of 50 patients with suspected CAD underwent stress-rest perfusion MRI, followed by LGE MRI with a 1.5-T system. Stress-rest perfusion MRI resulted in an area under the receiver-operating characteristic curve (AUC) of 0.92 for observer 1 and 0.84 for observer 2 with sensitivity and specificity of 89% (32/36) and 79% (11/14) by observer 1, 83% (30/36) and 71% (10/14) by observer 2, respectively, showing a moderate interobserver agreement (Cohen's kappa=0.49). While combination of stress-rest perfusion and LGE MRI did not result in improved accuracy for the prediction of flow-limiting obstructive CAD (AUC 0.81 for observer 1 and 0.80 for observer 2), the sensitivity was increased to 92% in both observers with a substantial interobserver agreement (kappa=0.70). Stress-rest myocardial perfusion MRI is an accurate diagnostic test for identifying patients with obstructive CAD.