Stress-ECG vs. CT coronary angiography for the diagnosis of coronary artery disease: a "real-world" experience

Purpose. This study aimed to evaluate the diagnostic accuracy of stress electrocardiogram (ECG) and computed tomography coronary angiography (CTCA) for the detection of significant coronary artery stenosis (>= 50%) in the real world using conventional CA as the reference standard. Materials and methods. A total of 236 consecutive patients (159 men, 77 women; mean age 62.8 +/- 10.2 years) at moderate risk and with suspected coronary artery disease (CAD) were enrolled in the study and underwent stress ECG, CTCA and CA. The CTCA scan was performed after i.v. administration of a 100-ml bolus of iodinated contrast material. The stress ECG and CTCA reports were used to evaluate diagnostic accuracy compared with CA in the detection of significant stenosis >= 50%. Results. We excluded 16 patients from the analysis because of the nondiagnostic quality of stress ECG and/or CTCA. The prevalence of disease demonstrated at CA was 62% (n=220), 51% in the population with comparable stress ECG and CTCA (n=147) and 84% in the population with equivocal stress ECG (n=73). Stress ECG was classified as equivocal in 73 cases (33.2%), positive in 69 (31.4%) and negative in 78 (35.5%). In the per-patient analysis, the diagnostic accuracy of stress ECG was sensitivity 47%, specificity 53%, positive predictive value (PPV) 51% and negative predictive value (NPV) 49%. On stress ECG, 40 (27.2%) patients were misclassified as negative, and 34 (23.1%) patients with nonsignificant stenosis were overestimated as positive. The diagnostic accuracy of CTCA was sensitivity 96%, specificity 65%, PPV 74% and NPV 94%. CTCA incorrectly classified three (2%) as negative and 25 (17%) as positive. The difference in diagnostic accuracy between stress ECG and CTCA was significant (p<0.01). Conclusions. CTCA in the real world has significantly higher diagnostic accuracy compared with stress ECG and could be used as a first-line study in patients at moderate risk.