Diagnostic Accuracy and Clinical Utility of Noninvasive Testing for Coronary Artery Disease

Background: Computed tomography coronary angiography (CTCA) has become a popular noninvasive test for diagnosing coronary artery disease. Objective: To compare the accuracy and clinical utility of stress testing and CTCA for identifying patients who require invasive coronary angiography (ICA). Design: Observational study. Setting: University medical center in Rotterdam, the Netherlands. Patients: 517 patients referred by their treating physicians for evaluation of chest symptoms by using stress testing or ICA. Intervention: Stress testing and CTCA in all patients. Measurements: Diagnostic accuracy of stress testing and CTCA compared with ICA; pretest probabilities of disease by Duke clinical score; and clinical utility of noninvasive testing, defined as a pretest or posttest probability that suggests how to proceed with testing (no further testing if <= 5%, proceed with ICA if between 5% and 90%, and refer directly for ICA if >= 90%). Results: Stress testing was not as accurate as CTCA; CTCA sensitivity approached 100%. In patients with a low (<20%) pretest probability of disease, negative stress test or CTCA results suggested no need for ICA. In patients with an intermediate (20% to 80%) pretest probability, a positive CTCA result suggested need to proceed with ICA (posttest probability, 93% [95% CI, 92% to 93%]) and a negative result suggested no need for further testing (posttest probability, 1% [CI, 1% to 1%]). Physicians could proceed directly with ICA in patients with a high (>80%) pretest probability (91% [CI, 90% to 92%]). Limitations: Referral and verification bias might have influenced findings. Stress testing provides functional information that may add value to that from anatomical (CTCA or ICA) imaging. Conclusion: Computed tomography coronary angiography seems most valuable in patients with intermediate pretest probability of disease, because the test can distinguish which of these patients need invasive angiography. These findings need to be confirmed before CTCA can be routinely recommended for these patients.