Noninvasive detection of coronary artery stenosis with 16-slice spiral computed tomography in a population at low to moderate risk for coronary artery disease

Abstract

To evaluate the feasibility and accuracy of 16-slice spiral computed tomography (MSCT) in detecting coronary artery lesions in a low to moderate risk population. MSCT was performed in 33/78 patients at low to moderate risk for coronary artery disease affected by mitral and/or aortic valve disease and scheduled for conventional coronary angiography for pre-surgery evaluation. The sensitivity (SE), specificity (SP), positive (PPV) and negative predictive values (NPV) to identify >or=50% obstructed lesions were evaluated. Moreover, a discriminant predictive statistical analysis was performed. MSCT visualised 428/479 (89%) segments identified at coronary angiography and the percentage of evaluable segments was 87%. By including all coronary branches, SE, SP, PPV and NPV were 69%, 96%, 58% and 97%, respectively. By excluding the segments with a diameter of <or=2 mm, SE and SP were 74% and 96%, respectively (MSCT correctly identified 17/23 lesions). In the evaluable coronary arteries, SE and SP were 86% and 88% for the left anterior descending coronary artery, 100% and 85% for the left circumflex coronary artery, and 86% and 89% for the right coronary artery, respectively. SP was 100% for the left main artery. When restricting the analysis to the patients, 22/23 patients were adequately evaluable, and SE, SP, PPV and NPV were 100%, 80%, 85% and 100%, respectively. Using the discriminant predictive statistical analysis of quantitative (age, heart rate and calcium score) and qualitative variables (presence or absence of critical lesions at MSCT), an overall accuracy of 85% was obtained. The percentage of adequately evaluable patients was low because of limitations due to calcifications. In selected populations at low to moderate risk for coronary artery disease, MSCT has a high NPV in the evaluation of coronary artery stenosis.