64-slice CT for diagnosis of coronary artery disease: a systematic review.

PURPOSE The purpose of this systematic review was to assess the accuracy of 64-slice CT coronary angiography for the diagnosis of coronary artery disease. METHODS We attempted to identify all published trials in all languages that used 64-slice CT to diagnose coronary artery disease. Results of 64-slice CT coronary angiography were compared with invasive coronary angiography or intravascular ultrasound. RESULTS Sensitivity of 64-slice CT for significant (> or =50%) stenosis, based on pooled data from all studies, was > or =90% in patient-based evaluations, named vessels, segments, and coronary artery bypass grafts, except the left circumflex (sensitivity 88%), distal segments (80%), and stents (88%). Specificity was 88% in patient-based evaluations, and > or =90% at individual sites. Positive predictive values for patient-based evaluations, left main coronary artery, and coronary artery bypass grafts ranged from 91% to 93%, but elsewhere ranged from 69% to 84%. Negative predictive values were 96% to 100%. Positive likelihood ratios for patient-based evaluations were 8.0 and, at specific sites, were > or =9.7. Negative likelihood ratios, except for distal segments, were <0.1. CONCLUSION Negative 64-slice CT reliably excluded significant coronary disease. However, the data suggest that stenoses shown on 64-slice CT require confirmation. Combining the results of 64-slice CT with a pre-CT clinical probability assessment would strengthen the diagnosis. Due to the risk of radiation-induced cancer, patients should be selected carefully for this test, and scan protocols should be optimized to minimize risk.