Negative predictive value of normal adenosine-stress cardiac MRI in the assessment of coronary artery disease and correlation with semiquantitative perfusion analysis.

Purpose: To prospectively determine the negative predictive value of normal adenosine stress cardiac MR (CMR) in routine patients referred for evaluation of coronary artery disease (CAD), predominantly with intermediate to high pretest risk. Materials and Methods: Consecutive patients referred for coronary angiography were examined in a 1.5 Tesla whole-body scanner before catheterization. A total of 158 patients with normal CMR on qualitative assessment were included, and semiquantitative perfusion analysis was performed. Significant CAD was regarded as luminal narrowing of >= 70% in coronary angiography. Results: In the 158 study patients, negative predictive value of normal adenosine-stress CMR for significant CAD was 96.2% (for stenosis >= 90%: 98.1%). True-negative and false-negative patients were comparable regarding clinical presentation, risk factors, and CMR findings. Semiquantitative perfusion analysis gave significantly prolonged arrival time index and peak time Index in the false-negative group. Using cutoff values >1.8 for arrival time index or >1.2 for peak time index, the CMR negative predictive value increased to 98.7% (for stenosis >= 90%: to 100%). Conclusion: The very high negative predictive value for CAD supports CMR-based decision making for the indication to coronary angiography. Semiquantitative perfusion analysis seems promising to identify the small group of CAD patients not detectable by qualitative CMR assessment.