BACKGROUND Exercise electrocardiography (ECG) is frequently used in the work-up of patients with suspected coronary artery disease (CAD), however the accuracy is reduced in women. Cardiovascular magnetic resonance (CMR) stress testing can accurately diagnose CAD in women. To date, a direct comparison of CMR to ECG has not been performed. RESULTS We prospectively enrolled 88 consecutive women with chest pain or other symptoms suggestive of CAD. Patients underwent a comprehensive clinical evaluation, exercise ECG, a CMR stress test including perfusion and infarct imaging, and x-ray coronary angiography (CA) within 24 hours. CAD was defined as stenosis ≥70% on quantitative analysis of CA.Exercise ECG, CMR and CA was completed in 68 females (age 66.4 ± 8.8 years, number of CAD risk factors 3.5±1.4). The prevalence of CAD on CA was 29%. The Duke treadmill score (DTS) in the entire group was -3.0±5.4 and was similar in those with and without CAD (-4.5±5.8 and -2.4±5.1; P=0.12). Sensitivity, specificity and accuracy for CAD diagnosis was higher for CMR compared with exercise ECG (sensitivities 85% and 50%, P=0.02, specificities 94% and 73%, P=0.01, and accuracies 91% and 66%, P=0.0007, respectively). Even after applying the DTS the accuracy of CMR was higher compared to exercise ECG (area under ROC curve 0.94±0.03 vs 0.56±0.07; P=0.0001). CONCLUSIONS In women with intermediate-to-high risk for CAD who are able to exercise and have interpretable resting ECG, CMR stress perfusion imaging has higher accuracy for the detection of relevant obstruction of the epicardial coronaries when directly compared to exercise ECG.