Purpose: The assessment of inducible wall motion abnormalities during high-dose dobutamine-stress cardiovascular magnetic resonance (DCMR) is well established for the identification of myocardial ischemia at 1.5 Tesla. Its feasibility at higher field strengths has not been reported. The present study was performed to prospectively determine the feasibility and diagnostic accuracy of DCMR at 3 Tesla for depicting hemodynamically significant coronary artery stenosis (> or = 50% diameter stenosis) in patients with suspected or known coronary artery disease (CAD).
Materials and methods: Thirty consecutive patients (6 women) (66 +/- 9.3 years) were scheduled for DCMR between January and May 2007 for detection of coronary artery disease. Patients were examined with a Philips Achieva 3 Tesla system (Philips Healthcare, Best, The Netherlands), using a spoiled gradient echo cine sequence. Technical parameters were: spatial resolution 2 x 2 x 8 mm3, 30 heart phases, spoiled gradient echo TR/TE: 4.5/2.6 msec, flip angle 15 degrees . Images were acquired at rest and stress in accordance with a standardized high-dose dobutamine-atropine protocol during short breath-holds in three short and three long-axis views. Dobutamine was administered using a standard protocol (10 microg increments every 3 minutes up to 40 microg dobutamine/kg body weight/minute plus atropine if required to reach target heart rate). The study protocol included administration of 0.1 mmol/kg/body weight Gd-DTPA before the cine images at rest were acquired to improve the image quality. The examination was terminated if new or worsening wall-motion abnormalities or chest pain occurred or when > 85% of age-predicted maximum heart rate was reached. Myocardial ischemia was defined as new onset of wall-motion abnormality in at least one segment. In addition, late gadolinium enhancement (LGE) was performed. Images were evaluated by two blinded readers. Diagnostic accuracy was determined with coronary angiography as the reference standard. Image quality and wall-motion at rest and maximum stress level were evaluated using a four-point scale.
Results: In 27 patients DCMR was performed successfully, no patient had to be excluded due to insufficient image quality. Twenty-two patients were examined by coronary angiography, which depicted significant stenosis in 68.2% of the patients. Patient-based sensitivity and specificity were 80.0% and 85.7% respectively and accuracy was 81.8%. Interobserver variability for assessment of wall motion abnormalities was 88% (kappa = 0.760; p < 0.0001). Negative and positive predictive values were 66.7% and 92.3%, respectively. No significant differences in average image quality at rest versus stress for short or long-axis cine images were found.
Conclusion: High-dose DCMR at 3T is feasible and an accurate method to depict significant coronary artery stenosis in patients with suspected or known CAD.