Session S62.5

Evaluation of Magnetic Resonance First-Pass Contrast-Enhanced Data Registration as a Basis for Quantification of Myocardial Perfusion

N Kachenoura*, P Cluzel, CA Cuenod, P Grenier,
F Frouin, D Balvay

INSERM
Paris , France

Rest-stress contrast enhanced cardiac magnetic resonance (CE-CMR) data are increasingly used for the evaluation of perfusion abnormalities related to flow-limiting coronary artery diseases (CAD). Quantification of myocardial perfusion is time consuming since it requires manual shifts of myocardial regions of interest to compensate for the respiratory-induced motion. Thus, the aim of this study was to implement an automated registration method and to test its efficiency at rest and during stress. Methods. We studied 10 patients undergoing a clinical CMR with contrast media for the evaluation of CAD. Three slices corresponding to perfusion data acquired at rest and stress were selected for each patient. For each dynamic sequence, a reference image on which we can visualize cardiac structures was selected, then a rigid registration algorithm based on gradient amplitude and orientation analysis was performed. The quality of this registration was first assessed qualitatively: a k-mean clustering was applied on dynamic series before and after registration. The ability to clearly identify heart structures on the resulted k-mean maps and the symmetry of the myocardial shape were visually evaluated. In addition, noise amplitude was quantified within the myocardium before and after registration, using a robust estimator* based on the assessment of the non correlated part of intensity time courses. Myocardium was manually delineated on the reference image. Results. Before registration the myocardial symmetry and heart structures identification were noted in 16/30 slices at rest and 5/30 slices during stress. On the remaining slices, registration provided a substantial improvement in symmetry and structure identification in 10/14 slices at rest and 20/25 slices during stress. Besides, noise amplitude (mean±SD) decreased from 48±26 to 28±10 at rest (p<0.05) and from 53±13 to 31±10 during stress (p<0.05). In conclusion, a robust registration method was successfully tested on rest and stress CMR perfusion data. It provides a valuable basis for quantitative assessment of CMR myocardial perfusion. *.New criteria for assessing fit quality in dynamic contrast-enhanced T1-weighted MRI for perfusion. Balvay D et al, Magn Reson Med.2005;54(4):868-77.

(Abstract Control Number: 203)