Meeting Banner
Abstract #0882

Motion, Resolution and Noise Thresholds for the Accurate Classification of Human Coronary Atherosclerotic Plaque by MRI

Paula Montesinos1, 2, Himanshu Bhat3, Guangping Dai4, Manuel Desco1, 2, Elfar Adalsteinsson5, Reza Nezafat6, David E. Sosnovik7

1Departamento de Bioingeniera e Ingeniera Aeroespacial, Universidad Carlos III de Madrid, Madrid, Spain; 2Instituto de Investigacin Sanitaria Gregorio Maran (IiSGM), Madrid, Spain; 3Siemens Medical Solutions, Charlestown, MA, United States; 4Massachusetts General Hospital, Charlestown, MA, United States; 5Martinos Center for Biomedical Imaging, Massachusetts General Hospital, Harvard Medical School, Charlestown, MA, United States; 6Beth Israel Deaconess Medical Center, Harvard Medical School, Boston, MA, United States; 7Martinos Center for Biomedical Imaging, Massachusetts General Hospital, Harvard Medical School, Charlestown, MA, United States

The requirements for the accurate classification of human coronary atherosclerotic plaque by MRI in vivo remain unknown. We aimed here to reconstruct a high resolution ex vivo ground truth dataset of human coronary atherosclerotic plaques with varying degrees of spatial resolution, SNR and motion. Our results indicate that a spatial resolution of 0.5 mm, SNR > 10 and motion correction in all 3 directions are needed for in vivo MRI of coronary plaque. We further show that the acquisition of the central 10-15% of k-space during a motion-free breathold could reduce artifacts even further.

Abstract PDF Presentation Video

Keywords

acceptable acceptance accuracy accurate accurately acquiring acquisition added adult aimed almost amounts applied artery artifact attainable audience beth biomedical calcification calcium central classification classify clinicians conditions coronary correction created deaconess degrees determine developed disease donor either eliminating error even extensive extremely features flash fluorocarbon foot free general ground head healthy heart hearts hospital human impact imperfect improved in vivo inadequate isotropic keep largest least limits little lumen maintain marked measure measured medical medium modeled morphologic motion navigator navigators needed noise original outer padding part particularly percent perfect pixels placing plaque plaques plots poor portion produced produces profile profiles progressively proved providing quality realistic recent reconstruct reconstructed reduce reduces reducing reduction remainder remaining remains representative require requirements requires residual resolution respectively respiratory root sampled school scientists seven simulate since slight solutions space spatial squared strategies studied subject subjective suppression target thresholds translation translational truth valuable varying ventricle vessel volunteer window zero