Meeting Banner
Abstract #0280

Human Statistical Atlas of Cardiac Fiber Architecture from DT-MRI

Herve Lombaert1,2, Jean-Marc Peyrat3, Stanislas Rapacchi4, Laurent Fanton5, Herve Delingette2, Nicholas Ayache2, Pierre Croisille4

1Ecole Polytechnique, Montreal, QC, Canada; 2Asclepios, INRIA, Sophia-Antipolis, France; 3Siemens Molecular Imaging, Oxford, United Kingdom; 4Creatis-LRMN, HCL, Lyon, France; 5Institut Universitaire de Mdecine Lgale, Lyon, France

A human statistical atlas of the cardiac fiber architecture is constructed from ex-vivo diffusion tensor images and is based on a set of 10 normal human hearts. To the best of our knowledge, this is the first time that such study has been conducted with human data. We have developed a semi-automated method where only minimal interactions are required for the segmentation of the myocardium, and where the registrations are fully automated via symmetric log domain diffeomorphic demons. The results on the variability of human cardiac fibers concur with studies on other mammals. The cardiac fiber orientation is indeed much more consistent across our population than the orientation of the cardiac laminar sheets.