Spatial registration of echocardiographic and positron emission tomographic heart studies

A method has been developed to match corresponding heart regions from functional echocardiographic (Echo) and metabolic fluorine-l8-fluoro-2-deoxy-d-glucose ([18F]FDG) positron emission tomography (PET) studies in individual patients. Echo and PET images are spatially correlated by determining homologous anatomical landmarks (the two papillary muscles and the inferior junction of the right ventricle), identifiable in images obtained by both acquisition modalities. Echo-PET image registration is first performed in the plane identified by the three landmarks, using a rigid rotate-translate scale model. The registration parameters are then used to transform the whole PET volume. This allows a consistent Echo-PET regional analysis, according to a segmental subdivision of the heart. The technique was tested on patients. The overlay of Echo and PET registered images proved the reliability of realignment of the three markers and a good spatial correlation of myocardial walls. This approach to image registration could be applied to other acquisition modalities (such as magnetic resonance imaging and single-photon emission tomography), provided that the three anatomical landmarks are visualized.