Separation of arteries and veins in 3D MR angiography using correlation analysis

Abstract

Multiphase contrast-enhanced 3D MR angiography (MRA) data sets allow the separate visualization of the arterial and venous pulmonary vasculature. However, due to short arterial-to-venous bolus transit times in the lung, the generation of pure venograms without arterial overlay is difficult. To suppress arterial signal in venograms, early arterial phase data are typically subtracted from peak venous phase images. In this study, a correlation algorithm is used to postprocess the multiphase 3D MRA data sets. The cross-correlation between a measured arterial or venous reference function and the local signal-time course is computed which highlights image locations with a similar signal-time curve as the reference function and suppresses constant signal. Conventional maximum intensity projections (MIP) are generated from the arterial and venous correlation maps. In a study with five volunteers, an increase in SNR by a factor of 2.1 (1.8) of arterial (venous) correlation MIP images over subtraction MIP images was observed.