Meeting Banner
Abstract #1162

An Improved Binning Strategy for 3D Image-Based Respiratory Motion Correction in Whole-Heart Coronary MRA

Davide Piccini1, Arne Littmann2, Hui Xue3, Jens Guehring2, Michael O. Zenge2

1Pattern Recognition Lab,, University of Erlangen-Nuremberg, Erlangen, Germany; 2MR Applications & Workflow Development,, Siemens AG, Healthcare Sector, Erlangen, Germany; 3Imaging and Visualization, Siemens Corporate Research, Princeton, NJ, United States

In recent approaches implementing image-based 3D respiratory motion correction for interleaved whole-heart coronary MRA, the interleaves are straightforwardly grouped in equally spaced bins on the respiratory cycle. This allows to obtain undersampled images with minimal intra-bin respiratory motion that can be registered to a reference. However, for bins with a low number of interleaves or non-uniform spatial distribution, image quality is degraded by streaking and undersampling artifacts. Hence, image registration might fail and such bins are usually discarded. In this work an optimized binning approach is proposed which maximizes the uniformity of the distribution of the interleaves. With this method 3D affine motion correction is allowed for all bins.

Abstract PDF

Keywords

according accuracy achieved acquisitions adapted adjacent affine allowed allows although angular anterior applied approximation artifact artifacts azimuthal better binning bins blood body challenges circle clinical coil combination compensation considered contained coronary corporate correct corrected correction correctly critical dataset datasets decreased degraded described development dimensional directly discarded displacement displayed distance distribution dominate either elements equally evaluation eventually example expected extracted feature final full gaps gated goal healthy heart henceforth idea implementing improved improvement induced inferior inferred interleave interleaved interleaves interval intra irregular limited linear matrix maximizing might minimize minimizing motion navigation navigator nevertheless novel optimized oriented outliers overlap pairs pool position positions posterior prepared previously procedure process product proposed quality radial readout readouts reason receiver recently recognition reduction referred reformatted registered registration reliable represented represents respectively respiratory scenario sector self several software space spatial sphere straightforward straightforwardly strategy successful sufficiently superior tenth trajectory transformation unchanged uniform uniformity unit visual visualize volunteers whole