Felix Schwab PhD, Presenter: Nothing to Disclose

Olaf Dietrich PhD, Abstract Co-Author: Nothing to Disclose

Florian Schwarz BS, Abstract Co-Author: Nothing to Disclose

Konstantin Nikolaou MD, Abstract Co-Author: Speakers Bureau, Siemens AG Speakers Bureau, Bracco Group Speakers Bureau, Bayer AG

Carmel Hayes PhD, Abstract Co-Author: Employee, Siemens AG

Daniel Theisen MD, Abstract Co-Author: Nothing to Disclose

Maximilian F. Reiser MD, Abstract Co-Author: Nothing to Disclose

In the context of cardiac functional cine MRI in cases of arrhythmia, faster real-time images can be obtained by optimizing the reference scan used for parallel imaging. The aim of this study is to investigate how k-space-based parallel MRI using a separately acquired reference scan or dynamic time-adaptive reference data (TGRAPPA) improves spatial and temporal resolution of single-shot cardiac MRI and whether segmentation of the left ventricle is possible despite the potential presence of motion artifacts.

5 subjects were measured on a clinical 3 Tesla scanner using a 32-element cardiac array. To fix the signal-to-noise ratio and analyze only the reconstruction artifacts, 3 single-shot real-time (rt) balanced SSFP protocols with equal spatial resolution and acceleration factor (R=5) but different reference scan methods were implemented. These 3 protocols use I) rtINT: an integrated GRAPPA reference scan with temporal resolution Tres=123ms; II) rtTGRAPPA: TGRAPPA (Tres=48.8ms); and III) rtSEP: a separate reference scan (Tres=48.8ms). Results were compared to those of a segmented cine SSFP sequence (segmINT) as standard of reference with integrated GRAPPA and Tres=30.42ms. Acquisition time for all rt protocols is below 30s, for segmented acquisition TA≈7min.

Qualitative assessment of the images shows that reconstruction artifacts are present in rtTGRAPPA and rtSEP, but hardly visible in rtINT. In general, fewer artifacts are observed when using rtSEP than when using rtTGRAPPA, where in some cases the most basal slices did reveal relevant artifacts. Mean values (standard deviation) of the ejection fraction for all techniques are: segmINT=64.0±7.4%, rtINT=38.3±9.7%, rtTGRAPPA=60.6±7.1%, and rtSEP =60.2±6.8%.

Free-breathing real-time cine MRI allows for accurate assessment of global systolic left ventricular function in reduced acquisition time. Motion artifacts arising in TGRAPPA or separate reference scans are shown not to impede the correct segmentation of the left ventricle. This knowledge can be used to obtain faster real-time images with higher spatial resolution in patients with cardiac arrhythmia.

http://media.rsna.org/media/abstract/2011/11015517/11015517_ar1q.jpg

Schwab, F, Dietrich, O, Schwarz, F, Nikolaou, K, Hayes, C, Theisen, D, Reiser, M, Comparison of GRAPPA-based Image Reconstruction Methods for Real-time Functional Cardiac Imaging at 3 T.  Radiological Society of North America 2011 Scientific Assembly and Annual Meeting, November 26 - December 2, 2011 ,Chicago IL. http://archive.rsna.org/2011/11015517.html