Translational Scientific Session: Fast Cardiac Imaging

Tuesday 23 April 2013

Hui Xue¹, Peter Kellman², Andrew E. Arai², and Michael Schacht Hansen^3
1Siemens Corporate Research, Princeton, New Jersey, United States, ²National Institutes of Health, National Heart, Lung and Blood Institute, Bethesda, Maryland, United States, ³National Institutes of Health, Bethesda, Maryland, United States

While the methods do exist for retrospectively reconstructing high temporal resolution cardiac cine images from real-time free-breathing acquisition, they are not clinical applicable, mainly due to the lengthy acquisition (>30s) required for good image quality. To significantly shorten the required acquisition duration, we propose to utilize the non-linear k-space reconstruction for shortened retrospective real-time cine imaging. In-vivo test shows proposed workflow can achieve good image quality comparable to segmented cine with only 10s of real-time data, which largely improves the usability of retrospective reconstruction of real-time cine.

Haris Saybasili^1,2, Daniel A. Herzka³, Kestutis Barkauskas⁴, Nicole Seiberlich⁴, and Mark A. Griswold^1,4
1Radiology, Case Western Reserve University, Cleveland, OH, United States, ²Biomedical Engineering, Johns Hopkins University School of Medicine, Baltimore, MD, United States, ³Biomedical Engineering, Johns Hopkins University, Baltimore, MD, United States, ⁴Biomedical Engineering, Case Western Reserve University, Cleveland, OH, United States

A hybrid (CPU- and GPU-based), faster-than-acquisition through-time radial GRAPPA reconstruction was previously demonstrated for 15 coil, rate 8 (16 projections, 128x128 matrix) radial datasets. However, because of the increased number of acquisition coils on modern scanners, single-GPU radial GRAPPA reconstructions were challenging for low-latency, real-time MRI with high number of acquisition coils. We present a completely automated, multi-node (group of workstations connected via network), multi-GPU radial GRAPPA implementation that can reconstruct 32-coil 16 projection radial datasets much faster than acquisition. Images from 32 coil, 16x256 data (acquisition time 42ms/frame) were reconstructed in 11.2 ms/frame using four nodes (two GPUs on each).

Robert Manka^1,2, Rolf Gebker³, Lukas Wissmann¹, Roy Jogiya⁴, Manish Motawani⁵, Michael Frick⁶, Sebastian Reinartz⁶, Bernhard Schnackenburg⁷, Eike Nagel⁴, Sven Plein⁵, and Sebastian Kozerke^1
1University and ETH Zurich, Zurich, Switzerland, ²University Hospital Zurich, Zurich, Switzerland, ³German Heart Institute Berlin, Berlin, Germany, ⁴King's College London, London, United Kingdom, ⁵University of Leeds, Leeds, United Kingdom, ⁶University Hospital RWTH Aachen, Aachen, Germany, ⁷Philips Healthcare, Clinical Sciences, Berlin, Germany

Cardiac magnetic resonance imaging enables noninvasive assessment of myocardial perfusion. However, standard 2D multi-slice CMR perfusion techniques only provide limited coverage and hence prohibit computation of myocardial ischemic burden. Recently, 3D CMR perfusion has proven highly diagnostic in two single-center studies. In this preliminary assessment of our multi-center study 3D CMR perfusion imaging proved highly diagnostic for the detection of functionally significant CAD as defined by FFR.

Ganesh Adluru¹ and Edward V.R. DiBella^1
1Radiology, University of Utah, Salt lake city, Utah, United States

Ungated cardiac perfusion MRI can offer a simple and an efficient means of data acquisition especially for patients with arrhythmias. However image quality of the reconstructed images can be affected from cardiac and respiratory motion. Here we propose a new reconstruction framework that can offer improved image quality when there is significant inter-frame motion in the data. By using model-reference images and a bi-directional diffeomorphic registration tool, motion is estimated and suppressed leading to improved signal sparsity in the temporal dimension. Results presented using in-vivo data show the promise of the proposed framework.

Mehmet Akçakaya¹, Tamer A. Basha¹, Raymond H. Chan¹, Warren J. Manning¹, and Reza Nezafat^1
1Beth Israel Deaconess Medical Center, Harvard Medical School, Boston, MA, United States

Despite significant advances over the past decade, coronary MRI still faces major challenges, including lengthy acquisition time, low signal-to-noise-ratio (SNR), and limited spatial resolution. Recently, it was shown that higher spatial resolution in sub-mm range may improve the detection of stenosis in targeted coronary MRI. However, higher spatial resolution also results in lower SNR and increased acquisition time, hindering its practicality. In this study, we sought to investigate whether an advanced compressed sensing technique can be applied to 6-fold accelerated sub-mm whole-heart coronary MRI, and compared it to parallel imaging, the state-of-the-art accelerated imaging technique in coronary MRI.

R. Reeve Ingle¹, Holden H. Wu², Nii Okai Addy¹, Joseph Y. Cheng¹, Bob S. Hu^1,3, and Dwight G. Nishimura^1
1Electrical Engineering, Stanford University, Stanford, California, United States, ²Department of Radiology, University of California, Los Angeles, California, United States, ³Palo Alto Medical Foundation, Palo Alto, California, United States

A new application of an autofocusing nonrigid motion correction technique is demonstrated for coronary artery imaging using a free-breathing 3D cones non-Cartesian sequence. Phantom, volunteer, and patient studies are conducted to analyze the performance of the autofocusing motion correction technique. The proposed technique is shown to yield significant improvements in the depiction of the coronary arteries in volunteer and patient studies.

Jing Liu¹, Petter Dyverfeldt², Michael Hope¹, and David A. Saloner^1
1University of California San Francisco, San Francisco, CA, United States, ²Linkoping University, Linkoping, Sweden

This study investigated time-resolved variable-density random sampling scheme for highly accelerating 4D flow cardiovascular imaging. With our proposed Sample-selective sliding window reconstruction for TIme-Resolved Random UndersamPling (STIRRUP) method, high quality flow images were achieved. STIRRUP reconstruction results were also used as improved initial solution for the parallel imaging method SPIRIT (iiSPIRIT), achieving accurate and robust flow images. Our qualitative and quantitative evaluations of the results demonstrated the potential of our methods to achieve highly accelerated 4D flow imaging with maintained accuracy.

Alejandro Roldán-Alzate¹, Alex Frydrychowicz², Oliver Wieben^1,3, and Scott B. Reeder^1,3
1Radiology, University of Wisconsin, Madison, WI, United States, ²Radiology, Universitätsklinikum Schleswig-Holstein, Lübeck, Germany, ³Medical Physics, University of Wisconsin, Madison, WI, United States

Portal hypertension is an end-stage complication of cirrhosis that leads to dramatic and complex alterations in the hemodynamics of the liver. Phase contrast 4D-flow MRI methods hold great promise to overcome the challenges associated with comprehensive non-invasive flow measurements in the abdomen. This study demonstrates the ability of radial 4D flow MRI to quantify changes in hepatic and mesenteric flow in patients with portal hypertension who are undergoing a meal challenge. Expect flow changes in portal venous, hepatic arterial and mesenteric vasculature were observed, including altered flow patterns, and flow in varices and the azygos vein.

Jerome Yerly^1,2, Richard Frayne^2,3, and R. Marc Lebel^2,4
1Electrical and Computer Engineering, University of Calgary, Calgary, AB, Canada, ²Seaman Family MR Research Centre, Foothills Medical Centre, Calgary, AB, Canada, ³Radiology and Clinical Neurosciences, University of Calgary, Calgary, AB, Canada, ⁴Applied Science Laboratory, GE Healthcare, Calgary, AB, Canada

Acute ischemic stroke requires extremely rapid imaging methods in order to effectively diagnose the occluded vessel(s) and provide timely intervention. Time-of-flight (TOF) angiography is central to this diagnosis. We investigate methods for optimizing TOF imaging given a 30 second acquisition window and demonstrate that highly accelerated methods provide significantly improved image quality relative to fully sampled data.

Robert X. Smith¹, Lirong Yan¹, Stanislas Rapacchi², Yiqun Xue³, Subashini Srinivasan², Daniel B. Ennis⁴, Peng Hu⁴, Hee Kwon Song³, and Danny J.J. Wang^1,2
1Neurology, UCLA, Los Angeles, CA, United States, ²Radiology, UCLA, Los Angeles, CA, United States, ³Radiology, University of Pennsylvania, Philadelphia, PA, United States, ⁴Radiology, University of California Los Angeles, Los Angeles, CA, United States

A new method for non-contrast enhanced time-resolved dynamic MRA (dMRA) is presented. A 3D stack-of-stars dynamic radial acquisition with golden angle view increment is combined with k-space weighted image contrast (KWIC) to achieve high spatial and temporal resolution dMRA images with significantly reduced scan time when compared with conventional methods. The acquired 4D dMRA demonstrates excellent image quality without discernable temporal blurring compared to standard Cartesian based dMRA approach.