Haben Berhane¹, Michael Scott¹, Ashitha Pathrose¹, Patrick McCarthy², Chris Malaisrie², Bradley Allen³, Ryan Avery³, and Michael Markl^1
1Biomedical Engineering, Northwestern University, Chicago, IL, United States, ²Cardiac Surgery, Northwestern University, Chicago, IL, United States, ³Radiology, Northwestern University, Evanston, IL, United States

4D flow MRI provides comprehensive assessment of cardiovascular hemodynamics. However, the current clinical usage of standard 4D flow MRI is hindered by long scan times and extensive, time-consuming post-processing such as eddy current corrections, noise masking, and 3D vessel segmentation. We seek to address this by developing a fully automated image reconstruction and post-processing pipeline for highly-accelerated aortic 4D flow MRI (R=5.7-10.2). The fully automated pipeline was shown to provide good-to-excellent agreement in quantitative and hemodynamic measures to conventional 4D flow MRI (GRAPPA, R=2) with manual post-processing. Additionally, the pipeline only requires <10 minutes compared to 20 minutes manually.

Ruponti Nath¹, Amirkhosro Kazemi¹, Marcus Stoddard², and Amir Amini^1
1ECE, University of Louisville, Louisville, KY, United States, ²Cardiovascular Division, Robley Rex VA Medical Centre, Louisville, KY, United States

We propose a novel deep learning based approach to estimate pressure drop inside a stenotic valve from 4D Flow MRI velocities. A neural network architecture learns the  relationship of three directional velocities and predicts pressure as output.  The network was tested on real 4D flow MRI data of aortic valvular flow both  in-vitro and in vivo. Estimated in-vitro pressure drop by proposed method  shows (3-5)% relative pressure drop error with corresponding CFD pressure. Estimated In-vivo pressure drop was also compared with doppler Echocardiography and simplified and modified Bernoulli at peak systole timepoints in 10 patients with aortic stenosis.

Ozden Kilinc¹, Stanley Chu¹, Elizabeth K. Weiss^1,2, Justin Baraboo^1,2, Anthony Maroun¹, Ning Jin³, Kelvin Chow^1,4, Xiaoming Bi⁴, Rachel Davids⁴, Chris Mehta⁵, S. Chris Malaisrie⁵, Andrew Hoel⁶, Michael Markl^1,2, and Bradley D. Allen^1
1Department of Radiology, Feinberg School of Medicine, Northwestern University, Chicago, IL, United States, ²Department of Biomedical Engineering, Northwestern University, Chicago, IL, United States, ³Cardiovascular MR R&D, Siemens Medical Solutions USA, Inc., Cleveland, OH, United States, ⁴Cardiovascular MR R&D, Siemens Medical Solutions USA, Inc., Chicago, IL, United States, ⁵Division of Cardiac Surgery, Feinberg School of Medicine, Northwestern University, Chicago, IL, United States, ⁶Division of Vascular Surgery, Feinberg School of Medicine, Northwestern University, Chicago, IL, United States

4D flow MRI is an established quantitative MR imaging technique for evaluating complex blood flow patterns. However, clinical applications of standard 4D flow MRI technique is limited by long scan times associated with multidimensional imaging. Compressed sensing (CS) acceleration can significantly reduce scan time through dramatic data undersampling, but it is not clear how this undersampling impacts assessment of flow patterns in type B aortic dissection (TBAD). In this study, we investigate performance of highly accelerated CS 4D flow MRI at two different acceleration factor levels (R=7.7 and 10.2) to evaluate aortic flow dynamics in volunteers and patients with TBAD.

Jianing Tang¹, Samantha J Ma², and Lirong Yan^1,3
1USC Mark and Mary Stevens Neuroimaging and Informatics Institute, Keck School of Medicine, University of Southern California, Los Angeles, CA, United States, ²Siemens Medical Solutions USA, Inc., Los Angeles, CA, United States, ³Department of Neurology, Keck School of Medicine, University of Southern California, Los Angeles, CA, United States

A dual VENC strategy was applied in 2D high-resolution phase-contrast MRI to improve the sensitivity of arterial pulsatility measurement in cerebral perforating arteries at 7T. Compared to a single VENC of 40cm/s, significant increase in the number of detected perforators was observed using a dual-VENC approach with VENC of 20cm/s and 40cm/s. Reproducibility of arterial pulsatility measurement in perforating arteries was evaluated in test-retest experiments. Good agreement was achieved between test-retest pulsatility measurements. 

Ang Zhou¹, Huiming Dong¹, Joseph Leach¹, Chengcheng Zhu², David Saloner¹, Michael Hope¹, and Dimitrios Mitsouras^1
1Department of Radiology and Biomedical Imaging, University of California San Francisco, San Francisco, CA, United States, ²Department of Radiology, University of Washington, Seattle, WA, United States

Abdominal Aortic Aneurysms (AAAs) are common in older men. Guidelines recommend repair at an aneurysm diameter greater than 5.5 cm, but some smaller aneurysms experience catastrophic rupture. Aneurysm features beyond diameter have been explored as markers of disease progression. We used DCE MRI to explore the inhomogeneity of contrast kinetics, including K^trans, in the aneurysm wall and assess if these markers independently predict AAA progression. Significant differences in DCE metrics were identified around the AAA circumference. Area under enhancement-time curve and enhancement slope correlated with growth rate but not diameter, suggesting that DCE may provide diameter-independent information regarding AAA risk.

Haben Berhane¹, Jackson E Moore¹, Ann Ragin¹, Eric Russell², Fan Caprio³, Susanne Schnell⁴, Sameer A Ansari², and Michael Markl^1
1Biomedical Engineering, Northwestern University, Chicago, IL, United States, ²Radiology, Northwestern University, Chicago, IL, United States, ³Neurology, Northwestern University, Chicago, IL, United States, ⁴Radiology, University of Greifswald, Greifswald, Germany

4D flow MRI provides a comprehensive assessment of hemodynamics through the 3D visualization and quantification of blood flow. However, its current clinical usage is hindered by long scan times. Recent developments have shown deep learning to be highly effective in accelerating image reconstruction, but no study has shown the its effectiveness in reconstructing highly undersampled cerebrovascular 4D flow MRI data. As such, we developed a CNN for the reconstruction of kt-accelerated intracranial dual-venc 4D flow MRI (R=5). We found that the CNN showed excellent SSIM values (0.94[0.93-0.95]) and moderate-to-good net flow and peak velocity agreement with the conventionally reconstructed data.

Amirkhosro Kazemi^1,2, Sean Callahan^1,2, Ruponti Nath^1,2, Marcus Stoddard^2,3, and Amir A. Amini^1,2
1Electrical and Computer Engineering, University of Louisville, Louisville, KY, United States, ²Robley Rex VA Medical Center, Louisville, KY, United States, ³Cardiovascular Division, University of Louisville, Louisville, KY, United States

We propose vortex ring as a new indicator of the location of maximum pressure drop along a phantom model of arterial stenosis. We investigated details of the flow structure of post stenotic jet with presence of the vortex ring in stenotic phantoms with 87% reduction in area using CFD velocities and 4D flow MRI. The pressure, velocity, and vorticity fields were quantitatively analyzed with the presence of the vortex ring in the spatiotemporal domains. It was found that the location of the vortex ring is associated with the location of the maximum pressure drop, axial velocity, and vorticity magnitude.

Takashi Fujiwara¹, Erin K Englund¹, Mehdi H Moghari¹, Brian Fonseca², Lorna P Browne¹, and Alex J Barker^1,3
1Department of Radiology, Children's Hospital Colorado, University of Colorado Anschutz Medical Campus, Aurora, CO, United States, ²Department of Pediatrics, Children's Hospital Colorado, University of Colorado Anschutz Medical Campus, Aurora, CO, United States, ³Department of Bioengineering, University of Colorado Anschutz Medical Campus, Aurora, CO, United States

The ratio of pulmonary to systemic blood flow (Qp/Qs) is a clinically important index to estimate the existence and amount of cardiovascular shunt flow and to determine the need for surgical intervention for congenital heart disease patients. Accelerated 4D flow MRI is capable of measuring Qp/Qs in a single acquisition, but its accuracy and reliability are not well established. We compared conventional 2D phase-contrast MRI and two 4D flow MRI acquisitions with different acceleration techniques to investigate accuracy of Qp/Qs quantification with accelerated 4D flow in adult healthy volunteers (N=10). Consequently, we found no statistical differences in Qp/Qs.

Chi Zhang^1,2, Ludovica Romanin^3,4, Davide Piccini^3,4, Steen Moeller², Matthias Stuber^3,5, and Mehmet Akçakaya^1,2
1Department of Electrical and Computer Engineering, University of Minnesota, Minneapolis, MN, United States, ²Center for Magnetic Resonance Research, University of Minnesota, Minneapolis, MN, United States, ³Department of Diagnostic and Interventional Radiology, Lausanne University Hospital and University of Lausanne, Lausanne, Switzerland, ⁴Advanced Clinical Imaging Technology, Siemens Healthcare AG, Bern, Switzerland, ⁵Center for Biomedical Imaging, Lausanne, Switzerland

SImilarity-driven Multi-dimensional Binning Algorithm (SIMBA) is a recently proposed technique for identifying motion-consistent clusters in free-running whole-heart MRA acquisitions, where the clusters are subsequently reconstructed using conventional approaches. Physics-guided deep learning (PG-DL) reconstruction has gained popularity with its superior performance at higher acceleration rates and may further improve the image quality of free-running whole-heart MRA in conjunction with SIMBA. However, PG-DL is difficult to apply to 3D non-Cartesian acquisitions due to hardware limitations. In this work we enable distributed memory-efficient PG-DL for large-scale SIMBA datasets, showing preliminary results in which PG-DL improves image quality compared to conventional methods.

Erin K Englund¹, Takashi Fujiwara¹, Kelly B Jarvis², Michael Markl², Daniel Enge³, Jane E.B. Reusch⁴, and Alex J Barker^1
1Radiology, University of Colorado Anschutz Medical Campus, Aurora, CO, United States, ²Radiology, Northwestern University, Chicago, IL, United States, ³Bioengineering, University of Colorado Anschutz Medical Campus, Aurora, CO, United States, ⁴Endocrinology, University of Colorado Anschutz Medical Campus, Aurora, CO, United States

Pulse wave velocity (PWV) is an important measure of cardiovascular health, related to vascular stiffness. Aortic PWV can be evaluated with a variety of MRI methods, however a direct comparison of PWV results quantified from 4D flow, 2D phase contrast (PC), and 1D projection PC MRI has not been performed. Here, PWV derived from 4D, 2D, and 1D-PC MRI acquisitions and associated analysis strategies were compared in the aortic arch. Relative agreement was observed among methods. 2D-PC PWV derived from the flow-area approach had substantial variability, likely due to the limited spatial/temporal resolution, and user-dependence of analysis.

Kaiyu Zhang¹, Zhensen Chen², Li Chen¹, Gador Canton¹, Duygu Baylam Geleri¹, Kristi Pimentel¹, Niranjan Balu¹, Thomas Hatsukami¹, and Chun Yuan^1
1Vascular Imaging Lab and BioMolecular Imaging Center, Department of Radiology, University of Washington, Seattle, WA, United States, ²Institute of Science and Technology for Brain-Inspired Intelligence, Fudan University, Shanghai, China

Association between brain tissue level perfusion and cognitive decline has been previously documented. However, contribution of blood flow and artery patency in medium-to-large arteries and its relation to tissue level perfusion and cognition remain unclear. In this study, 3D-TOF, ASL, and SNAP, an MRA technique sensitive to slow blood flow, were acquired at baseline and 12-months follow-up. A vascular map construction software was used to measure changes in medium-to-large arterial features. Changes in SNAP artery length and branch number during follow-up were significantly associated with changes in cognitive function, while no such association was found with ASL tissue level perfusion.

Yitong Yang¹, Jackson Hair¹, Jerome Yerly², Davide Piccini³, Matthias Stuber⁴, and John Oshinski^1,5
1Biomedical Engineering, Emory University/Georgia Institute of Technology, Atlanta, GA, United States, ²Diagnostic and Interventional Radiology, Lausanne University Hospital, Lausanne, Switzerland, ³Siemens Healthcare, Lausanne, Switzerland, ⁴Lausanne University Hospital, Lausanne, Switzerland, ⁵Radiology, Emory University School of Medicine, Atlanta, GA, United States

Under-sampled reconstruction of retrospectively gated 5D free-running acquisition is time-consuming and computationally intensive. In this work, a limited reconstruction method is proposed which provides an efficient way to compute a static cardiac volume from free-breathing, ECG-free 5D free-running CMR acquisitions. Structural similarity index measure is used to compare the limited reconstructions with the fully reconstructed image at the physiologic state of interest. The proposed limited reconstruction method achieves a SSIM of 0.9 using only 22% of the full reconstruction time and an SSIM of 0.95 using 44% of the full reconstruction.  

Kwang Eun Jang^1,2, Dwight G. Nishimura², and Shreyas S. Vasanawala^3
1Department of Bioengineering, Stanford University, Stanford, CA, United States, ²Magnetic Resonance Systems Research Lab (MRSRL), Department of Electrical Engineering, Stanford University, Stanford, CA, United States, ³Department of Radiology, Stanford University, Stanford, CA, United States

For free-breathing 3D coronary angiography, we consider a new multiphase acquisition that consists of a series of short imaging blocks to reduce motion blur in the acquisition window. However, this approach poses two challenges: (1) because the alterations of readout gradient waveforms become large due to the shortened imaging window, eddy current artifacts appear in steady-state free precession imaging, and (2) motion among cardiac phases needs to be addressed. In this work, we propose a new ordering scheme to establish smooth transitions between conic interleaves. Phase-to-phase motion is estimated using self image-based navigators and corrected by motion-based reconstruction.

Fei Xiong¹, Ning Jin², Anna Lena Emrich ³, SarahRose Hall³, Jeanie Marie Ruddy³, Jonathan Aldinger⁴, Gabrielle Young ⁴, U Joseph Schoepf⁴, Daniel Giese⁵, Tilman Emrich ⁴, and Akos Varga-Szemes^4
1Cardiovascular MR R&D, Siemens Medical Solutions USA, Inc, Charleston, SC, United States, ²Cardiovascular MR R&D, Siemens Medical Solutions USA, Inc, Chicago, IL, United States, ³Department of Surgery, Medical University of South Carolina, Charleston, SC, United States, ⁴Department of Radiology and Radiological Science, Medical University of South Carolina, Charleston, SC, United States, ⁵MR Cardiovascular Predevelopment, Siemens Healthcare GmbH, Erlangen, Germany

We aim to demonstrate the feasibility of a highly accelerated, free-breathing RT PC technique empowered by compressed sensing (CS) and shared velocity encoding (SVE) which allows a scan time less than 5s and fast inline image reconstruction.

Stanley Chu¹, Ozden Kilinc ², Elizabeth Weiss², Kelly Jarvis², Christopher Mehta ², Chris Malaisrie ², Andrew Hoel², Maurice Pradella ², Bradley Allen², and Michael Markl^2,3,4
1Radiology, Northwestern University, Chicago, IL, United States, ²Northwestern University, Chicago, IL, United States, ³Siemens Healtineers, Erlangen, Germany, ⁴Cryolife Inc., Kennesaw, GA, United States

Increasing aorta diameter in type B aortic dissection (TBAD) patients is associated with adverse clinical outcomes. Our study evaluates true and false lumen (TL, FL) hemodynamics (kinetic energy (KE), maximum velocity (MV), forward flow (FF), and reverse flow (RF) in baseline and follow-up thoracic aorta 4D flow (4DF) MRI. Baseline aortic diameter does not correlate with aortic growth rate, while FL-total KE, FF, and RF correlate negatively. At follow-up, larger increases in FL-total KE, total FF, KE ratio and FL-mean MV correlate positively with aortic growth rate.  Investigating hemodynamic changes in follow-up 4DF may be useful for evaluating TBAD patients.