Increasing Speed of Acquisition for Pediatric Imaging
Adrienne Campbell-Washburn1
1National Heart, Lung, and Blood Institute, National Institutes of Health, Bethesda, MD, United States

Synopsis

This presentation will provide an overview of fast imaging methods used for pediatric cardiovascular MR. Specifically, it will focus on: New developments in data undersampling paired with advanced image reconstruction, non-Cartesian acquisitions, rapid free-breathing methods, multi-parametric imaging approaches, and applications of machine learning for rapid imaging.

Highlights

To capture the dynamics of cardiac motion, cardiac MRI demands rapid acquisitions. This is even more challenging for pediatric applications where heart rates are faster, higher spatial resolution is required, and patients tend to be less compliant to breath-hold instructions. Moreover, short exams are desirable to reduce, or avoid, anesthesia exposure in pediatric patients (1).

Numerous techniques have been developed for rapid acquisitions during cardiac imaging. This presentation will describe approaches to increase the speed of acquisitions, specifically focusing on the following:

  • Undersampled acquisitions paired with advanced reconstructions
  • Non-Cartesian data acquisitions
  • Continuous free-breathing acquisitions
  • Multi-parametric imaging
  • Machine learning developments for rapid imaging

Recent developments in data undersampling

The most straightforward method to speed up image acquisition is to undersample k-space. The ability to highly undersample data is closely coupled with advanced reconstruction methods. In recent years, data undersampling has moved beyond standard parallel imaging methods as compressed sensing reconstructions have been adopted into clinical workflow.

Compressed sensing reconstructions can generate high quality images from highly undersampled data, which results in a significant reduction of image acquisition time. Compressed sensing reconstructions have been demonstrated to speed up image acquisitions including cine function, 4D flow measurements, and 3D contrast-enhanced MRA, in both pediatric and fetal patients (eg. (2-5)).

One limitation of compressed sensing reconstruction is the prohibitively long reconstruction time, which can constrain the seamless integration of advanced reconstructions into a manageable clinical workflow. Computational developments that use available computation resources and additional vendor support have helped to overcome this problem.

Non-Cartesian acquisitions

Non-Cartesian acquisitions, including spiral and radial data sampling, are also powerful methods to accelerate data acquisitions. These non-Cartesian sampling strategies are inherently efficient and robust to motion, making them well-suited to fast cardiac imaging. Spiral acquisitions have been critical in the development of real-time flow and real-time function measurements (4,6,7), and have also been applied for rapid angiography (8). Radial imaging has been critical in the development of 3D whole heart imaging (9), which has significant implications in pediatric congenital heart disease. More recently, perturbed spiral trajectories have been implemented to further improve compressed sensing reconstruction performance of this non-Cartesian data sampling approach for real-time flow measurements in children (10).

Rapid free-breathing acquisitions

Free-breathing acquisitions have become essential to avoid breath-holds and accelerate MRI acquisitions. Many recent publications describe the use of free-breathing methods combined with respiratory motion correction (MOCO). Free-breathing motion-corrected acquisitions have been implemented for many of the routine CMR sequences including late-gadolinium enhancement(11), which has been validated in children (12), as well as T1-mapping (13), T2 mapping (14), and T2*-mapping (15).

Alternatively, continuous free-breathing acquisitions can be re-binned into multiple motion or contrast states. One prominent example of continuous free-breathing acquisitions is the “XD-GRASP” framework (16). Another approach which is entirely self-gated was developed by Sopra et al for free-running acquisition for 5D whole heart coronary MRA imaging in adults (17). These re-binned approaches typically result in undersampled data sets which are reconstructed using compressed sensing.

One example that combines MOCO and re-binnng with reconstruction parallelized on the cloud is the free-breathing cine imaging developed by Xue et al (18). This approach has a streamlined clinical workflow integration and has been successfully applied for rapid LV and RV functional measurements in children (19,20).

Newer approaches further leverage computational capabilities. “Extreme MRI” is a continuous non-gated 3D cone acquisition that is reconstructed using novel computational approaches (21). A stochastic reconstruction is used to reduce the number of non-uniform FFTs and compressed low rank matrix model of the data is used to reduce memory requirements. This innovative method allows the reconstruction of data that would otherwise be 100s of GB and too computationally cumbersome. Extreme MRI was demonstrated for DCE imaging of the chest in pediatric patients. This approach offers a new framework for acquisition and reconstruction of multi-dimensional data sets.

Multi-parametric imaging

MR fingerprinting (22) is an established technique for simultaneous quantification of multiple MRI parameters, usually T1 and T2. This technique uses a variable acquisition strategy to sensitize the MRI signal to parameters of-interest. Cardiac MRF was initially implemented as a 16 heart-beat breath-hold with a combination of inversion recovery preparations and T2-preparations for measurement of T1, T2 and M0 in a single slice (23). New developments in cardiac MRF have included additional corrections for hardware imperfections (24), simultaneous multi-slice acquisitions (25), inclusion of fat quantification (26), and motion resolved acquisitions (27).

Building on continuous free-breathing acquisitions, MR multitasking (28) was recently developed for motion-resolved multi-contrast imaging. This technique represents physiological motion as multiple dimension and uses low-rank tensor imaging for resolving motion states. This method was demonstrated for T1 mapping, T1-T2 mapping, and myocardial perfusion imaging.

While these multi-parametric imaging methods have not been validated for pediatric patients, they may offer advantages for rapid exams in this patient population.

Machine learning

Recent advancements in artificial intelligence for MRI offer potential to accelerate data acquisition through automated planning; to improve image reconstruction algorithms and speed; and to automate image processing, most commonly for segmentation. Most automated acquisition and image segmentation algorithms, to-date, have been developed using CMR data from adults, but translation to pediatrics and congenital heart disease is expected in the near future (29).

Some algorithms have been specifically developed for image reconstruction in pediatric CMR. These include a deep learning reconstruction applied to remove artifacts in real-time radial MRI in patients with congenital heart disease (30), and a deep learning method for super resolution reconstruction of low-resolution 3D whole heart acquisition in congenital heart disease (31).

Conclusions

Significant developments in rapid pediatric imaging have been made. Ongoing research in undersampled acquisitions, advanced reconstructions, non-Cartesian sampling, free-breathing implementations, multi-parametric imaging and machine learning will continue to impact the efficiency of the pediatric CMR workflow.

Acknowledgements

No acknowledgement found.

References

  1. Dong SZ, Zhu M, Bulas D. Techniques for minimizing sedation in pediatric MRI. J Magn Reson Imaging 2019;50(4):1047-1054.
  2. Cheng JY, Hanneman K, Zhang T, Alley MT, Lai P, Tamir JI, Uecker M, Pauly JM, Lustig M, Vasanawala SS. Comprehensive motion-compensated highly accelerated 4D flow MRI with ferumoxytol enhancement for pediatric congenital heart disease. J Magn Reson Imaging 2016;43(6):1355-1368.
  3. Roy CW, Seed M, Kingdom JC, Macgowan CK. Motion compensated cine CMR of the fetal heart using radial undersampling and compressed sensing. J Cardiovasc Magn Reson 2017;19(1):29.
  4. Steeden JA, Kowalik GT, Tann O, Hughes M, Mortensen KH, Muthurangu V. Real-time assessment of right and left ventricular volumes and function in children using high spatiotemporal resolution spiral bSSFP with compressed sensing. J Cardiovasc Magn Reson 2018;20(1):79.
  5. Moghari MH, Annese D, Geva T, Powell AJ. Three-dimensional heart locator and compressed sensing for whole-heart MR angiography. Magn Reson Med 2016;75(5):2086-2093.
  6. Kowalik GT, Steeden JA, Pandya B, Odille F, Atkinson D, Taylor A, Muthurangu V. Real-time flow with fast GPU reconstruction for continuous assessment of cardiac output. J Magn Reson Imaging 2012;36(6):1477-1482.
  7. Steeden JA, Atkinson D, Taylor AM, Muthurangu V. Assessing vascular response to exercise using a combination of real-time spiral phase contrast MR and noninvasive blood pressure measurements. J Magn Reson Imaging 2010;31(4):997-1003.
  8. Kowalik GT, Steeden JA, Atkinson D, Montalt-Tordera J, Mortensen KH, Muthurangu V. Golden ratio stack of spirals for flexible angiographic imaging: Proof of concept in congenital heart disease. Magn Reson Med 2019;81(1):90-101.
  9. Nam S, Akcakaya M, Basha T, Stehning C, Manning WJ, Tarokh V, Nezafat R. Compressed sensing reconstruction for whole-heart imaging with 3D radial trajectories: a graphics processing unit implementation. Magn Reson Med 2013;69(1):91-102.
  10. Kowalik GT, Knight D, Steeden JA, Muthurangu V. Perturbed spiral real-time phase-contrast MR with compressive sensing reconstruction for assessment of flow in children. Magn Reson Med 2019.
  11. Kellman P, Larson AC, Hsu LY, Chung YC, Simonetti OP, McVeigh ER, Arai AE. Motion-corrected free-breathing delayed enhancement imaging of myocardial infarction. Magn Reson Med 2005;53(1):194-200.
  12. Olivieri L, Cross R, O'Brien KJ, Xue H, Kellman P, Hansen MS. Free-breathing motion-corrected late-gadolinium-enhancement imaging improves image quality in children. Pediatr Radiol 2016;46(7):983-990.
  13. Chow K, Yang Y, Shaw P, Kramer CM, Salerno M. Robust free-breathing SASHA T1 mapping with high-contrast image registration. J Cardiovasc Magn Reson 2016;18(1):47.
  14. Basha TA, Bellm S, Roujol S, Kato S, Nezafat R. Free-breathing slice-interleaved myocardial T2 mapping with slice-selective T2 magnetization preparation. Magn Reson Med 2016;76(2):555-565.
  15. Kellman P, Xue H, Spottiswoode BS, Sandino CM, Hansen MS, Abdel-Gadir A, Treibel TA, Rosmini S, Mancini C, Bandettini WP, McGill LA, Gatehouse P, Moon JC, Pennell DJ, Arai AE. Free-breathing T2* mapping using respiratory motion corrected averaging. J Cardiovasc Magn Reson 2015;17(1):3.
  16. Feng L, Axel L, Chandarana H, Block KT, Sodickson DK, Otazo R. XD-GRASP: Golden-angle radial MRI with reconstruction of extra motion-state dimensions using compressed sensing. Magn Reson Med 2016;75(2):775-788.
  17. Di Sopra L, Piccini D, Coppo S, Stuber M, Yerly J. An automated approach to fully self-gated free-running cardiac and respiratory motion-resolved 5D whole-heart MRI. Magn Reson Med 2019;82(6):2118-2132.
  18. Xue H, Inati S, Sørensen TS, Kellman P, Hansen MS. Distributed MRI reconstruction using gadgetron-based cloud computing. Magn Reson Med 2014.
  19. Cross R, Olivieri L, O'Brien K, Kellman P, Xue H, Hansen M. Improved workflow for quantification of left ventricular volumes and mass using free-breathing motion corrected cine imaging. J Cardiovasc Magn Reson 2016;18:10.
  20. Merlocco A, Olivieri L, Kellman P, Xue H, Cross R. Improved Workflow for Quantification of Right Ventricular Volumes Using Free-Breathing Motion Corrected Cine Imaging. Pediatr Cardiol 2019;40(1):79-88.
  21. Ong F, Zhu X, Cheng JY, Johnson K, Larson P, Vasanawala SS, Lustig M. Extreme MRI: Large-Scale Volumetric Dynamic Imaging from Continuous Non-Gated Acquisitions. . arXiv 2019.
  22. Ma D, Gulani V, Seiberlich N, Liu K, Sunshine JL, Duerk JL, Griswold MA. Magnetic resonance fingerprinting. Nature 2013;495(7440):187-192.
  23. Hamilton JI, Jiang Y, Chen Y, Ma D, Lo WC, Griswold M, Seiberlich N. MR fingerprinting for rapid quantification of myocardial T1 , T2 , and proton spin density. Magn Reson Med 2017;77(4):1446-1458.
  24. Hamilton JI, Jiang Y, Ma D, Lo WC, Gulani V, Griswold M, Seiberlich N. Investigating and reducing the effects of confounding factors for robust T1 and T2 mapping with cardiac MR fingerprinting. Magn Reson Imaging 2018;53:40-51.
  25. Hamilton JI, Jiang Y, Ma D, Chen Y, Lo WC, Griswold M, Seiberlich N. Simultaneous multislice cardiac magnetic resonance fingerprinting using low rank reconstruction. NMR Biomed 2019;32(2):e4041.
  26. Jaubert O, Cruz G, Bustin A, Schneider T, Lavin B, Koken P, Hajhosseiny R, Doneva M, Rueckert D, Botnar RM, Prieto C. Water-fat Dixon cardiac magnetic resonance fingerprinting. Magn Reson Med 2020;83(6):2107-2123.
  27. Jaubert O, Cruz G, Bustin A, Schneider T, Koken P, Doneva M, Botnar RM, Prieto C. Cardiac motion resolved Magnetic Resonance Fingerprinting with joint reconstruction: jMORE-MRF. Proceedings of the 27th Annual Meeting of the International Society for Magnetic Resonance in Medicine 2019:0808.
  28. Christodoulou AG, Shaw JL, Nguyen C, Yang Q, Xie Y, Wang N, Li D. Magnetic resonance multitasking for motion-resolved quantitative cardiovascular imaging. Nat Biomed Eng 2018;2(4):215-226.
  29. Arafati A, Hu P, Finn JP, Rickers C, Cheng AL, Jafarkhani H, Kheradvar A. Artificial intelligence in pediatric and adult congenital cardiac MRI: an unmet clinical need. Cardiovasc Diagn Ther 2019;9(Suppl 2):S310-S325.
  30. Hauptmann A, Arridge S, Lucka F, Muthurangu V, Steeden JA. Real-time cardiovascular MR with spatio-temporal artifact suppression using deep learning-proof of concept in congenital heart disease. Magn Reson Med 2019;81(2):1143-1156.
  31. Steeden JA, Quail M, Gotschy A, Hauptmann A, Arridge S, Jones R, Muthurangu V. Rapid Whole-Heart CMR with Single Volume Super-resolution. arXiv 2019.
Proc. Intl. Soc. Mag. Reson. Med. 28 (2020)