Synopsis
Keywords: Image acquisition: Multiparametric, Image acquisition: MR Fingerprinting, Image acquisition: Motion Correction
Multi-parametric mapping techniques are gaining attention for their potential to overcome some of magnetic resonance imaging’s clinical limitations. The major advantages of multi-parametric mapping lie within their ability to simultaneously capture multiple “features” such as motions, T1/T2/T2* relaxation, etc. This talk aims to introduce advanced multi-parametric mapping techniques, describing key concepts that many of these techniques share to produce co-registered, high quality MR images in less time and with less requirements for MR technologists.
Introduction
Quantitative magnetic resonance imaging (MRI) is a technique that provides important insights into the human body and is used to manage various diseases. This imaging modality offers a multitude of contrast generating mechanisms, including T1, T2, T2*, ECV, and fat fraction, which reflect both the macrostructure and microstructure of tissues. The conventional MRI procedure, however, acquires qualitative images with limited accuracy due to coil positioning and inter-scanner and inter-reader variability. To overcome this issue, several accelerated techniques have been developed, leading to the emergence of quantitative MRI, which has become increasingly important in clinical applications1.
Moving organs, such as the heart, lungs, and abdominal organs, introduce severe motion artifacts, which degrade image quality. One method to reduce cardiac-motion-induced artifacts is ECG gating, but this significantly increases scan time, particularly for quantitative MRI, which requires multiple contrast weightings. Respiratory gating is another option for respiratory motion, but it also results in lengthy scans. Patients are often asked to hold their breath for 10-20 secs to limit respiratory motion, but this method is not always reliable and can reduce the resolution of quantitative maps, making diagnoses more difficult. To overcome these limitations, several motion-resolved multi-parametric mapping techniques have been developed, and these techniques will be discussed during the talk.HD-PROST
High-dimensionality Undersampled Patch-based Reconstruction (HD-PROST)2 has been validated in dynamic imaging in brain, cardiac and abdominal organs. A recent HD-PROST application, which recovers 3D cine, T1, and T2, was tested in both phantoms and 10 healthy volunteers1-3. The results showed that HD-PROST gave comparable results for ejection fraction (EF) as well as highly precise T1 and T2 measurements when compared to standard methods. Another study used a free-breathing 3D whole-heart sequence to visualize the coronary vasculature, and the images had an excellent agreement in visualizing the coronary vasculature and its distal segments compared with the fully sampled reference image. These images had good quality despite shorter scan times. HD-PROST has also been applied to reconstruct water- and fat-suppressed LGE images. Although the HD-PROST images were of diagnostic quality in 18/20 datasets with strong agreement in the location of enhancement when compared to standard LGE images, residual cardiac motion was still present. Further clinical studies are needed to help standardize the tuning of hyper-parameters for cardiac applications that intend to use HD-PROST or any of the sparse-sampling methods previously described to reduce the noise and aliasing artifacts caused by over-regularization.MR Multitasking
MR Multitasking1 is a multiparametric method that can capture cardiac and respiratory motion as well as myocardial relaxation properties,without ECG triggering. It has broder application from brain, cardiac, to abdominal organs. It captures the physiological and bulk motion by frequently acquiring the center k-space aream which forms the training data. The contrast dynamics and motion information are extracted from the training data. MR Multitasking has been used to measure 2D myocardial T1 and extracellular volumes (ECV), 2D/3D myocardial T1 and T2 joint mapping, 2D myocardial T1, T2, T2*, and fat-fraction joint mapping, and 3D carotid plaques and aortic strain in patients with thoracic aortic disease4-7. These studies demonstrated that the Multitasking framework can produce high quality images with reproducible values that are in good agreement with reference values. It is worth noting that despite the promising results from the proof-of-concept studies, the use of Multitasking sequence in clinical settings is still limited due to several challenges. These challenges include the need for optimization of acquisition parameters, such as spatial and temporal resolution, and the need for further development and validation of the reconstruction algorithms to ensure accurate and reproducible results.MR Fingerprinting
MR Fingerprinting has been an effective parametric mapping technique and has been used widely in cardiac and abdominal organs. The latest development of this technique aimed to improve the efficiency of cardiac magnetic resonance fingerprinting (MRF) by shortening the breathhold and diastolic acquisition window while performing simultaneous T1, T2, and proton spin density (M0) mapping, in order to reduce motion artifacts8. The authors developed a new reconstruction method called DIP-MRF8, which combines low-rank subspace modeling with a deep image prior. DIP-MRF uses a system of neural networks to generate spatial basis images and quantitative tissue property maps, with training performed using only the undersampled k-space measurements from the current scan. In simulations and experiments using a standardized phantom, healthy subjects, and patients with suspected cardiomyopathy, DIP-MRF demonstrated decreased normalized root-mean-square error compared to dictionary matching and a sparse and locally low rank (SLLR-MRF) reconstruction. The approach also yielded strong correlation with T1 and T2 reference values in the phantom and better suppression of noise and aliasing artifacts in vivo, with lower intersubject and intrasubject variability compared to dictionary matching and SLLR-MRF. Overall, DIP-MRF provides an innovative and promising approach for motion-resolved cardiac MRF tissue property mapping, which does not require pre-training with in vivo scans and enables a shortened breathhold and acquisition window.Acknowledgements
No acknowledgement found.References
1. Eyre, K., Lindsay, K., Razzaq, S., Chetrit, M., & Friedrich, M. (2022). Simultaneous multi-parametric acquisition and reconstruction techniques in cardiac magnetic resonance imaging: basic concepts and status of clinical development. Frontiers in Cardiovascular Medicine, 9.
2. Bustin, A., Lima da Cruz, G., Jaubert, O., Lopez, K., Botnar, R. M., & Prieto, C. (2019). High‐dimensionality undersampled patch‐based reconstruction (HD‐PROST) for accelerated multi‐contrast MRI. Magnetic resonance in medicine, 81(6), 3705-3719.
3. Qi, H., Bustin, A., Cruz, G., Jaubert, O., Chen, H., Botnar, R. M., & Prieto, C. (2019). Free-running simultaneous myocardial T1/T2 mapping and cine imaging with 3D whole-heart coverage and isotropic spatial resolution. Magnetic resonance imaging, 63, 159-169.
4. Christodoulou, A. G., Shaw, J. L., Nguyen, C., Yang, Q., Xie, Y., Wang, N., & Li, D. (2018). Magnetic resonance multitasking for motion-resolved quantitative cardiovascular imaging. Nature biomedical engineering, 2(4), 215-226.
5. Mao, X., Lee, H. L., Hu, Z., Cao, T., Han, F., Ma, S., ... & Christodoulou, A. G. (2022). Simultaneous multi-slice cardiac mr multitasking for motion-resolved, non-ECG, free-breathing T1-T2 mapping. Frontiers in Cardiovascular Medicine, 267.
6. Cao, T., Wang, N., Kwan, A. C., Lee, H. L., Mao, X., Xie, Y., ... & Li, D. (2022). Free‐breathing, non‐ECG, simultaneous myocardial T1, T2, T2*, and fat‐fraction mapping with motion‐resolved cardiovascular MR multitasking. Magnetic Resonance in Medicine, 88(4), 1748-1763.
7. Hu, Z., Xiao, J., Mao, X., Xie, Y., Kwan, A. C., Song, S. S., ... & Fan, Z. (2023). MR Multitasking‐based multi‐dimensional assessment of cardiovascular system (MT‐MACS) with extended spatial coverage and water‐fat separation. Magnetic Resonance in Medicine, 89(4), 1496-1505.
8. Hamilton, J. I. (2022). A self-supervised deep learning reconstruction for shortening the breathhold and acquisition window in cardiac magnetic resonance fingerprinting. Frontiers in Cardiovascular Medicine, 1665.