Motion compensated self supervised deep learning for highly accelerated 3D ultrashort Echo time pulmonary MRI

Spectrally selective and interleaved water imaging and fat imaging (siWIFI)

PURPOSE

To develop a novel imaging sequence that independently acquires water and fat images while being inherently insensitive to motion.

METHODS

The new sequence, termed spectrally selective and interleaved water imaging and fat imaging (siWIFI), uses a narrow bandwidth RF pulse for selective excitation of water and fat separately. The interleaved acquisition method ensures that the obtained water and fat images are inherently coregistered. A radial sampling strategy further reduces motion-induced artifacts. Phantoms with lipid concentrations ranging from 0% to 50% were scanned to measure fat fraction. Moreover, healthy volunteers were scanned to assess the in vivo feasibility of fat fraction measurement at the hip, knee, and liver. In vivo fat fraction measurements were compared with those from vendor-provided iterative decomposition of water and fat with echo asymmetry and least-squares estimation (IDEAL) scans. Furthermore, a magnetization transfer (MT) preparation module was incorporated to demonstrate the feasibility of simultaneous measurement of fat fraction and MT ratio utilizing the siWIFI framework.

RESULTS

The phantom fat fractions measured by siWIFI showed excellent correlation with lipid concentrations (R2 = 0.9995, p < 0.0001). In vivo studies demonstrated that the fat fractions obtained from siWIFI were comparable to those from IDEAL. Additionally, siWIFI demonstrates reduced motion artifacts from pulsatile flow in knee imaging compared to IDEAL scans and exhibits less sensitivity to respiratory motion in liver imaging compared to IDEAL scans without breath-hold. The knee imaging study demonstrated that MT-prepared siWIFI is capable of generating fat fraction and MT ratio maps simultaneously.

CONCLUSION

The proposed siWIFI sequence allows selective water-fat imaging and quantification with reduced motion artifacts.

A review of deep learning-based reconstruction methods for accelerated MRI using spatiotemporal and multi-contrast redundancies

Accelerated magnetic resonance imaging (MRI) has played an essential role in reducing data acquisition time for MRI. Acceleration can be achieved by acquiring fewer data points in k-space, which results in various artifacts in the image domain. Conventional reconstruction methods have resolved the artifacts by utilizing multi-coil information, but with limited robustness. Recently, numerous deep learning-based reconstruction methods have been developed, enabling outstanding reconstruction performances with higher acceleration. Advances in hardware and developments of specialized network architectures have produced such achievements. Besides, MRI signals contain various redundant information including multi-coil redundancy, multi-contrast redundancy, and spatiotemporal redundancy. Utilization of the redundant information combined with deep learning approaches allow not only higher acceleration, but also well-preserved details in the reconstructed images. Consequently, this review paper introduces the basic concepts of deep learning and conventional accelerated MRI reconstruction methods, followed by review of recent deep learning-based reconstruction methods that exploit various redundancies. Lastly, the paper concludes by discussing the challenges, limitations, and potential directions of future developments.

Knowledge-driven deep learning for fast MR imaging: Undersampled MR image reconstruction from supervised to un-supervised learning

Deep learning (DL) has emerged as a leading approach in accelerating MRI. It employs deep neural networks to extract knowledge from available datasets and then applies the trained networks to reconstruct accurate images from limited measurements. Unlike natural image restoration problems, MRI involves physics-based imaging processes, unique data properties, and diverse imaging tasks. This domain knowledge needs to be integrated with data-driven approaches. Our review will introduce the significant challenges faced by such knowledge-driven DL approaches in the context of fast MRI along with several notable solutions, which include learning neural networks and addressing different imaging application scenarios. The traits and trends of these techniques have also been given which have shifted from supervised learning to semi-supervised learning, and finally, to unsupervised learning methods. In addition, MR vendors' choices of DL reconstruction have been provided along with some discussions on open questions and future directions, which are critical for the reliable imaging systems.

Stop moving: MR motion correction as an opportunity for artificial intelligence

Subject motion is a long-standing problem of magnetic resonance imaging (MRI), which can seriously deteriorate the image quality. Various prospective and retrospective methods have been proposed for MRI motion correction, among which deep learning approaches have achieved state-of-the-art motion correction performance. This survey paper aims to provide a comprehensive review of deep learning-based MRI motion correction methods. Neural networks used for motion artifacts reduction and motion estimation in the image domain or frequency domain are detailed. Furthermore, besides motion-corrected MRI reconstruction, how estimated motion is applied in other downstream tasks is briefly introduced, aiming to strengthen the interaction between different research areas. Finally, we identify current limitations and point out future directions of deep learning-based MRI motion correction.

Comparison of weighting algorithms to mitigate respiratory motion in free-breathing neonatal pulmonary radial UTE-MRI

Background. Thoracoabdominal MRI is limited by respiratory motion, especially in populations who cannot perform breath-holds. One approach for reducing motion blurring in radially-acquired MRI is respiratory gating. Straightforward 'hard-gating' uses only data from a specified respiratory window and suffers from reduced SNR. Proposed 'soft-gating' reconstructions may improve scan efficiency but reduce motion correction by incorporating data with nonzero weight acquired outside the specified window. However, previous studies report conflicting benefits, and importantly the choice of soft-gated weighting algorithm and effect on image quality has not previously been explored. The purpose of this study is to map how variable soft-gated weighting functions and parameters affect signal and motion blurring in respiratory-gated reconstructions of radial lung MRI, using neonates as a model population.Methods. Ten neonatal inpatients with respiratory abnormalities were imaged using a 1.5 T neonatal-sized scanner and 3D radial ultrashort echo-time (UTE) sequence. Images were reconstructed using ungated, hard-gated, and several soft-gating weighting algorithms (exponential, sigmoid, inverse, and linear weighting decay outside the period of interest), with %Nprojrepresenting the relative amount of data included. The apparent SNR (aSNR) and motion blurring (measured by the maximum derivative of image intensity at the diaphragm, MDD) were compared between reconstructions.Results. Soft-gating functions produced higher aSNR and lower MDD than hard-gated images using equivalent %Nproj, as expected. aSNR was not identical between different gating schemes for given %Nproj. While aSNR was approximately linear with %Nprojfor each algorithm, MDD performance diverged between functions as %Nprojdecreased. Algorithm performance was relatively consistent between subjects, except in images with high noise.Conclusion. The algorithm selection for soft-gating has a notable effect on image quality of respiratory-gated MRI; the timing of included data across the respiratory phase, and not simply the amount of data, plays an important role in aSNR. The specific soft-gating function and parameters should be considered for a given imaging application's requirements of signal and sharpness.

A comprehensive set of ultrashort echo time magnetic resonance imaging biomarkers to assess cortical bone health: A feasibility study at clinical field strength

INTRODUCTION

Conventional bone imaging methods primarily use X-ray techniques to assess bone mineral density (BMD), focusing exclusively on the mineral phase. This approach lacks information about the organic phase and bone water content, resulting in an incomplete evaluation of bone health. Recent research highlights the potential of ultrashort echo time magnetic resonance imaging (UTE MRI) to measure cortical porosity and estimate BMD based on signal intensity. UTE MRI also provides insights into bone water distribution and matrix organization, enabling a comprehensive bone assessment with a single imaging technique. Our study aimed to establish quantifiable UTE MRI-based biomarkers at clinical field strength to estimate BMD and microarchitecture while quantifying bound water content and matrix organization.

METHODS

Femoral bones from 11 cadaveric specimens (n = 4 males 67-92 yrs of age, n = 7 females 70-95 yrs of age) underwent dual-echo UTE MRI (3.0 T, 0.45 mm resolution) with different echo times and high resolution peripheral quantitative computed tomography (HR-pQCT) imaging (60.7 μm voxel size). Following registration, a 4.5 mm HR-pQCT region of interest was divided into four quadrants and used across the multi-modal images. Statistical analysis involved Pearson correlation between UTE MRI porosity index and a signal-intensity technique used to estimate BMD with corresponding HR-pQCT measures. UTE MRI was used to calculate T1 relaxation time and a novel bound water index (BWI), compared across subregions using repeated measures ANOVA.

RESULTS

The UTE MRI-derived porosity index and signal-intensity-based estimated BMD correlated with the HR-pQCT variables (porosity: r = 0.73, p = 0.006; BMD: r = 0.79, p = 0.002). However, these correlations varied in strength when we examined each of the four quadrants (subregions, r = 0.11-0.71). T1 relaxometry and the BWI exhibited variations across the four subregions, though these differences were not statistically significant. Notably, we observed a strong negative correlation between T1 relaxation time and the BWI (r = -0.87, p = 0.0006).

CONCLUSION

UTE MRI shows promise for being an innocuous method for estimating cortical porosity and BMD parameters while also giving insight into bone hydration and matrix organization. This method offers the potential to equip clinicians with a more comprehensive array of imaging biomarkers to assess bone health without the need for invasive or ionizing procedures.

Deep Learning for Retrospective Motion Correction in MRI: A Comprehensive Review

Motion represents one of the major challenges in magnetic resonance imaging (MRI). Since the MR signal is acquired in frequency space, any motion of the imaged object leads to complex artefacts in the reconstructed image in addition to other MR imaging artefacts. Deep learning has been frequently proposed for motion correction at several stages of the reconstruction process. The wide range of MR acquisition sequences, anatomies and pathologies of interest, and motion patterns (rigid vs. deformable and random vs. regular) makes a comprehensive solution unlikely. To facilitate the transfer of ideas between different applications, this review provides a detailed overview of proposed methods for learning-based motion correction in MRI together with their common challenges and potentials. This review identifies differences and synergies in underlying data usage, architectures, training and evaluation strategies. We critically discuss general trends and outline future directions, with the aim to enhance interaction between different application areas and research fields.

Submillimeter lung MRI at 0.55 T using balanced steady-state free precession with half-radial dual-echo readout (bSTAR)

PURPOSE

To demonstrate the feasibility of high-resolution morphologic lung MRI at 0.55 T using a free-breathing balanced steady-state free precession half-radial dual-echo imaging technique (bSTAR).

METHODS

Self-gated free-breathing bSTAR (TE1 /TE2 /TR of 0.13/1.93/2.14 ms) lung imaging in five healthy volunteers and a patient with granulomatous lung disease was performed using a 0.55 T MR-scanner. A wobbling Archimedean spiral pole (WASP) trajectory was used to ensure a homogenous coverage of k-space over multiple breathing cycles. WASP uses short-duration interleaves randomly tilted by a small polar angle and rotated by a golden angle about the polar axis. Data were acquired continuously over 12:50 min. Respiratory-resolved images were reconstructed off-line using compressed sensing and retrospective self-gating. Reconstructions were performed with a nominal resolution of 0.9 mm and a reduced isotropic resolution of 1.75 mm corresponding to shorter simulated scan times of 8:34 and 4:17 min, respectively. Analysis of apparent SNR was performed in all volunteers and reconstruction settings.

RESULTS

The technique provided artifact-free morphologic lung images in all subjects. The short TR of bSTAR in conjunction with a field strength of 0.55 T resulted in a complete mitigation of off-resonance artifacts in the chest. Mean SNR values in healthy lung parenchyma for the 12:50 min scan were 3.6 ± 0.8 and 24.9 ± 6.2 for 0.9 mm and 1.75 mm reconstructions, respectively.

CONCLUSION

This study demonstrates the feasibility of morphologic lung MRI with a submillimeter isotropic spatial resolution in human subjects with bSTAR at 0.55 T.