Compressive manifold learning: Estimating one-dimensional respiratory motion directly from undersampled k-space data

Validation and quantification of left ventricular function during exercise and free breathing from real-time cardiac magnetic resonance images

Exercise cardiovascular magnetic resonance (CMR) can unmask cardiac pathology not evident at rest. Real-time CMR in free breathing can be used, but respiratory motion may compromise quantification of left ventricular (LV) function. We aimed to develop and validate a post-processing algorithm that semi-automatically sorts real-time CMR images according to breathing to facilitate quantification of LV function in free breathing exercise. A semi-automatic algorithm utilizing manifold learning (Laplacian Eigenmaps) was developed for respiratory sorting. Feasibility was tested in eight healthy volunteers and eight patients who underwent ECG-gated and real-time CMR at rest. Additionally, volunteers performed exercise CMR at 60% of maximum heart rate. The algorithm was validated for exercise by comparing LV mass during exercise to rest. Respiratory sorting to end expiration and end inspiration (processing time 20 to 40 min) succeeded in all research participants. Bias ± SD for LV mass was 0 ± 5 g when comparing real-time CMR at rest, and 0 ± 7 g when comparing real-time CMR during exercise to ECG-gated at rest. This study presents a semi-automatic algorithm to retrospectively perform respiratory sorting in free breathing real-time CMR. This can facilitate implementation of exercise CMR with non-ECG-gated free breathing real-time imaging, without any additional physiological input.

Subsampled brain MRI reconstruction by generative adversarial neural networks

A main challenge in magnetic resonance imaging (MRI) is speeding up scan time. Beyond improving patient experience and reducing operational costs, faster scans are essential for time-sensitive imaging, such as fetal, cardiac, or functional MRI, where temporal resolution is important and target movement is unavoidable, yet must be reduced. Current MRI acquisition methods speed up scan time at the expense of lower spatial resolution and costlier hardware. We introduce a practical, software-only framework, based on deep learning, for accelerating MRI acquisition, while maintaining anatomically meaningful imaging. This is accomplished by MRI subsampling followed by estimating the missing k-space samples via generative adversarial neural networks. A generator-discriminator interplay enables the introduction of an adversarial cost in addition to fidelity and image-quality losses used for optimizing the reconstruction. Promising reconstruction results are obtained from feasible sampling patterns of up to a fivefold acceleration of diverse brain MRIs, from a large publicly available dataset of healthy adult scans as well as multimodal acquisitions of multiple sclerosis patients and dynamic contrast-enhanced MRI (DCE-MRI) sequences of stroke and tumor patients. Clinical usability of the reconstructed MRI scans is assessed by performing either lesion or healthy tissue segmentation and comparing the results to those obtained by using the original, fully sampled images. Reconstruction quality and usability of the DCE-MRI sequences is demonstrated by calculating the pharmacokinetic (PK) parameters. The proposed MRI reconstruction approach is shown to outperform state-of-the-art methods for all datasets tested in terms of the peak signal-to-noise ratio (PSNR), the structural similarity index (SSIM), as well as either the mean squared error (MSE) with respect to the PK parameters, calculated for the fully sampled DCE-MRI sequences, or the segmentation compatibility, measured in terms of Dice scores and Hausdorff distance. The code is available on GitHub.

Respiratory motion correction for liver contrast-enhanced ultrasound by automatic selection of a reference image

PURPOSE

Respiratory motion correction is necessary for the quantitative analysis of liver contrast-enhanced ultrasound (CEUS) image sequences. Most respiratory motion correction methods are based on the dual mode of CEUS image sequences, including contrast and grayscale image sequences. Due to free-breathing motion, the acquired two-dimensional (2D) ultrasound cine might show the in-plane and out-of-plane motion of tumors. The registration of an entire 2D ultrasound contrast image sequence based on out-of-plane images is ineffective. For the respiratory motion correction of CEUS sequences, the reference image is usually considered the standard for the deletion of any out-of-plane images. Most methods used for the selection of the reference image are subjective in nature. Here, a quantitative selection method for an optimal reference image from CEUS image sequences in the B mode and contrast mode was explored.

METHODS

The original high-dimensional ultrasound grayscale image data were mapped into a two-dimensional space using Laplacian Eigenmaps (LE), and K-means clustering was adopted. The center image of the larger cluster with a near-peak contrast intensity was considered the optimal ultrasound reference image. In the ultrasound grayscale image sequence, the images with the maximum correlations to the reference image in the same time interval were selected as the corrected image sequence. The effectiveness of this proposed method was then validated on 18 CEUS cases of VX2 tumors in rabbit livers.

RESULTS

Correction smoothed the time-intensity curves (TICs) extracted from the region of interest of the CEUS image sequences. Before correction, the average of the total mean structural similarity (TMSSIM) and the average of the mean correlation coefficient (MCC) from the image sequences were 0.45 ± 0.11 and 0.67 ± 0.16, respectively, and after correction, the average TMSSIM and MCC increased (P < 0.001) by 31% to 0.59 ± 0.11 and by 21% to 0.81 ± 0.11, respectively. The average deviation value (DV) index of the TICs from the image sequences prior to correction was 92.16 ± 18.12, and correction reduced the average to 31.71 ± 7.31. The average TMSSIM and MCC values after correction using the mean frame of the reference image (MBMFRI) were clearly lower than those after correction using the proposed method (P < 0.001). Moreover, the average DV after correction using the MBMFRI was obviously higher than that after correction using the proposed method (P < 0.001).

CONCLUSIONS

The breathing frequency of rabbits is notably faster than that of human beings, but the proposed correction method could reduce the effect of the respiratory motion in the CEUS image sequences. The reference image was selected quantitatively, which could improve the accuracy of the quantitative analysis of rabbit liver CEUS sequences using the reference image method based on the current standard of manual selection and the MBMFRI. This easy-to-operate method can potentially be used in both animal studies and clinical applications.

Reconstruction techniques for cardiac cine MRI

The present survey describes the state-of-the-art techniques for dynamic cardiac magnetic resonance image reconstruction. Additionally, clinical relevance, main challenges, and future trends of this image modality are outlined. Thus, this paper aims to provide a general vision about cine MRI as the standard procedure in functional evaluation of the heart, focusing on technical methodologies.

Motion robust high resolution 3D free-breathing pulmonary MRI using dynamic 3D image self-navigator

PURPOSE

To achieve motion robust high resolution 3D free-breathing pulmonary MRI utilizing a novel dynamic 3D image navigator derived directly from imaging data.

METHODS

Five-minute free-breathing scans were acquired with a 3D ultrashort echo time (UTE) sequence with 1.25 mm isotropic resolution. From this data, dynamic 3D self-navigating images were reconstructed under locally low rank (LLR) constraints and used for motion compensation with one of two methods: a soft-gating technique to penalize the respiratory motion induced data inconsistency, and a respiratory motion-resolved technique to provide images of all respiratory motion states.

RESULTS

Respiratory motion estimation derived from the proposed dynamic 3D self-navigator of 7.5 mm isotropic reconstruction resolution and a temporal resolution of 300 ms was successful for estimating complex respiratory motion patterns. This estimation improved image quality compared to respiratory belt and DC-based navigators. Respiratory motion compensation with soft-gating and respiratory motion-resolved techniques provided good image quality from highly undersampled data in volunteers and clinical patients.

CONCLUSION

An optimized 3D UTE sequence combined with the proposed reconstruction methods can provide high-resolution motion robust pulmonary MRI. Feasibility was shown in patients who had irregular breathing patterns in which our approach could depict clinically relevant pulmonary pathologies. Magn Reson Med 79:2954-2967, 2018. © 2017 International Society for Magnetic Resonance in Medicine.

Respiratory motion model based on the noise covariance matrix of a receive array

PURPOSE

Tracking of the internal anatomy by means of a motion model that uses the MR-derived motion fields and noise covariance matrix (NCM) dynamic as a surrogate signal.

METHODS

A 2D respiratory motion model was developed based on the MR-derived motion fields and the NCM of a receive array used in MRI. Temporal dynamics of the NCM were used as a motion surrogate for a linear correspondence motion model. The model performance was tested on five healthy volunteers with a liver as the target. The motion fields were calculated from the cineMR frames with an optical flow registration tool.

RESULTS

The model estimated the liver motion with an average residual error of 2.3 mm (13% of the motion amplitude). The model formation takes 3 min and the model latency was 0.5 s in the current implementation. The limiting factor for the latency is the current update time of the NCM (0.48 s), which in principle can be reduced to 0.004 s with an alternative way to determine the NCM.

CONCLUSIONS

The 2D respiratory motion of the liver can be effectively estimated with the linear motion model that uses the temporal behavior of the NCM as motion surrogate. Magn Reson Med 79:1730-1735, 2018. © 2017 International Society for Magnetic Resonance in Medicine.

High-Resolution Self-Gated Dynamic Abdominal MRI Using Manifold Alignment

We present a novel retrospective self-gating method based on manifold alignment (MA), which enables reconstruction of free breathing, high spatial, and temporal resolution abdominal magnetic resonance imaging sequences. Based on a radial golden-angle acquisition trajectory, our method enables a multidimensional self-gating signal to be extracted from the k -space data for more accurate motion representation. The k -space radial profiles are evenly divided into a number of overlapping groups based on their radial angles. MA is then used to simultaneously learn and align the low dimensional manifolds of all groups, and embed them into a common manifold. In the manifold, k -space profiles that represent similar respiratory positions are close to each other. Image reconstruction is performed by combining radial profiles with evenly distributed angles that are close in the manifold. Our method was evaluated on both 2-D and 3-D synthetic and in vivo data sets. On the synthetic data sets, our method achieved high correlation with the ground truth in terms of image intensity and virtual navigator values. Using the in vivo data, compared with a state-of-the-art approach based on the center of k -space gating, our method was able to make use of much richer profile data for self-gating, resulting in statistically significantly better quantitative measurements in terms of organ sharpness and image gradient entropy.

Automatic assessment of average diaphragm motion trajectory from 4DCT images through machine learning

To automatically estimate average diaphragm motion trajectory (ADMT) based on four-dimensional computed tomography (4DCT), facilitating clinical assessment of respiratory motion and motion variation and retrospective motion study. We have developed an effective motion extraction approach and a machine-learning-based algorithm to estimate the ADMT. Eleven patients with 22 sets of 4DCT images (4DCT1 at simulation and 4DCT2 at treatment) were studied. After automatically segmenting the lungs, the differential volume-per-slice (dVPS) curves of the left and right lungs were calculated as a function of slice number for each phase with respective to the full-exhalation. After 5-slice moving average was performed, the discrete cosine transform (DCT) was applied to analyze the dVPS curves in frequency domain. The dimensionality of the spectrum data was reduced by using several lowest frequency coefficients (f_v) to account for most of the spectrum energy (Σf_v²). Multiple linear regression (MLR) method was then applied to determine the weights of these frequencies by fitting the ground truth-the measured ADMT, which are represented by three pivot points of the diaphragm on each side. The 'leave-one-out' cross validation method was employed to analyze the statistical performance of the prediction results in three image sets: 4DCT1, 4DCT2, and 4DCT1 + 4DCT2. Seven lowest frequencies in DCT domain were found to be sufficient to approximate the patient dVPS curves (R = 91%-96% in MLR fitting). The mean error in the predicted ADMT using leave-one-out method was 0.3 ± 1.9 mm for the left-side diaphragm and 0.0 ± 1.4 mm for the right-side diaphragm. The prediction error is lower in 4DCT2 than 4DCT1, and is the lowest in 4DCT1 and 4DCT2 combined. This frequency-analysis-based machine learning technique was employed to predict the ADMT automatically with an acceptable error (0.2 ± 1.6 mm). This volumetric approach is not affected by the presence of the lung tumors, providing an automatic robust tool to evaluate diaphragm motion.