Single Breath-Hold 3-Dimensional Magnetic Resonance Elastography Depicts Liver Fibrosis and Inflammation in Obese Patients.

OBJECTIVES: Three-dimensional (3D) magnetic resonance elastography (MRE) measures liver fibrosis and inflammation but requires several breath-holds that hamper clinical acceptance. The aim of this study was to evaluate the technical and clinical feasibility of a single breath-hold 3D MRE sequence as a means of measuring liver fibrosis and inflammation in obese patients. METHODS: From November 2020 to December 2021, subjects were prospectively enrolled and divided into 2 groups. Group 1 included healthy volunteers (n = 10) who served as controls to compare the single breath-hold 3D MRE sequence with a multiple-breath-hold 3D MRE sequence. Group 2 included liver patients (n = 10) who served as participants to evaluate the clinical feasibility of the single breath-hold 3D MRE sequence in measuring liver fibrosis and inflammation. Controls and participants were scanned at 60 Hz mechanical excitation with the single breath-hold 3D MRE sequence to retrieve the magnitude of the complex-valued shear modulus (|G*| [kPa]), the shear wave speed (Cs [m/s]), and the loss modulus (G" [kPa]). The controls were also scanned with a multiple-breath-hold 3D MRE sequence for comparison, and the participants had histopathology (Ishak scores) for correlation with Cs and G". RESULTS: For the 10 controls, 5 were female, and the mean age and body mass index were 33.1 ± 9.5 years and 23.0 ± 2.1 kg/m 2 , respectively. For the 10 participants, 8 were female, and the mean age and body mass index were 45.1 ± 16.5 years and 33.1 ± 4.0 kg/m 2 (obese range), respectively. All participants were suspected of having nonalcoholic fatty liver disease. Bland-Altman analysis of the comparison in controls shows there are nonsignificant differences in |G*|, Cs, and G" below 6.5%, suggesting good consensus between the 2 sequences. For the participants, Cs and G" correlated significantly with Ishak fibrosis and inflammation grades, respectively ( ρ = 0.95, P < 0.001, and ρ = 0.84, P = 0.002). CONCLUSION: The single breath-hold 3D MRE sequence may be effective in measuring liver fibrosis and inflammation in obese patients.

Native-resolution myocardial principal Eulerian strain mapping using convolutional neural networks and Tagged Magnetic Resonance Imaging.

BACKGROUND: Assessment of regional myocardial function at native pixel-level resolution can play a crucial role in recognizing the early signs of the decline in regional myocardial function. Extensive data processing in existing techniques limits the effective resolution and accuracy of the generated strain maps. The purpose of this study is to compute myocardial principal strain maps εp1 and εp2 from tagged MRI (tMRI) at the native image resolution using deep-learning local patch convolutional neural network (CNN) models (DeepStrain). METHODS: For network training, validation, and testing, realistic tMRI datasets were generated and consisted of 53,606 cine images simulating the heart, the liver, blood pool, and backgrounds, including ranges of shapes, positions, motion patterns, noise, and strain. In addition, 102 in-vivo image datasets from three healthy subjects, and three Pulmonary Arterial Hypertension patients, were acquired and used to assess the network's in-vivo performance. Four convolutional neural networks were trained for mapping input tagging patterns to corresponding ground-truth principal strains using different cost functions. Strain maps using harmonic phase analysis (HARP) were obtained with various spectral filtering settings for comparison. CNN and HARP strain maps were compared at the pixel level versus the ground-truth and versus the least-loss in-vivo maps using Pearson correlation coefficients (R) and the median error and Inter-Quartile Range (IQR) histograms. RESULTS: CNN-based local patch DeepStrain maps at a phantom resolution of 1.1mm × 1.1 mm and in-vivo resolution of 2.1mm × 1.6 mm were artifact-free with multiple fold improvement with εp1 ground-truth median error of 0.009(0.007) vs. 0.32(0.385) using HARP and εp2 ground-truth error of 0.016(0.021) vs. 0.181(0.08) using HARP. CNN-based strain maps showed substantially higher agreement with the ground-truth maps with correlation coefficients R > 0.91 for εp1 and εp2 compared to R < 0.21 and R < 0.82 for HARP-generated maps, respectively. CONCLUSION: CNN-generated Eulerian strain mapping permits artifact-free visualization of myocardial function at the native image resolution.

Direct pixel to pixel principal strain mapping from tagging MRI using end to end deep convolutional neural network (DeepStrain).

Regional soft tissue mechanical strain offers crucial insights into tissue's mechanical function and vital indicators for different related disorders. Tagging magnetic resonance imaging (tMRI) has been the standard method for assessing the mechanical characteristics of organs such as the heart, the liver, and the brain. However, constructing accurate artifact-free pixelwise strain maps at the native resolution of the tagged images has for decades been a challenging unsolved task. In this work, we developed an end-to-end deep-learning framework for pixel-to-pixel mapping of the two-dimensional Eulerian principal strains [Formula: see text] and [Formula: see text] directly from 1-1 spatial modulation of magnetization (SPAMM) tMRI at native image resolution using convolutional neural network (CNN). Four different deep learning conditional generative adversarial network (cGAN) approaches were examined. Validations were performed using Monte Carlo computational model simulations, and in-vivo datasets, and compared to the harmonic phase (HARP) method, a conventional and validated method for tMRI analysis, with six different filter settings. Principal strain maps of Monte Carlo tMRI simulations with various anatomical, functional, and imaging parameters demonstrate artifact-free solid agreements with the corresponding ground-truth maps. Correlations with the ground-truth strain maps were R = 0.90 and 0.92 for the best-proposed cGAN approach compared to R = 0.12 and 0.73 for the best HARP method for [Formula: see text] and [Formula: see text], respectively. The proposed cGAN approach's error was substantially lower than the error in the best HARP method at all strain ranges. In-vivo results are presented for both healthy subjects and patients with cardiac conditions (Pulmonary Hypertension). Strain maps, obtained directly from their corresponding tagged MR images, depict for the first time anatomical, functional, and temporal details at pixelwise native high resolution with unprecedented clarity. This work demonstrates the feasibility of using the deep learning cGAN for direct myocardial and liver Eulerian strain mapping from tMRI at native image resolution with minimal artifacts.

Adaptive smoothing of high angular resolution diffusion-weighted imaging data by generalized cross-validation improves Q-ball orientation distribution function reconstruction.

Q-ball imaging (QBI) is a high angular resolution diffusion-weighted imaging (HARDI) technique for reconstructing the orientation distribution function (ODF). Some form of smoothing or regularization is typically required in the ODF reconstruction from low signal-to-noise ratio HARDI data. The amount of smoothing or regularization is usually set a priori at the discretion of the investigator. In this article, we apply an adaptive and objective means of smoothing the raw HARDI data using the smoothing splines on the sphere method with generalized cross-validation (GCV) to estimate the diffusivity profile in each voxel. Subsequently, we reconstruct the ODF, from the smoothed data, based on the Funk-Radon transform (FRT) used in QBI. The spline method was applied to both simulated data and in vivo human brain data. Simulated data show that the smoothing splines on the sphere method with GCV smoothing reduces the mean squared error in estimates of the ODF as compared with the standard analytical QBI approach. The human data demonstrate the utility of the method for estimating smooth ODFs.

Utility of axial and radial diffusivity from diffusion tensor MRI as markers of neurodegeneration in amyotrophic lateral sclerosis.

OBJECTIVE: To investigate changes in the diffusion tensor imaging measures, axial diffusivity and radial diffusivity, in addition to the more commonly used fractional anisotropy and mean diffusivity, in patients with amyotrophic lateral sclerosis (ALS) using the voxel-based statistical analysis tool, tract based spatial statistics. METHODS: We studied 12 patients with ALS and 19 normal controls using diffusion tensor imaging; tract based spatial statistics was applied to study changes in fractional anisotropy, mean diffusivity, axial diffusivity and radial diffusivity values in brain white matter tracts. ALS patients were evaluated using clinical examination, administration of the revised ALS functional rating scale and measurement of the forced vital capacity. RESULTS: In ALS patients, we found significant increases in axial diffusivity, radial diffusivity, and mean diffusivity and significant decreases in fractional anisotropy. Increases in axial diffusivity and radial diffusivity were more widespread and more prominent in the corticospinal tract than the decreases in fractional anisotropy. The decreases in fractional anisotropy were evident only in the corona radiata and genu of the corpus callosum. CONCLUSION: In ALS, axial diffusivity and radial diffusivity may be useful diffusion tensor imaging-derived indices to consider in addition to fractional anisotropy and mean diffusivity to aid in demonstrating neurodegenerative changes.

An information theoretic approach characterizing diffusion anisotropy in diffusion-weighted magnetic resonance images.

We propose an alternative approach that does not rely on tensor models for characterizing diffusion anisotropy from diffusion-weighted magnetic resonance images. Information content inherent in the diffusion attenuation values are the only measures needed for our characterization. We explore the information content inherent in these values. We calculate Shannon's entropy on the diffusion attenuation values measured across the applied diffusion-sensitizing gradient directions. This method is evaluated with data generated with different diffusion gradient encoding schemes demonstrating the validity of our approach and its potential use to better differentiate between brain tissue types over tensor-based measures.