Neural network enhanced 3D turbo spin echo for MR intracranial vessel wall imaging

PurposeTo improve the signal-to-noise ratio (SNR) and image sharpness for whole brain isotropic 0.5 mm three-dimensional (3D) T₁ weighted (T₁w) turbo spin echo (TSE) intracranial vessel wall imaging (IVWI) at 3 T.MethodsThe variable flip angle (VFA) method enables useful optimization across scan efficiency, SNR and relaxation induced point spread function (PSF) for TSE imaging. A convolutional neural network (CNN) was developed to retrospectively enhance the acquired TSE image with PSF blurring. The previously developed VFA method to increase SNR at the expense of blur can be combined with the presented PSF correction to yield long echo train length (ETL) scan while the acquired image remains high SNR and sharp. The overall approach can enable an optimized solution for accelerated whole brain high-resolution 3D T₁w TSE IVWI. Its performance was evaluated on healthy volunteers and patients.ResultsThe PSF blurred image acquired by a long ETL scan can be enhanced by CNN to restore similar sharpness as a short ETL scan, which outperforms the traditional linear PSF enhancement approach. For accelerated whole brain IVWI on volunteers, the optimized isotropic 0.5 mm 3D T₁w TSE sequence with CNN based PSF enhancement provides sufficient flow suppression and improved image quality. Preliminary results on patients further demonstrated its improved delineation for intracranial vessel wall and plaque morphology.ConclusionThe CNN enhanced VFA TSE imaging enables an overall image quality improvement for high-resolution 3D T₁w IVWI, and may provide a better tradeoff across scan efficiency, SNR and PSF for 3D TSE acquisitions.     

MRI of liver: a comparison of CNR enhancement using high dose and low dose ferumoxide infusion in patients with colorectal liver metastases

The purpose of this study was to compare the effects of high dose (HD) and low dose (LD) ferumoxides infusions on lesion-to-liver contrast-to-noise ratio (CNR) using four different T(2)-weighted MR sequences. Seventy-three patients with known colorectal liver metastases underwent T(2)-weighted fast spin echo (FSE) imaging before and after ferumoxides. After ferumoxides, T(2)-weighted dual echo (DE) and T(2)-weighted GRE FLASH images were also obtained. To evaluate the relationship between TE length and lesion-to-liver CNR, the same FLASH sequence was repeated in 18 LD patients after lengthening the TE. Ferumoxides was administered at a dose of 15 micromol/kg (HD) and 7.5 micromol/kg (LD) in 45 and 28 patients, respectively. The effects of HD and LD ferumoxides infusions were measured as the percentage signal intensity change (PSIC) in the liver and lesions, lesion-to-liver CNR and the change in lesion-to-liver CNR (DeltaCNR). In both LD and HD groups, all CNR values obtained after SPIO were significantly greater than those observed with unenhanced FSE (p < 0.01). There was no significant difference between the mean CNR values obtained with either dose for any sequence. With the FLASH sequence, CNR increased progressively with longer TE. At the longest TE of 26 ms, mean CNR was higher than that recorded with any of the other sequences. Although mean liver PSIC was significantly greater in the HD group than in the LD group (p < 0.01) because the mean lesion PSIC was also greater in the HD group, the mean DeltaCNR after ferumoxides was not significantly different in the two groups. LD SPIO enhanced MR significantly increases lesion-to-liver CNR compared with unenhanced images. At 1. 0 T, HD and LD ferumoxides infusions produce comparable lesion-to-liver CNR. Our results suggest that at 1.0 T ferumoxides may be administered at a dose of 7.5 micromol/kg without loss of image quality.     

A novel technique for imaging with inhomogeneous fields

We introduce a simple, efficient, low-SAR method for magnetic resonance imaging in the presence of a static field with a permanent, and possibly large gradient. The technique, which is called slant-slice imaging is essentially a spin-echo imaging sequence except that the imaging slice is oriented such that the static field gradient can be used in conjunction with applied gradients during readout. Data are collected for 2D slices. Unlike single point imaging techniques, entire lines of k-space are acquired with each readout. The slant-slice pulse sequence is used to obtain high quality images, using a clinical scanner to simulate a static field with a large permanent gradient. The effects of the inhomogeneity are quantified by two parameters nu and q, which are useful for assessing the utility of a magnet design for 3D-MR imaging.     

Spiral MRI和图像处理

介绍了在Bruker Biospec 47/30 超导核磁共振成象仪(4.7 T)上实现Spiral快速成像及图像处理系统. 图像处理系统基于PC技术构建而成,主要功能包括：1） 将以Spiral形式采集到的时域磁共振信号转化为适用于快速傅立叶变换的笛卡尔网格(Cartesian)形式（网格化处理）;2）二维快速傅立叶变换(2D-FFT,图像重建);3）由化学位移偏置或磁场不均匀引起得偏共振效应(off-resonance effect)的校正;4）图像分析. 该软件适用于包括以多片多回波在内的各种采样方式得到的Spiral图像的重建和分析,也适用于常规成像数据的重建和分析. 所得到的图像可以以数据方式保存以供再次读入,也能够以TIF、GIF、JPG、BM等格式辅出为图像文件.     

Initial in vivo rodent sodium and proton MR imaging at 21.1 T

The first in vivo sodium and proton magnetic resonance (MR) images and localized spectra of rodents were attained using the wide bore (105 mm) high resolution 21.1-T magnet, built and operated at the National High Magnetic Field Laboratory (Tallahassee, FL, USA). Head images of normal mice (C57BL/6J) and Fisher rats (∼250 g) were acquired with custom designed radiofrequency probes at frequencies of 237/900 MHz for sodium and proton, respectively. Sodium MR imaging resolutions of ∼0.125 μl for mouse and rat heads were achieved by using a 3D back-projection pulse sequence. A gain in SNR of ∼3 for sodium and ∼2 times for proton were found relative to corresponding MR images acquired at 9.4 T. 3D Fast Low Angle Shot (FLASH) proton mouse images (50×50×50 μm³) were acquired in 90 min and corresponding rat images (100×100×100 μm³) within a total time of 120 min. Both in vivo large rodent MR imaging and localized spectroscopy at the extremely high field of 21.1 T are feasible and demonstrate improved resolution and sensitivity valuable for structural and functional brain analysis.     

Effects of static and radiofrequency magnetic field inhomogeneity in ultra-high field magnetic resonance imaging

To characterize the severe static (B(0)) and radiofrequency (B(1)) magnetic field inhomogeneity in ultra-high field (> or =7 T) magnetic resonance imaging, gradient echo (GE) and spin echo (SE) images of in vivo and postmortem human brains were acquired. The B(0) and B(1) inhomogeneity were experimentally mapped and/or numerically simulated, and correlated with the image artifacts. Whereas B(0) inhomogeneity affects predominantly GE images near air/tissue interfaces, B(1) inhomogeneity affects SE images more severely and shows non-intuitive patterns. Mapping of the B(0) and B(1) inhomogeneity is important in characterizing image artifacts. This will help develop better B(0) and B(1) inhomogeneity correction methods.     

Comparative study of fast MR imaging: quantitative analysis on image quality and efficiency among various time frames and contrast behaviors

The purpose of this study is to quantitatively compare the image quality and efficiency provided by widely available fast MR imaging pulse sequences. A composite phantom with various T1 and T2 values and subjected to periodic motion was imaged at 1.5 T. The fast MRI sequences evaluated included fast spin-echo (FSE), single shot fast spin-echo (SSFSE), echo-planar imaging (EPI), multi-slice gradient recalled (MPGR), fast MPGR (FMPGR), and fast multi-slice spoiled gradient echo (FMPSPGR). T1-weighted (T1WI), T2-weighted (T2WI), proton-density-weighted (PDWI), and T2*-weighted (T2*WI) images were evaluated in breath-hold and non-breath-hold time frames. Analysis included measurement of image signal-to-noise ratio (SNR), contrast-to-noise ratio (CNR), nonuniformity, ghosting ratio, SNR per unit time and CNR per unit time. Among fast T2WI sequences, FSE with breath-hold time frame resulted in the highest image quality and in superior SNR and CNR efficiency by a factor of 5 or 6 as compared with conventional spin echo sequence. Among fast T1WI sequences, FMPGR and FMPSPGR both with non-breath-hold time frame produced the highest image quality and SNR and CNR efficiency by a factor of greater than 5 as compared with conventional spin echo. Among fast PDWI and T2*WI sequences, FSE produced the highest SNR and CNR, and was maximally efficient with a factors of greater than 6 as compared with conventional spin echo.     

Optimization of 3D MP-RAGE for neonatal brain imaging at 3.0 T

Three-dimensional (3D) magnetic resonance imaging (MRI) has shown great potential for studying the impact of prematurity and pathology on brain development. We have investigated the potential of optimized T1-weighted 3D magnetization-prepared rapid gradient-echo imaging (MP-RAGE) for obtaining contrast between white matter (WM) and gray matter (GM) in neonates at 3 T. Using numerical simulations, we predicted that the inversion time (TI) for obtaining strongest contrast at 3 T is approximately 2 s for neonates, whereas for adults, this value is approximately 1.3 s. The optimal neonatal TI value was found to be insensitive to reasonable variations of the assumed T1 relaxation times. The maximum theoretical contrast for neonates was found to be approximately one third of that for adults. Using the optimized TI values, MP-RAGE images were obtained from seven neonates and seven adults at 3 T, and the contrast-to-noise ratio (CNR) was measured for WM versus five GM regions. Compared to adults, neonates exhibited lower CNR between cortical GM and WM and showed a different pattern of regional variation in CNR. These results emphasize the importance of sequence optimization specifically for neonates and demonstrate the challenge in obtaining strong contrast in neonatal brain with T1-weighted 3D imaging.     

Parallel imaging for first-pass myocardial perfusion

Two parallel imaging methods used for first-pass myocardial perfusion imaging were compared in terms of signal-to-noise ratio (SNR), contrast-to-noise ratio (CNR) and image artifacts. One used adaptive Time-adaptive SENSitivity Encoding (TSENSE) and the other used GeneRalized Autocalibrating Partially Parallel Acquisition (GRAPPA), which are both applied to a gradient-echo sequence. Both methods were tested on 12 patients with coronary artery disease. The order of perfusion sequences was inverted in every other patient. Image acquisition was started during the administration of a contrast bolus followed by a 20-ml saline flush (3 ml/s), and the next perfusion was started at least 15 min thereafter using an identical bolus. An acceleration rate of 2 was used in both methods, and acquisition was performed during breath-holding. Significantly higher SNR, CNR and image quality were obtained with GRAPPA images than with TSENSE images. GRAPPA, however, did not yield a higher CNR when applied after the second bolus. GRAPPA perfusion imaging produced larger differences between subjects than did TSENSE. Compared to TSENSE, GRAPPA produced significantly better CNR on the first bolus. More consistent SNR and CNR were obtained from TSENSE images than from GRAPPA images, indicating that the diagnostic value of TSENSE may be better.     

A novel, fast entropy-minimization algorithm for bias field correction in MR images

A novel, fast entropy-minimization algorithm for bias field correction in magnetic resonance (MR) images is suggested to correct the intensity inhomogeneity degradation of MR images that has become an increasing problem with the use of phased-array coils. Four important modifications were made to the conventional algorithm: (a) implementation of a modified two-step sampling strategy for stacked 2D image data sets, which included reducing the size of the measured image on each slice with a simple averaging method without changing the number of slices and then using a binary mask generated by a histogram threshold method to define the sampled voxels in the reduced image; (b) improvement of the efficiency of the correction function by using a Legendre polynomial as an orthogonal base function polynomial; (c) use of a nonparametric Parzen window estimator with a Gaussian kernel to calculate the probability density function and Shannon entropy directly from the image data; and (d) performing entropy minimization with a conjugate gradient method. Results showed that this algorithm could correct different types of MR images from different types of coils acquired at different field strengths very efficiently and with decreased computational load.     

Retrospective cardiac gating: a review of technical aspects and future directions

The advent of short TR gradient-echo imaging has made it possible to acquire cine images of the heart with conventional whole body MRI scanners. In this paper, technical details of the data collection and image reconstruction process for cine MRI using retrospective cardiac gating are presented. Specifically, the following issues are discussed: data sorting and interpolation; time resolution; motion compensation and phase information; the type of steady state sequence including optimal flip angle; respiratory motion and correction; and the potential of 3D imaging.     

An investigation of motion correction algorithms for pediatric spinal cord DTI in healthy subjects and patients with spinal cord injury

Patient and physiological motion can cause artifacts in DTI of the spinal cord which can impact image quality and diffusion indices. The purpose of this investigation was to determine a reliable motion correction method for pediatric spinal cord DTI and show effects of motion correction on DTI parameters in healthy subjects and patients with spinal cord injury. Ten healthy subjects and ten subjects with spinal cord injury were scanned using a 3 T scanner. Images were acquired with an inner field-of-view DTI sequence covering cervical spine levels C1 to C7. Images were corrected for motion using two types of transformation (rigid and affine) and three cost functions. Corrected images and transformations were examined qualitatively and quantitatively using in-house developed code. Fractional anisotropy (FA) and mean diffusivity (MD) indices were calculated and tested for statistical significance pre- and post- motion correction. Images corrected using rigid methods showed improvements in image quality, while affine methods frequently showed residual distortions in corrected images. Blinded evaluation of pre and post correction images showed significant improvement in cord homogeneity and edge conspicuity in corrected images (p < 0.0001). The average FA changes were statistically significant (p < 0.0001) in the spinal cord injury group, while healthy subjects showed less FA change and were not significant. In both healthy subjects and subjects with spinal cord injury, quantitative and qualitative analysis showed the rigid scaled-least-squares registration technique to be the most reliable and effective in improving image quality.     

Magnetic resonance imaging of coronary arteries and heart valves in a living mouse: techniques and preliminary results

New investigations in MRI of a mouse heart showed high-contrast cardiac images and thereby the possibility of doing functional cardiac studies of in vivo mice. But is MRI, in addition, capable of visualizing microstructures such as the coronary arteries and the heart valves of a living mouse? To answer this question, 2D and 3D gradient echo sequences with and without flow compensation were used to image the coronary arteries. To increase signal-to-noise ratio, a birdcage resonator was optimized for mouse heart imaging. Contrast between blood and myocardium was achieved through the inflow effect. A segmented three-dimensional FLASH sequence acquired with a multiple overlap thin slab technique showed the best results. With this technique an isotropic resolution of 100 microm was achieved. The left coronary artery could be visualized up to the apex of the heart. This is demonstrated with short axis views and 3D surface reconstructions of the mouse heart. The four cardiac valves were also visible with the 3D method.     

Artifact-free black-blood cine cardiac imaging in a single breath-hold

Fast imaging using the STimulated Echo Acquisition Mode (STEAM) sequence can produce cine images of the heart with black-blood contrast. Nevertheless, correction of deformation-related artifacts is required in order to maintain myocardial signal throughout the cardiac cycle. Recent work by our group has eliminated this artifact by combining two STEAM sequences acquired with two different demodulation gradients. Unfortunately, these two STEAM sequences were acquired on two separate breath-holds; thus, scan time doubled. In this work, we present a technique to reduce the total scan time by one half, without sacrificing image quality. The technique is based on interleaving two demodulations within one acquisition in order to obtain quality cine images of the heart in a single breath-hold. The technique was tested on animal models and human subjects, and the impact of interleaved acquisition on image quality was studied using quantitative and qualitative measures.     

A calibration method for stereo vision sensor with large FOV based on 1D targets

Large FOV (field of view) stereo vision sensor is of great importance in the measurement of large free-form surface. Before using it, the intrinsic and structure parameters of cameras should be calibrated. Traditional methods are mainly based on planar or 3D targets, which are usually expensive and difficult to manufacture especially for large dimension ones. Compared to that the method proposed in this paper is based on 1D (one dimensional) targets, which are easy to operate and with high efficiency. First two 1D targets with multiple feature points are placed randomly, and the cameras acquire multiple images of the targets from different angles of view. With the fixed angle between vectors defined by the two 1D targets we can establish the objective function with intrinsic parameters, which can be later solved by the optimization method. Then the stereo vision sensor with two calibrated cameras is set up, which acquire multiple images of another 1D target with two feature points in unrestrained motion. The initial values of the structure parameters are estimated by the linear method for the known distance between two feature points on the 1D target, while the optimal ones and intrinsic parameters of the stereo vision sensor are estimated with non-linear optimization method by establishing the minimizing function involving all the parameters. The experimental results show that the measurement precision of the stereo vision sensor is 0.046 mm with the working distance of about 3500 mm and the measurement scale of about 4000 mm×3000 mm. The method in this paper is proved suitable for calibration of stereo vision sensor of large-scale measurement field for its easy operation and high efficiency.     

Breath-held MR Cholangiopancreatography (MRCP) using a 3D Dixon fat–water separated balanced steady state free precession sequence

A novel 3D breath-held Dixon fat–water separated balanced steady state free precession (b-SSFP) sequence for MR cholangiopancreatography (MRCP) is described and its potential clinical utility assessed in a series of patients. The main motivation is to develop a robust breath-held alternative to the respiratory gated 3D Fast Spin Echo (FSE) sequence, the current clinical sequence of choice for MRCP. Respiratory gated acquisitions are susceptible to motion artifacts and blurring in patients with significant diaphragmatic drift, erratic respiratory rhythms or sleep apnea. A two point Dixon fat–water separation scheme was developed which eliminates signal loss arising from B₀ inhomogeneity effects and minimizes artifacts from perturbation of the b-SSFP steady state. Preliminary results from qualitative analysis of 49 patients demonstrate robust performance of the 3D Dixon b-SSFP sequence with diagnostic image quality acquired in a 20–24 s breath-hold.     

Optimal techniques for magnetic resonance imaging of the liver using a respiratory navigator-gated three-dimensional spoiled gradient-recalled echo sequence

Purpose

To optimize the navigator-gating technique for the acquisition of high-quality three-dimensional spoiled gradient-recalled echo (3D SPGR) images of the liver during free breathing.

Materials and methods

Ten healthy volunteers underwent 3D SPGR magnetic resonance imaging of the liver using a conventional navigator-gated 3D SPGR (cNAV-3D-SPGR) sequence or an enhanced navigator-gated 3D SPGR (eNAV-3D-SPGR) sequence. No exogenous contrast agent was used. A 20-ms wait period was inserted between the 3D SPGR acquisition component and navigator component of the eNAV-3D-SPGR sequence to allow T1 recovery. Visual evaluation and calculation of the signal-to-noise ratio were performed to compare image quality between the imaging techniques.

Result

The eNAV-3D-SPGR sequence provided better noise properties than the cNAV-3D-SPGR sequence visually and quantitatively. Navigator gating with an acceptance window of 2 mm effectively inhibited respiratory motion artifacts. The widening of the window to 6 mm shortened the acquisition time but increased motion artifacts, resulting in degradation of overall image quality. Neither slice tracking nor incorporation of short breath holding successfully compensated for the widening of the window.

Conclusion

The eNAV-3D-SPGR sequence with an acceptance window of 2 mm provides high-quality 3D SPGR images of the liver.     

A multi-plane scout sequence using flash imaging

A new pulse sequence is presented for multi-plane imaging which simultaneously acquires images in the axial, coronal, and sagittal planes. Combined with FLASH imaging or other similar techniques, T1- and T2-weighted images in all three planes can be acquired in approximately 12 and 30 seconds, respectively. The sequence is easily implemented by modifying a multi-slice sequence to provide a rotation of gradient assignments between slice excitations. This sequence has been used extensively as a rapid "scout" scan. Several clinical examples are included.     

A comparison of fast spin echo and gradient field echo sequences

Optimal angle, fast repeat time, gradient field echo imaging techniques such as FISP (Fast Imaging with Steady Precession) and FLASH (Fast Low Angle Shot) often fail to discriminate disease from healthy tissue for two main reasons. First, T1 and T2 of the affected tissue may increase such that the ratio of T1 to T2 remains nearly unchanged, hence there is no contrast change with FISP. Second, T2 weighted gradient field echo images suffer severely from T2* signal and resolution loss leading to a reduction in C/N. Although FLASH imaging with two separate angles can, in principle, extract the longer T1 tumors, contrast is often not good. To overcome the inhomogeneity and contrast problems, we have implemented a FAst optimal angle spin-echo sequence with a short TE(FATE). For the first echo, FATE has the same contrast properties as FLASH with a slight decrease in signal intensity. The advantage is that the intensity of the signal does not suffer from T2* signal decay, hence improved contrast and disease detection via T2 weighted FATE images is possible. Contrast-to-noise in lesion detection is also considered for CE FAST (Contrast Enhanced Fast), a T2-weighted version of FISP, and HYBRID.     

Estimation of mutual information objective function based on Fourier shift theorem: an application to eddy current distortion correction in diffusion tensor imaging

Diffusion tensor imaging requires correction of eddy current distortion in diffusion-weighted images. An effective retrospective correction approach is to transform a diffusion-weighted image to maximize the mutual information (MI) between the transformed diffusion-weighted image and the corresponding T2-weighted image. In the literature, either linear interpolation or partial volume interpolation is applied to estimate the MI objective function. However, these interpolation methods induce artifacts to the MI objective function, thus compromising correction results. In this work, the MI objective function is estimated based on interpolation using Fourier shift theorem. This method eliminates the artifacts incurred with the aforementioned interpolation methods. The algorithm is further improved by approximating pixel values using their nearest neighbors in the up-sampled spatial domain, resulting in dramatically increased computational efficiency without compromising the correction results. The effects of varying the number of quantization levels and using Parzen window filtering to smooth the MI objective function are also investigated to obtain optimized algorithm parameters. The diffusion tensor image quality after applying the proposed distortion correction method is significantly improved visually.