Multichannel MRI Research Articles

Deep-learning-based motion correction using multichannel MRI data: a study using simulated artifacts in the fastMRI dataset.

Deep learning presents a generalizable solution for motion correction requiring no pulse sequence modifications or additional hardware, but previous networks have all been applied to coil-combined data. Multichannel MRI data provide a degree of spatial encoding that may be useful for motion correction. We hypothesize that incorporating deep learning for motion correction prior to coil combination will improve results. A conditional generative adversarial network was trained using simulated rigid motion artifacts in brain images acquired at multiple sites with multiple contrasts (not limited to healthy subjects). We compared the performance of deep-learning-based motion correction on individual channel images (single-channel model) with that performed after coil combination (channel-combined model). We also investigate simultaneous motion correction of all channel data from an image volume (multichannel model). The single-channel model significantly (p < 0.0001) improved mean absolute error, with an average 50.9% improvement compared with the uncorrected images. This was significantly (p < 0.0001) better than the 36.3% improvement achieved by the channel-combined model (conventional approach). The multichannel model provided no significant improvement in quantitative measures of image quality compared with the uncorrected images. Results were independent of the presence of pathology, and generalizable to a new center unseen during training. Performing motion correction on single-channel images prior to coil combination provided an improvement in performance compared with conventional deep-learning-based motion correction. Improved deep learning methods for retrospective correction of motion-affected MR images could reduce the need for repeat scans if applied in a clinical setting.

New Theory and Faster Computations for Subspace-Based Sensitivity Map Estimation in Multichannel MRI.

Sensitivity map estimation is important in many multichannel MRI applications. Subspace-based sensitivity map estimation methods like ESPIRiT are popular and perform well, though can be computationally expensive and their theoretical principles can be nontrivial to understand. In the first part of this work, we present a novel theoretical derivation of subspace-based sensitivity map estimation based on a linear-predictability/structured low-rank modeling perspective. This results in an estimation approach that is equivalent to ESPIRiT, but with distinct theory that may be more intuitive for some readers. In the second part of this work, we propose and evaluate a set of computational acceleration approaches (collectively known as PISCO) that can enable substantial improvements in computation time (up to ∼100× in the examples we show) and memory for subspace-based sensitivity map estimation.

Deep learning for the prediction of type 2 diabetes mellitus from neck-to-knee Dixon MRI in the UK biobank

Rationale and objectivesWe evaluate the automatic identification of type 2 diabetes from neck-to-knee, two-point Dixon MRI scans with 3D convolutional neural networks on a large, population-based dataset. To this end, we assess the best combination of MRI contrasts and stations for diabetes prediction, and the benefit of integrating risk factors. Materials and methodsSubjects with type 2 diabetes mellitus have been identified in the prospective UK Biobank Imaging study, and a matched control sample has been created to avoid confounding bias. Five-fold cross-validation is used for the evaluation. All scans from the two-point Dixon neck-to-knee sequence have been standardized. A neural network that considers multi-channel MRI input was developed and integrates clinical information in tabular format. An ensemble strategy is used to combine multi-station MRI predictions. A subset with quantitative fat measurements is identified for comparison to prior approaches. ResultsMRI scans from 3406 subjects (mean age, 66.2 years ± 7.1 [standard deviation]; 1128 women) were analyzed with 1703 diabetics. A balanced accuracy of 78.7 %, AUC ROC of 0.872, and an average precision of 0.878 was obtained for the classification of diabetes. The ensemble over multiple Dixon MRI stations yields better performance than selecting the individually best station. Moreover, combining fat and water scans as multi-channel inputs to the networks improves upon just using single contrasts as input. Integrating clinical information about known risk factors of diabetes in the network boosts the performance across all stations and the ensemble. The neural network achieved superior results compared to the prediction based on quantitative MRI measurements. ConclusionsThe developed deep learning model accurately predicted type 2 diabetes from neck-to-knee two-point Dixon MRI scans.

INDICES OF NEUROPLASTICITY AND FUNCTIONAL REORGANIZATION IN MOTOR FUNCTIO RECOVERY AFTER STROKE

The paper reviews data on changes of various indices calculated from multi-channel EEG, MRI, fMRI, and DTI data obtained from post-stroke patients during motor function recovery. The indices are most frequently discussed in literature on the topic of both motor rehabilitation in general and using BCI-based procedures in particular. The dynamics of the indices considered reflects the changes in interhemispheric imbalance during movement, the contribution of different areas and their interaction during motor execution as well as structural reorganization. The role of damaged and intact hemispheres and particular areas in motor recovery is discussed.

A Self-Attention-Guided 3D Deep Residual Network With Big Transfer to Predict Local Failure in Brain Metastasis After Radiotherapy Using Multi-Channel MRI

A noticeable proportion of larger brain metastases (BMs) are not locally controlled after stereotactic radiotherapy, and it may take months before local progression is apparent on standard follow-up imaging. This work proposes and investigates new explainable deep-learning models to predict the radiotherapy outcome for BM. A novel self-attention-guided 3D residual network is introduced for predicting the outcome of local failure (LF) after radiotherapy using the baseline treatment-planning MRI. The 3D self-attention modules facilitate capturing long-range intra/inter slice dependencies which are often overlooked by convolution layers. The proposed model was compared to a vanilla 3D residual network and 3D residual network with CBAM attention in terms of performance in outcome prediction. A training recipe was adapted for the outcome prediction models during pretraining and training the down-stream task based on the recently proposed big transfer principles. A novel 3D visualization module was coupled with the model to demonstrate the impact of various intra/peri-lesion regions on volumetric multi-channel MRI upon the network's prediction. The proposed self-attention-guided 3D residual network outperforms the vanilla residual network and the residual network with CBAM attention in accuracy, F1-score, and AUC. The visualization results show the importance of peri-lesional characteristics on treatment-planning MRI in predicting local outcome after radiotherapy. This study demonstrates the potential of self-attention-guided deep-learning features derived from volumetric MRI in radiotherapy outcome prediction for BM. The insights obtained via the developed visualization module for individual lesions can possibly be applied during radiotherapy planning to decrease the chance of LF.

Transfer learning enhanced generative adversarial networks for multi-channel MRI reconstruction

Deep learning based generative adversarial networks (GAN) can effectively perform image reconstruction with under-sampled MR data. In general, a large number of training samples are required to improve the reconstruction performance of a certain model. However, in real clinical applications, it is difficult to obtain tens of thousands of raw patient data to train the model since saving k-space data is not in the routine clinical flow. Therefore, enhancing the generalizability of a network based on small samples is urgently needed. In this study, three novel applications were explored based on parallel imaging combined with the GAN model (PI-GAN) and transfer learning. The model was pre-trained with public Calgary brain images and then fine-tuned for use in (1) patients with tumors in our center; (2) different anatomies, including knee and liver; (3) different k-space sampling masks with acceleration factors (AFs) of 2 and 6. As for the brain tumor dataset, the transfer learning results could remove the artifacts found in PI-GAN and yield smoother brain edges. The transfer learning results for the knee and liver were superior to those of the PI-GAN model trained with its own dataset using a smaller number of training cases. However, the learning procedure converged more slowly in the knee datasets compared to the learning in the brain tumor datasets. The reconstruction performance was improved by transfer learning both in the models with AFs of 2 and 6. Of these two models, the one with AF = 2 showed better results. The results also showed that transfer learning with the pre-trained model could solve the problem of inconsistency between the training and test datasets and facilitate generalization to unseen data.

A Frequency Translation System for Multi-Channel, Multi-Nuclear MR Spectroscopy.

Most MRI scanners are equipped to receive signals from 1H array coils but few support multi-channel reception for other nuclei. Using receive arrays can provide significant SNR benefits, usually exploited to enable accelerated imaging, but the extension of these arrays to non-1H nuclei has received less attention because of the relative lack of broadband array receivers. Non-1H nuclei often have low sensitivity and stand to benefit greatly from the increase in SNR that arrays can provide. This paper presents a cost-effective approach for adapting standard 1H multi-channel array receivers for use with other nuclei - in this case, 13C. A frequency translation system has been developed that uses active mixers residing at the magnet bore to convert the received signal from a non-1H array to the 1H frequency for reception by the host system receiver. This system has been demonstrated at 4.7T and 7T while preserving SNR and isolation. 1H decoupling, particularly important for 13C detection, can be straightforwardly accommodated. Frequency translation can convert 1H-only multi-channel receivers for use with other nuclei while maintaining SNR and channel isolation while still enabling 1H decoupling. This work allows existing multi-channel MRI receivers to be adapted to receive signals from nuclei other than 1H, allowing for the use of receive arrays for in vivo multi-nuclear NMR.

Joint multi-contrast variational network reconstruction (jVN) with application to rapid 2D and 3D imaging.

To improve the image quality of highly accelerated multi-channel MRI data by learning a joint variational network that reconstructs multiple clinical contrasts jointly. Data from our multi-contrast acquisition were embedded into the variational network architecture where shared anatomical information is exchanged by mixing the input contrasts. Complementary k-space sampling across imaging contrasts and Bunch-Phase/Wave-Encoding were used for data acquisition to improve the reconstruction at high accelerations. At 3T, our joint variational network approach across T1w, T2w and T2-FLAIR-weighted brain scans was tested for retrospective under-sampling atR=6 (2D) andR=4×4 (3D) acceleration. Prospective acceleration was also performed for 3D data where the combined acquisition time for whole brain coverage at 1mm isotropic resolution across three contrasts was less than 3 min. Across all test datasets, our joint multi-contrast network better preserved fine anatomical details with reduced image-blurring when compared to the corresponding single-contrast reconstructions. Improvement in image quality was also obtained through complementary k-space sampling and Bunch-Phase/Wave-Encoding where the synergistic combination yielded the overall best performance as evidenced by exemplary slices and quantitative error metrics. By leveraging shared anatomical structures across the jointly reconstructed scans, our joint multi-contrast approach learnt more efficient regularizers, which helped to retain natural image appearance and avoid over-smoothing. When synergistically combined with advanced encoding techniques, the performance was further improved, enabling up to R=16-fold acceleration with good image quality. This should help pave the way to very rapid high-resolution brain exams.

Radiomic profiles in diffuse glioma reveal distinct subtypes with prognostic value.

To evaluate a radiomic approach for the stratification of diffuse gliomas with distinct prognosis and provide additional resolution of their clinicopathological and molecular characteristics. For this retrospective study, a total of 704 radiomic features were extracted from the multi-channel MRI data of 166 diffuse gliomas. Survival-associated radiomic features were identified and submitted to distinguish glioma subtypes using consensus clustering. Multi-layered molecular data were used to observe the different clinical and molecular characteristics between radiomic subtypes. The relative profiles of an array of immune cell infiltrations were measured gene set variation analysis approach to explore differences in tumor immune microenvironment. A total of 6 categories, including 318 radiomic features were significantly correlated with the overall survival of glioma patients. Two subgroups with distinct prognosis were separated by consensus clustering of radiomic features that significantly associated with survival. Histological stage and molecular factors, including IDH status and MGMT promoter methylation status were significant differences between the two subtypes. Furthermore, gene functional enrichment analysis and immune infiltration pattern analysis also hinted that the inferior prognosis subtype may more response to immunotherapy. A radiomic model derived from multi-parameter MRI of the gliomas was successful in the risk stratification of diffuse glioma patients. These data suggested that radiomics provided an alternative approach for survival estimation and may improve clinical decision-making.

Compressed sensing reconstruction of 7 Tesla 23Na multi-channel breast data using 1H MRI constraint

PurposeTo reduce acquisition time and to improve image quality in sodium magnetic resonance imaging (23Na MRI) using an iterative reconstruction algorithm for multi-channel data sets based on compressed sensing (CS) with anatomical 1H prior knowledge. MethodsAn iterative reconstruction for 23Na MRI with multi-channel receiver coils is presented. Based on CS it utilizes a second order total variation (TV(2)), adopted by anatomical weighting factors (AnaWeTV(2)) obtained from a high-resolution 1H image. A support region is included as additional regularization. Simulated and measured 23Na multi-channel data sets (n = 3) of the female breast acquired at 7 T with different undersampling factors (USF = 1.8/3.6/7.2/14.4) were reconstructed and compared to a conventional gridding reconstruction. The structural similarity was used to assess image quality of the reconstructed simulated data sets and to optimize the weighting factors for the CS reconstruction. ResultsCompared with a conventional TV(2), the AnaWeTV(2) reconstruction leads to an improved image quality due to preserving of known structure and reduced partial volume effects. An additional incorporated support region shows further improvements for high USFs. Since the decrease in image quality with higher USFs is less pronounced compared to a conventional gridding reconstruction, proposed algorithm is beneficial especially for higher USFs. Acquisition time can be reduced by a factor of 4 (USF = 7.2), while image quality is still similar to a nearly fully sampled (USF = 1.8) gridding reconstructed data set. ConclusionEspecially for high USFs, the proposed algorithm allows improved image quality for multi-channel 23Na MRI data sets.

Uniform combined reconstruction of multichannel 7T knee MRI receive coil data without the use of a reference scan.

MR image intensity nonuniformity is often observed at 7T. Reference scans from the body coil used for uniformity correction at lower field strengths are typically not available at 7T. To evaluate the efficacy of a novel algorithm, Uniform Combined Reconstruction (UNICORN), to correct receive coil-induced nonuniformity in musculoskeletal 7T MRI without the use of a reference scan. Retrospective image analysis study. MRI data of 20 subjects was retrospectively processed offline. Field Strength/Sequence: Knees of 20 subjects were imaged at 7T with a single-channel transmit, 28-channel phased-array receive knee coil. A turbo-spin-echo sequence was used to acquire 33 series of images. Three fellowship-trained musculoskeletal radiologists with cumulative experience of 42 years reviewed the images. The uniformity, contrast, signal-to-noise ratio (SNR), and overall image quality were evaluated for images with no postprocessing, images processed with N4 bias field correction algorithm, and the UNICORN algorithm. Intraclass correlation coefficient (ICC) was used for measuring the interrater reliability. ICC and 95% confidence intervals (CIs) were calculated using the R statistical package employing a two-way mixed-effects model based on a mean rating (k = 3) for absolute agreement. The Wilcoxon signed-rank test with continuity correction was used for analyzing the overall image quality scores. UNICORN was preferred among the three methods evaluated for uniformity in 97.9% of the pooled ratings, with excellent interrater agreement (ICC of 0.98, CI 0.97-0.99). UNICORN was also rated better than N4 for contrast and equivalent to N4 in SNR with ICCs of 0.80 (CI 0.72-0.86) and 0.67 (CI 0.54-0.77), respectively. The overall image quality scores for UNICORN were significantly higher than N4 (P < 6 × 10-13 ), with good to excellent interrater agreement (ICC 0.90, CI 0.86-0.93). Without the use of a reference scan, UNICORN provides better image uniformity, contrast, and overall image quality at 7T compared with the N4 bias field-correction algorithm. 4 Technical Efficacy: Stage 1 J. Magn. Reson. Imaging 2019;50:1534-1544.

MODEL BASED IMAGE RECONSTRUCTION USING DEEP LEARNED PRIORS (MODL).

We introduce a model-based image reconstruction framework, where we use a deep convolution neural network (CNN) based regularization prior. We rely on a recursive algorithm, which alternates between a CNN based denoising step and enforcement of data consistency. Unrolling the recursive algorithm yields a deep network that is trained using backpropagation. The unique aspect of this method is the use of the same CNN weights at each iteration, which makes the resulting structure consistent with the model-based formulation. Also, this approach reduces the number of trainable parameters, which hence lower the amount of training data needed. The use of a forward model also reduces the size of the network and enables the exploitation additional prior information available from calibration data. The use of the framework for multichannel MRI reconstruction provides improved reconstructions, compared to other state-of-the-art methods.

Compressed sensing MRI reconstruction from 3D multichannel data using GPUs.

To accelerate iterative reconstructions of compressed sensing (CS) MRI from 3D multichannel data using graphics processing units (GPUs). The sparsity of MRI signals and parallel array receivers can reduce the data acquisition requirements. However, iterative CS reconstructions from data acquired using an array system may take a significantly long time, especially for a large number of parallel channels. This paper presents an efficient method for CS-MRI reconstruction from 3D multichannel data using GPUs. In this method, CS reconstructions were simultaneously processed in a channel-by-channel fashion on the GPU, in which the computations of multiple-channel 3D-CS reconstructions are highly parallelized. The final image was then produced by a sum-of-squares method on the central processing unit. Implementation details including algorithm, data/memory management, and parallelization schemes are reported in the paper. Both simulated data and in vivo MRI array data were tested. The results showed that the proposed method can significantly improve the image reconstruction efficiency, typically shortening the runtime by a factor of 30. Using low-cost GPUs and an efficient algorithm allowed the 3D multislice compressive-sensing reconstruction to be performed in less than 1 s. The rapid reconstructions are expected to help bring high-dimensional, multichannel parallel CS MRI closer to clinical applications. Magn Reson Med 78:2265-2274, 2017. © 2017 International Society for Magnetic Resonance in Medicine.

Efficient multi-scale 3D CNN with fully connected CRF for accurate brain lesion segmentation.

We propose a dual pathway, 11-layers deep, three-dimensional Convolutional Neural Network for the challenging task of brain lesion segmentation. The devised architecture is the result of an in-depth analysis of the limitations of current networks proposed for similar applications. To overcome the computational burden of processing 3D medical scans, we have devised an efficient and effective dense training scheme which joins the processing of adjacent image patches into one pass through the network while automatically adapting to the inherent class imbalance present in the data. Further, we analyze the development of deeper, thus more discriminative 3D CNNs. In order to incorporate both local and larger contextual information, we employ a dual pathway architecture that processes the input images at multiple scales simultaneously. For post-processing of the network's soft segmentation, we use a 3D fully connected Conditional Random Field which effectively removes false positives. Our pipeline is extensively evaluated on three challenging tasks of lesion segmentation in multi-channel MRI patient data with traumatic brain injuries, brain tumours, and ischemic stroke. We improve on the state-of-the-art for all three applications, with top ranking performance on the public benchmarks BRATS 2015 and ISLES 2015. Our method is computationally efficient, which allows its adoption in a variety of research and clinical settings. The source code of our implementation is made publicly available.

Vectorial total generalized variation for accelerated multi-channel multi-contrast MRI

PurposeTo develop and implement an efficient reconstruction technique to improve accelerated multi-channel multi-contrast MRI. Theory and MethodsThe vectorial total generalized variation (TGV) operator is used as a regularizer for the sensitivity encoding (SENSE) technique to improve image quality of multi-channel multi-contrast MRI. The alternating direction method of multipliers (ADMM) is used to efficiently reconstruct the data. The performance of the proposed method (MC-TGV-SENSE) is assessed on two healthy volunteers at several acceleration factors. ResultsAs demonstrated on the in vivo results, MC-TGV-SENSE had the lowest root-mean-square error (RMSE), highest structural similarity index, and best visual quality at all acceleration factors, compared to other methods under consideration. MC-TGV-SENSE yielded up to 17.3% relative RMSE reduction compared to the widely used total variation regularized SENSE. Furthermore, we observed that the reconstruction time of MC-TGV-SENSE is reduced by approximately a factor of two with comparable RMSEs by using the proposed ADMM-based algorithm as opposed to the more commonly used Chambolle–Pock primal-dual algorithm for the TGV-based reconstruction. ConclusionMC-TGV-SENSE is a better alternative than the existing reconstruction methods for accelerated multi-channel multi-contrast MRI. The proposed method exploits shared information among the images (MC), mitigates staircasing artifacts (TGV), and uses the encoding power of multiple receiver coils (SENSE).

3D active surfaces for liver segmentation in multisequence MRI images

Biopsies for diagnosis can sometimes be replaced by non-invasive techniques such as CT and MRI. Surgeons require accurate and efficient methods that allow proper segmentation of the organs in order to ensure the most reliable intervention planning. Automated liver segmentation is a difficult and open problem where CT has been more widely explored than MRI. MRI liver segmentation represents a challenge due to the presence of characteristic artifacts, such as partial volumes, noise and low contrast. In this paper, we present a novel method for multichannel MRI automatic liver segmentation. The proposed method consists of the minimization of a 3D active surface by means of the dual approach to the variational formulation of the underlying problem. This active surface evolves over a probability map that is based on a new compact descriptor comprising spatial and multisequence information which is further modeled by means of a liver statistical model. This proposed 3D active surface approach naturally integrates volumetric regularization in the statistical model. The advantages of the compact visual descriptor together with the proposed approach result in a fast and accurate 3D segmentation method. The method was tested on 18 healthy liver studies and results were compared to a gold standard made by expert radiologists. Comparisons with other state-of-the-art approaches are provided by means of nine well established quality metrics. The obtained results improve these methodologies, achieving a Dice Similarity Coefficient of 98.59.

Spherical Deconvolution of Multichannel Diffusion MRI Data with Non-Gaussian Noise Models and Spatial Regularization.

Spherical deconvolution (SD) methods are widely used to estimate the intra-voxel white-matter fiber orientations from diffusion MRI data. However, while some of these methods assume a zero-mean Gaussian distribution for the underlying noise, its real distribution is known to be non-Gaussian and to depend on many factors such as the number of coils and the methodology used to combine multichannel MRI signals. Indeed, the two prevailing methods for multichannel signal combination lead to noise patterns better described by Rician and noncentral Chi distributions. Here we develop a Robust and Unbiased Model-BAsed Spherical Deconvolution (RUMBA-SD) technique, intended to deal with realistic MRI noise, based on a Richardson-Lucy (RL) algorithm adapted to Rician and noncentral Chi likelihood models. To quantify the benefits of using proper noise models, RUMBA-SD was compared with dRL-SD, a well-established method based on the RL algorithm for Gaussian noise. Another aim of the study was to quantify the impact of including a total variation (TV) spatial regularization term in the estimation framework. To do this, we developed TV spatially-regularized versions of both RUMBA-SD and dRL-SD algorithms. The evaluation was performed by comparing various quality metrics on 132 three-dimensional synthetic phantoms involving different inter-fiber angles and volume fractions, which were contaminated with noise mimicking patterns generated by data processing in multichannel scanners. The results demonstrate that the inclusion of proper likelihood models leads to an increased ability to resolve fiber crossings with smaller inter-fiber angles and to better detect non-dominant fibers. The inclusion of TV regularization dramatically improved the resolution power of both techniques. The above findings were also verified in human brain data.

Image reconstruction from phased-array data based on multichannel blind deconvolution

In this paper we consider image reconstruction from fully sampled multichannel phased array MRI data without knowledge of the coil sensitivities. To overcome the non-uniformity of the conventional sum-of-square reconstruction, a new framework based on multichannel blind deconvolution (MBD) is developed for joint estimation of the image function and the sensitivity functions in image domain. The proposed approach addresses the non-uniqueness of the MBD problem by exploiting the smoothness of both functions in the image domain through regularization. Results using simulation, phantom and in vivo experiments demonstrate that the reconstructions by the proposed algorithm are more uniform than those by the existing methods.

Inductively coupled wireless RF coil arrays

As the number of coils increases in multi-channel MRI receiver-coil arrays, RF cables and connectors become increasingly bulky and heavy, degrading patient comfort and slowing workflow. Inductive coupling of signals provides an attractive “wireless” approach, with the potential to reduce coil weight and cost while simplifying patient setup. In this work, multi-channel inductively coupled anterior arrays were developed and characterized for 1.5T imaging. These comprised MR receiver coils inductively (or “wirelessly”) linked to secondary or “sniffer” coils whose outputs were transmitted via preamps to the MR system cabinet. The induced currents in the imaging coils were blocked by passive diode circuits during RF transmit. The imaging arrays were totally passive, obviating the need to deliver power to the coils, and providing lightweight, untethered signal reception with easily positioned coils. Single-shot fast spin echo images were acquired from 5 volunteers using a 7-element inductively coupled coil array and a conventionally cabled 7-element coil array of identical geometry, with the inductively-coupled array showing a relative signal-to-noise ratio of 0.86 +/− 0.07. The concept was extended to a larger 9-element coil array to demonstrate the effect of coil element size on signal transfer and RF-transmit blocking.

Multichannel MRI labeling of mammalian cells by switchable nanocarriers for hyperpolarized xenon.

We demonstrate a concept for multichannel MRI cell-labeling using encapsulated laser-polarized xenon. Conceptually different Xe trapping properties of two nanocarriers, namely macrocyclic cages as individual hosts or compartmentalization into nanodroplets, ensure a large chemical shift separation for Xe bound in either of the carriers even after cellular internalization. Two differently labeled mammalian cell populations were imaged by frequency selective saturation transfer resulting in a switchable "two-color" xenon-MRI contrast at micro- to nanomolar Xe carrier concentrations.