Determination of the appropriate b value and number of gradient directions for high‐angular‐resolution diffusion‐weighted imaging

High‐angular‐resolution diffusion‐weighted imaging (HARDI) is one of the most common MRI acquisition schemes for use with higher order models of diffusion. However, the optimal b value and number of diffusion‐weighted (DW) directions for HARDI are still undetermined, primarily as a result of the large number of available reconstruction methods and corresponding parameters, making it impossible to identify a single criterion by which to assess performance. In this study, we estimate the minimum number of DW directions and optimal b values required for HARDI by focusing on the angular frequency content of the DW signal itself. The spherical harmonic (SH) series provides the spherical analogue of the Fourier series, and can hence be used to examine the angular frequency content of the DW signal. Using high‐quality data acquired along 500 directions over a range of b values, we estimate that SH terms above l = 8 are negligible in practice for b values up to 5000 s/mm², implying that a minimum of 45 DW directions is sufficient to fully characterise the DW signal. l > 0 SH terms were found to increase as a function of b value, levelling off at b = 3000 s/mm², suggesting that this value already provides the highest achievable angular resolution. In practice, it is recommended to acquire more than the minimum of 45 DW directions to avoid issues with imperfections in the uniformity of the DW gradient directions and to meet signal‐to‐noise requirements of the intended reconstruction method. Copyright © 2013 John Wiley & Sons, Ltd.     

An intravoxel oriented flow model for diffusion‐weighted imaging of the kidney

By combining intravoxel incoherent motion (IVIM) and diffusion tensor imaging (DTI) we introduce a new diffusion model called intravoxel oriented flow (IVOF) that accounts for anisotropy of diffusion and the flow‐related signal. An IVOF model using a simplified apparent flow fraction tensor (IVOF_f) is applied to diffusion‐weighted imaging of human kidneys. The kidneys of 13 healthy volunteers were examined on a 3 T scanner. Diffusion‐weighted images were acquired with six b values between 0 and 800 s/mm² and 30 diffusion directions. Diffusivity and flow fraction were calculated for different diffusion models. The Akaike information criterion was used to compare the model fit of the proposed IVOF_f model to IVIM and DTI. In the majority of voxels the proposed IVOF_f model with a simplified apparent flow fraction tensor performs better than IVIM and DTI. Mean diffusivity is significantly higher in DTI compared with models that account for the flow‐related signal. The fractional anisotropy of diffusion is significantly reduced when flow fraction is considered to be anisotropic. Anisotropy of the apparent flow fraction tensor is significantly higher in the renal medulla than in the cortex region. The IVOF_f model describes diffusion‐weighted data in the human kidney more accurately than IVIM or DTI. The apparent flow fraction in the kidney proved to be anisotropic.     

A multi‐compartmental SE‐BOLD interpretation for stimulus‐related signal changes in diffusion‐weighted functional MRI

A new interpretation is proposed for stimulus‐induced signal changes in diffusion‐weighted functional MRI. T₂‐weighted spin‐echo echo‐planar images were acquired at different diffusion‐weightings while visual stimulation was presented to human volunteers. The amplitudes of the positive stimulus‐correlated response and post‐stimulus undershoot (PSU) in the functional time‐courses were found to follow different trends as a function of b‐value. Data were analysed using a three‐compartment signal model, with one compartment being purely vascular and the other two dominated by fast‐ and slow‐diffusing molecules in the brain tissue. The diffusion coefficients of the tissue were assumed to be constant throughout the experiments. It is shown that the stimulus‐induced signal changes can be decomposed into independent contributions originating from each of the three compartments. After decomposition, the fast‐diffusion phase displays a substantial PSU, while the slow‐diffusion phase demonstrates a highly reproducible and stimulus‐correlated time‐course with minimal undershoot. The decomposed responses are interpreted in terms of the spin‐echo blood oxygenation level dependent (SE‐BOLD) effect, and it is proposed that the signal produced by fast‐ and slow‐diffusing molecules reflect a sensitivity to susceptibility changes in arteriole/venule‐ and capillary‐sized vessels, respectively. This interpretation suggests that diffusion‐weighted SE‐BOLD imaging may provide subtle information about the haemodynamic and neuronal responses. Copyright © 2009 John Wiley & Sons, Ltd.     

The quest for high spatial resolution diffusion‐weighted imaging of the human brain in vivo

Diffusion‐weighted imaging, a contrast unique to MRI, is used for assessment of tissue microstructure in vivo. However, this exquisite sensitivity to finer scales far above imaging resolution comes at the cost of vulnerability to errors caused by sources of motion other than diffusion motion. Addressing the issue of motion has traditionally limited diffusion‐weighted imaging to a few acquisition techniques and, as a consequence, to poorer spatial resolution than other MRI applications. Advances in MRI imaging methodology have allowed diffusion‐weighted MRI to push to ever higher spatial resolution. In this review we focus on the pulse sequences and associated techniques under development that have pushed the limits of image quality and spatial resolution in diffusion‐weighted MRI.     

Single‐ and double‐Diffusion encoding MRI for studying ex vivo apparent axon diameter distribution in spinal cord white matter

Mapping average axon diameter (AAD) and axon diameter distribution (ADD) in neuronal tissues non‐invasively is a challenging task that may have a tremendous effect on our understanding of the normal and diseased central nervous system (CNS). Water diffusion is used to probe microstructure in neuronal tissues, however, the different water populations and barriers that are present in these tissues turn this into a complex task. Therefore, it is not surprising that recently we have witnessed a burst in the development of new approaches and models that attempt to obtain, non‐invasively, detailed microstructural information in the CNS. In this work, we aim at challenging and comparing the microstructural information obtained from single diffusion encoding (SDE) with double diffusion encoding (DDE) MRI. We first applied SDE and DDE MR spectroscopy (MRS) on microcapillary phantoms and then applied SDE and DDE MRI on an ex vivo porcine spinal cord (SC), using similar experimental conditions. The obtained diffusion MRI data were fitted by the same theoretical model, assuming that the signal in every voxel can be approximated as the superposition of a Gaussian‐diffusing component and a series of restricted components having infinite cylindrical geometries. The diffusion MRI results were then compared with histological findings. We found a good agreement between the fittings and the experimental data in white matter (WM) voxels of the SC in both diffusion MRI methods. The microstructural information and apparent AADs extracted from SDE MRI were found to be similar or somewhat larger than those extracted from DDE MRI especially when the diffusion time was set to 40 ms. The apparent ADDs extracted from SDE and DDE MRI show reasonable agreement but somewhat weaker correspondence was observed between the diffusion MRI results and histology. The apparent subtle differences between the microstructural information obtained from SDE and DDE MRI are briefly discussed.     

Fast low‐angle shot diffusion tensor imaging with stimulated echo encoding in the muscle of rabbit shank

In the past, spin‐echo (SE) echo planar imaging(EPI)‐based diffusion tensor imaging (DTI) has been widely used to study the fiber structure of skeletal muscles in vivo. However, this sequence has several shortcomings when measuring restricted diffusion in small animals, such as its sensitivity to susceptibility‐related distortions and a relatively short applicable diffusion time. To address these limitations, in the current work, a stimulated echo acquisition mode (STEAM) MRI technique, in combination with fast low‐angle shot (FLASH) readout (turbo‐STEAM MRI), was implemented and adjusted for DTI in skeletal muscles. Signal preparation using stimulated echoes enables longer effective diffusion times, and thus the detection of restricted diffusion within muscular tissue with intracellular distances up to 100 µm. Furthermore, it has a reduced penalty for fast T₂ muscle signal decay, but at the expense of 50% signal loss compared with a SE preparation. Turbo‐STEAM MRI facilitates high‐resolution DTI of skeletal muscle without introducing susceptibility‐related distortions. To demonstrate its applicability, we carried out rabbit in vivo measurements on a human whole‐body 3 T scanner. DTI parameters of the shank muscles were extracted, including the apparent diffusion coefficient, fractional anisotropy, eigenvalues and eigenvectors. Eigenvectors were used to calculate maps of structural parameters, such as the planar index and the polar coordinates θ and ? of the largest eigenvector. These parameters were compared between three muscles. θ and ? showed clear differences between the three muscles, reflecting different pennation angles of the underlying fiber structures. Fiber tractography was performed to visualize and analyze the architecture of skeletal pennate muscles. Optimization of tracking parameters and utilization of T₂‐weighted images for improved muscle boundary detection enabled the determination of additional parameters, such as the mean fiber length. The presented results support the applicability of turbo‐STEAM MRI as a promising method for quantitative DTI analysis and fiber tractography in skeletal muscles. Copyright © 2013 John Wiley & Sons, Ltd.     

Comparison of interpolation methods of predominant cardiomyocyte orientation from in vivo and ex vivo cardiac diffusion tensor imaging data

Cardiac electrophysiology and cardiac mechanics both depend on the average cardiomyocyte long-axis orientation. In the realm of personalized medicine, knowledge of the patient-specific changes in cardiac microstructure plays a crucial role. Patient-specific computational modelling has emerged as a tool to better understand disease progression. In vivo cardiac diffusion tensor imaging (cDTI) is a vital tool to non-destructively measure the average cardiomyocyte long-axis orientation in the heart. However, cDTI suffers from long scan times, rendering volumetric, high-resolution acquisitions challenging. Consequently, interpolation techniques are needed to populate bio-mechanical models with patient-specific average cardiomyocyte long-axis orientations. In this work, we compare five interpolation techniques applied to in vivo and ex vivo porcine input data. We compare two tensor interpolation approaches, one rule-based approximation, and two data-driven, low-rank models. We demonstrate the advantage of tensor interpolation techniques, resulting in lower interpolation errors than do low-rank models and rule-based methods adapted to cDTI data. In an ex vivo comparison, we study the influence of three imaging parameters that can be traded off against acquisition time: in-plane resolution, signal to noise ratio, and number of acquired short-axis imaging slices.     

BOLD background gradient contributions in diffusion‐weighted fMRI—comparison of spin‐echo and twice‐refocused spin‐echo sequences

The interaction (‘cross terms’) between diffusion‐weighting gradients and susceptibility‐induced background gradient fields around vessels has an impact on apparent diffusion coefficient (ADC) measurements and diffusion‐weighted functional magnetic resonance imaging (DFMRI) experiments. Monte‐Carlo (MC) simulations numerically integrating the Bloch equations for a large number of random walks in a vascular model were used to investigate to what extent such interactions would influence the extravascular signal change as well as the ADC change observed in DFMRI experiments. The vascular model consists of a set of independent, randomly oriented, infinite cylinders whose internal magnetic susceptibility varies as the state changes between rest and activation. In such a network, the cross terms result in the observation of a functional increase in ADC accompanied by a descending percent signal change with increasing diffusion weighting. It is shown that the twice‐refocused spin‐echo sequence permits sufficient yet not total suppression of such effects compared to the standard Stejskal‐Tanner spin‐echo diffusion weighting under experimentally relevant conditions. Copyright © 2010 John Wiley & Sons, Ltd.     

Intraventricular temperature measured by diffusion‐weighted imaging compared with brain parenchymal temperature measured by MRS in vivo

We examined and compared the temperatures of the intraventricular cerebrospinal fluid (T_v) and the brain parenchyma (T_p) using MRI, with reference to the tympanic membrane temperature (T_t) in healthy subjects. We estimated T_v and T_p values from data gathered simultaneously by MR diffusion‐weighted imaging (DWI) and MRS, respectively, in 35 healthy volunteers (17 males, 18 females; age 25–78 years). We also obtained T_t values just before each MR examination to evaluate the relationships among the three temperatures. There were significant positive correlations between T_v and T_p (R = 0.611, p < 0.001). The correlation was also significant after correction for T_t (R = 0.642, p < 0.001). There was no significant correlation between T_v and T_t or between T_p and T_t in the men or the women. Negative correlations were found between T_v and age and between T_p and age in the males but not females. DWI thermometry seems to reflect the intracranial environment as accurately as MRS thermometry. An age‐dependent decline in temperature was evident in our male subjects by both DWI and MRS thermometry, probably due to the decrease in cerebral metabolism with age. Copyright © 2016 John Wiley & Sons, Ltd.     

Tensor estimation for double‐pulsed diffusional kurtosis imaging

Double‐pulsed diffusional kurtosis imaging (DP‐DKI) represents the double diffusion encoding (DDE) MRI signal in terms of six‐dimensional (6D) diffusion and kurtosis tensors. Here a method for estimating these tensors from experimental data is described. A standard numerical algorithm for tensor estimation from conventional (i.e. single diffusion encoding) diffusional kurtosis imaging (DKI) data is generalized to DP‐DKI. This algorithm is based on a weighted least squares (WLS) fit of the signal model to the data combined with constraints designed to minimize unphysical parameter estimates. The numerical algorithm then takes the form of a quadratic programming problem. The principal change required to adapt the conventional DKI fitting algorithm to DP‐DKI is replacing the three‐dimensional diffusion and kurtosis tensors with the 6D tensors needed for DP‐DKI. In this way, the 6D diffusion and kurtosis tensors for DP‐DKI can be conveniently estimated from DDE data by using constrained WLS, providing a practical means for condensing DDE measurements into well‐defined mathematical constructs that may be useful for interpreting and applying DDE MRI. Data from healthy volunteers for brain are used to demonstrate the DP‐DKI tensor estimation algorithm. In particular, representative parametric maps of selected tensor‐derived rotational invariants are presented.     

DTI of human skeletal muscle: the effects of diffusion encoding parameters,signal‐to‐noise ratio and T2 on tensor indices and fiber tracts

In this study, we have performed simulations to address the effects of diffusion encoding parameters, signal‐to‐noise ratio (SNR) and T₂ on skeletal muscle diffusion tensor indices and fiber tracts. Where appropriate, simulations were corroborated and validated by in vivo diffusion tensor imaging (DTI) of human skeletal muscle. Specifically, we have addressed: (i) the accuracy and precision of the diffusion parameters and eigenvectors at different SNR levels; (ii) the effects of the diffusion gradient direction encoding scheme; (iii) the optimal b value for diffusion tensor estimation; (iv) the effects of changes in skeletal muscle T₂; and, finally, the influence of SNR on fiber tractography and derived (v) fiber lengths, (vi) pennation angles and (vii) fiber curvatures. We conclude that accurate DTI of skeletal muscle requires an SNR of at least 25, a b value of between 400 and 500 s/mm², and data acquired with at least 12 diffusion gradient directions homogeneously distributed on half a sphere. Furthermore, for DTI studies focusing on skeletal muscle injury or pathology, apparent changes in the diffusion parameters need to be interpreted with great care in view of the confounding effects of T₂, particularly for moderate to low SNR values. Copyright © 2013 John Wiley & Sons, Ltd.     

Interrelations of muscle functional MRI,diffusion‐weighted MRI and 31P‐MRS in exercised lower back muscles

Exercise‐induced changes of transverse proton relaxation time (T₂), tissue perfusion and metabolic turnover were investigated in the lower back muscles of volunteers by applying muscle functional MRI (mfMRI) and diffusion‐weighted imaging (DWI) before and after as well as dynamic ³¹P‐MRS during the exercise. Inner (M. multifidus, MF) and outer lower back muscles (M. erector spinae, ES) were examined in 14 healthy young men performing a sustained isometric trunk‐extension. Significant phosphocreatine (PCr) depletions ranging from 30% (ES) to 34% (MF) and Pi accumulations between 95% (left ES) and 120%–140% (MF muscles and right ES) were observed during the exercise, which were accompanied by significantly decreased pH values in all muscles (?pH ≈ –0.05). Baseline T₂ values were similar across all investigated muscles (approximately 27 ms at 3 T), but revealed right–left asymmetric increases (T₂,_inc) after the exercise (right ES/MF: T₂,_inc = 11.8/9.7%; left ES/MF: T₂,_inc = 4.6/8.9%). Analyzed muscles also showed load‐induced increases in molecular diffusion D (p = .007) and perfusion fraction f (p = .002). The latter parameter was significantly higher in the MF than in the ES muscles both at rest and post exercise. Changes in PCr (p = .03), diffusion (p < .01) and perfusion (p = .03) were strongly associated with T_2,inc, and linear mixed model analysis revealed that changes in PCr and perfusion both affect T_2,inc (p < .001). These findings support previous assumptions that T₂ changes are not only an intra‐cellular phenomenon resulting from metabolic stress but are also affected by increased perfusion in loaded muscles. Copyright © 2014 John Wiley & Sons, Ltd.     

High‐resolution diffusion tensor imaging in cervical spondylotic myelopathy: a preliminary follow‐up study

Diffusion imaging is a promising technique as it can provide microstructural tissue information and thus potentially show viable changes in spinal cord. However, the traditional single‐shot imaging method is limited as a result of various image artifacts. In order to improve measurement accuracy, we used a newly developed, multi‐shot, high‐resolution, diffusion tensor imaging (DTI) method to investigate diffusion metric changes and compare them with T₂‐weighted (T2W) images before and after decompressive surgery for cervical spondylotic myelopathy (CSM). T2W imaging, single‐shot DTI and multi‐shot DTI were employed to scan seven patients with CSM before and 3 months after decompressive surgery. High signal intensities were scored using the T2 W images. DTI metrics, including fractional anisotropy (FA), axial diffusivity (AD), radial diffusivity (RD) and mean diffusivity (MD), were quantified and compared pre‐ and post‐surgery. In addition, the relationship between imaging metrics and neurological assessments was examined. The reproducibility of multi‐shot DTI was also assessed in 10 healthy volunteers. Post‐surgery, the mean grade of cervical canal stenosis was reduced from grade 3 to normal after 3 months. Compared with single‐shot DTI, multi‐shot DTI provided better images with lower artifact levels, especially following surgery, as a result of reduced artifacts from metal implants. The new method also showed acceptable reproducibility. Both FA and RD values from the new acquisition showed significant differences post‐surgery (FA, p = 0.026; RD, p = 0.048). These changes were consistent with neurological assessments. In contrast, T2W images did not show significant changes before and after surgery. Multi‐shot diffusion imaging showed improved image quality over single‐shot DWI, and presented superior performance in diagnosis and recovery monitoring for patients with CSM compared with T2W imaging. DTI metrics can reflect the pathological conditions of spondylotic spinal cord quantitatively and may serve as a sensitive biomarker for potential CSM management.     

Double‐pulsed diffusional kurtosis imaging

Diffusional kurtosis imaging (DKI) is extended to double‐pulsed‐field‐gradient (d‐PFG) diffusion MRI sequences. This gives a practical approach for acquiring and analyzing d‐PFG data. In particular, the leading d‐PFG effects, beyond what conventional single‐pulsed field gradient (s‐PFG) provides, are interpreted in terms of the kurtosis for a diffusion displacement probability density function (dPDF) in a six‐dimensional (6D) space. The 6D diffusional kurtosis encodes the unique information provided by d‐PFG sequences up to second order in the b‐value. This observation leads to a compact expression for the signal magnitude, and it suggests novel data acquisition and analysis methods. Double‐pulsed DKI (DP‐DKI) is demonstrated for in vivo mouse brain with d‐PFG data obtained at 7 T. Copyright © 2014 John Wiley & Sons, Ltd.     

Diffusion tensor imaging and T2 relaxometry of bilateral lumbar nerve roots: feasibility of in‐plane imaging

Lower back pain is a common problem frequently encountered without specific biomarkers that correlate well with an individual patient's pain generators. MRI quantification of diffusion and T₂ relaxation properties may provide novel insight into the mechanical and inflammatory changes that occur in the lumbosacral nerve roots in patients with lower back pain. Accurate imaging of the spinal nerve roots is difficult because of their small caliber and oblique course in all three planes. Two‐dimensional in‐plane imaging of the lumbosacral nerve roots requires oblique coronal imaging with large field of view (FOV) in both dimensions, resulting in severe geometric distortions using single‐shot echo planar imaging (EPI) techniques. The present work describes initial success using a reduced‐FOV single‐shot spin‐echo EPI acquisition to obtain in‐plane diffusion tensor imaging (DTI) and T₂ mapping of the bilateral lumbar nerve roots at the L4 level of healthy subjects, minimizing partial volume effects, breathing artifacts and geometric distortions. A significant variation in DTI and T₂ mapping metrics is also reported along the course of the normal nerve root. The fractional anisotropy is statistically significantly lower in the dorsal root ganglia (0.287 ± 0.068) than in more distal regions in the spinal nerve (0.402 ± 0.040) (p < 10^–5). The T₂ relaxation value is statistically significantly higher in the dorsal root ganglia (78.0 ± 11.9 ms) than in more distal regions in the spinal nerve (59.5 ± 7.4 ms) (p < 10^–5). The quantification of nerve root DTI and T₂ properties using the proposed methodology may identify the specific site of any degenerative and inflammatory changes along the nerve roots of patients with lower back pain. Copyright © 2012 John Wiley & Sons, Ltd.     

Diffusion‐weighted imaging of the prostate and rectal wall: comparison of biexponential and monoexponential modelled diffusion and associated perfusion coefficients

This study compares parameters from monoexponential and biexponential modelling of diffusion‐weighted imaging of normal and malignant prostate tissue and normal rectal wall tissues. Fifty men with Stage Ic prostate cancer were studied using endorectal T₂‐weighted imaging and diffusion‐weighted imaging with 11 diffusion‐sensitive values (b‐values = 0, 1, 2, 4, 10, 20, 50, 100, 200, 400, 800 s/mm²). Regions of interest were drawn within non‐malignant central gland and peripheral zone, malignant prostate tissue and normal rectal wall tissue. Both a monoexponential and biexponential model was fitted over various b‐value ranges, giving an apparent diffusion coefficient (ADC) from the monoexponential model and a diffusion coefficient, perfusion coefficient and perfusion fraction from the biexponential model. In all tissues, over the full range of b‐values, the ADC from the monoexponential model was significantly higher than the corresponding diffusion coefficient from the biexponential model. As the minimum b‐value increased, the ADC decreased and was equal to the diffusion coefficient for some b‐value ranges. The biexponential model best described the data when low b‐values were included, suggesting that there is a fast perfusion component. Neither model could distinguish between benign prostate tissues on the basis of diffusion coefficients, but the rectal wall tissue and malignant prostate tissue had significantly lower diffusion coefficients than normal prostate tissues. Perfusion coefficients and fractions were highly variable within the population, so their clinical utility may be limited, but removal of this variable perfusion component from reported diffusion coefficients is important when attributing clinical differences to diffusion within tissues. Copyright © 2008 John Wiley & Sons, Ltd.     

Detection of cancer in cervical tissue biopsies using mobile lipid resonances measured with diffusion‐weighted 1H magnetic resonance spectroscopy

The purpose of this study was to implement a diffusion‐weighted sequence for visualisation of mobile lipid resonances (MLR) using high resolution magic angle spinning (HR‐MAS) ¹H MRS and to evaluate its use in establishing differences between tissues from patients with cervical carcinoma that contain cancer from those that do not. A stimulated echo sequence with bipolar gradients was modified to allow T₁ and T₂ measurements and optimised by recording signal loss in HR‐MAS spectra as a function of gradient strength in model lipids and tissues. Diffusion coefficients, T₁ and apparent T₂ relaxation times were measured in model lipid systems. MLR profiles were characterised in relation to T₁ and apparent T₂ relaxation in human cervical cancer tissue samples. Diffusion‐weighted (DW) spectra of cervical biopsies were quantified and peak areas analysed using linear discriminant analysis (LDA). The optimised sequence reduced spectral overlap by suppressing signals originating from low molecular weight metabolites and non‐lipid contributions. Significantly improved MLR visualisation allowed visualisation of peaks at 0.9, 1.3, 1.6, 2.0, 2.3, 2.8, 4.3 and 5.3 ppm. MLR analysis of DW spectra showed at least six peaks arising from saturated and unsaturated lipids and those arising from triglycerides. Significant differences in samples containing histologically confirmed cancer were seen for peaks at 0.9 (p < 0.006), 1.3 (p < 0.04), 2.0 (p < 0.03), 2.8 (p < 0.003) and 4.3 ppm (p < 0.0002). LDA analysis of MLR peaks from DW spectra almost completely separated two clusters of cervical biopsies (cancer, ‘no‐cancer’), reflecting underlying differences in MLR composition. Generated Receiver Operating Characteristic (ROC) curves and calculated area under the curve (0.962) validated high sensitivity and specificity of the technique. Diffusion‐weighting of HR‐MAS spectroscopic sequences is a useful method for characterising MLR in cancer tissues and displays an accumulation of lipids arising during tumourigenesis and an increase in the unsaturated lipid and triglyceride peaks with respect to saturated MLR. Copyright © 2009 John Wiley & Sons, Ltd.     

Generative adversarial network‐based super‐resolution of diffusion‐weighted imaging: Application to tumour radiomics in breast cancer

Diffusion‐weighted imaging (DWI) is increasingly used to guide the clinical management of patients with breast tumours. However, accurate tumour characterization with DWI and the corresponding apparent diffusion coefficient (ADC) maps are challenging due to their limited resolution. This study aimed to produce super‐resolution (SR) ADC images and to assess the clinical utility of these SR images by performing a radiomic analysis for predicting the histologic grade and Ki‐67 expression status of breast cancer. To this end, 322 samples of dynamic enhanced magnetic resonance imaging (DCE‐MRI) and the corresponding DWI data were collected. A SR generative adversarial (SRGAN) and an enhanced deep SR (EDSR) network along with the bicubic interpolation were utilized to generate SR‐ADC images from which radiomic features were extracted. The dataset was randomly separated into a development dataset (n = 222) to establish a deep SR model using DCE‐MRI and a validation dataset (n = 100) to improve the resolution of ADC images. This random separation of datasets was performed 10 times, and the results were averaged. The EDSR method was significantly better than the SRGAN and bicubic methods in terms of objective quality criteria. Univariate and multivariate predictive models of radiomic features were established to determine the area under the receiver operating characteristic curve (AUC). Individual features from the tumour SR‐ADC images showed a higher performance with the EDSR and SRGAN methods than with the bicubic method and the original images. Multivariate analysis of the collective radiomics showed that the EDSR‐ and SRGAN‐based SR‐ADC images performed better than the bicubic method and original images in predicting either Ki‐67 expression levels (AUCs of 0.818 and 0.801, respectively) or the tumour grade (AUCs of 0.826 and 0.828, respectively). This work demonstrates that in addition to improving the resolution of ADC images, deep SR networks can also improve tumour image‐based diagnosis in breast cancer.     

3D variable‐density SPARKLING trajectories for high‐resolution T2*‐weighted magnetic resonance imaging

We have recently proposed a new optimization algorithm called SPARKLING (Spreading Projection Algorithm for Rapid K‐space sampLING) to design efficient compressive sampling patterns for magnetic resonance imaging (MRI). This method has a few advantages over conventional non‐Cartesian trajectories such as radial lines or spirals: i) it allows to sample the k‐space along any arbitrary density while the other two are restricted to radial densities and ii) it optimizes the gradient waveforms for a given readout time. Here, we introduce an extension of the SPARKLING method for 3D imaging by considering both stacks‐of‐SPARKLING and fully 3D SPARKLING trajectories. Our method allowed to achieve an isotropic resolution of 600 μm in just 45 seconds for T2? ‐weighted ex vivo brain imaging at 7 Tesla over a field‐of‐view of 200 × 200 × 140 mm³ . Preliminary in vivo human brain data shows that a stack‐of‐SPARKLING is less subject to off‐resonance artifacts than a stack‐of‐spirals.     

High‐resolution diffusion tensor imaging of the human kidneys using a free‐breathing,multi‐slice,targeted field of view approach

Fractional anisotropy (FA) obtained by diffusion tensor imaging (DTI) can be used to image the kidneys without any contrast media. FA of the medulla has been shown to correlate with kidney function. It is expected that higher spatial resolution would improve the depiction of small structures within the kidney. However, the achievement of high spatial resolution in renal DTI remains challenging as a result of respiratory motion and susceptibility to diffusion imaging artefacts. In this study, a targeted field of view (TFOV) method was used to obtain high‐resolution FA maps and colour‐coded diffusion tensor orientations, together with measures of the medullary and cortical FA, in 12 healthy subjects. Subjects were scanned with two implementations (dual and single kidney) of a TFOV DTI method. DTI scans were performed during free breathing with a navigator‐triggered sequence. Results showed high consistency in the greyscale FA, colour‐coded FA and diffusion tensors across subjects and between dual‐ and single‐kidney scans, which have in‐plane voxel sizes of 2 × 2 mm² and 1.2 × 1.2 mm², respectively. The ability to acquire multiple contiguous slices allowed the medulla and cortical FA to be quantified over the entire kidney volume. The mean medulla and cortical FA values were 0.38 ± 0.017 and 0.21 ± 0.019, respectively, for the dual‐kidney scan, and 0.35 ± 0.032 and 0.20 ± 0.014, respectively, for the single‐kidney scan. The mean FA between the medulla and cortex was significantly different (p < 0.001) for both dual‐ and single‐kidney implementations. High‐spatial‐resolution DTI shows promise for improving the characterization and non‐invasive assessment of kidney function. © 2014 The Authors. NMR in Biomedicine published by John Wiley & Sons, Ltd.