Liver proton density fat fraction (PDFF), the ratio between fat-only and overall proton densities, is an extensively validated biomarker associated with several diseases. In recent years, numerous deep learning-based methods for estimating PDFF have been proposed to optimize acquisition and post-processing times without sacrificing accuracy, compared to conventional methods. However, the lack of interpretability and the often poor generalizability of these DL-based models undermine the adoption of such techniques in clinical practice. In this work, we propose an Artificial Intelligence-based Decomposition of water and fat with Echo Asymmetry and Least-squares (AI-DEAL) method, designed to estimate both proton density fat fraction (PDFF) and the associated uncertainty maps. Once trained, AI-DEAL performs a one-shot MRI water-fat separation by first calculating the nonlinear confounder variables, R 2 ∗ and off-resonance field. It then employs a weighted least squares approach to compute water-only and fat-only signals, along with their corresponding covariance matrix, which are subsequently used to derive the PDFF and its associated uncertainty. We validated our method using in vivo liver CSE-MRI, a fat-water phantom, and a numerical phantom. AI-DEAL demonstrated PDFF biases of 0.25% and -0.12% at two liver ROIs, outperforming state-of-the-art deep learning-based techniques. Although trained using in vivo data, our method exhibited PDFF biases of -3.43% in the fat-water phantom and -0.22% in the numerical phantom with no added noise. The latter bias remained approximately constant when noise was introduced. Furthermore, the estimated uncertainties showed good agreement with the observed errors and the variations within each ROI, highlighting their potential value for assessing the reliability of the resulting PDFF maps. • MRI water-fat separation requires long scan and post-processing times. • Deep learning approaches enable time optimization but have poor interpretability and generalisability. • AI-DEAL allows simultaneous water-fat separation and uncertainty quantification even from fat-water phantoms, despite being trained on liver MRI. • Our method demonstrated robustness to noise, and uncertainty estimates showed sensitivity to commonly observed artifacts. • The errors of the measured fat fractions showed good agreement with uncertainty estimates and noise level.

Even after early repair of aortic coarctation (AoCo), life expectancy is reduced due to complications such as hypertension. Invasive diagnostic catheterization is used to evaluate peak-to-peak pressure gradients (PGpp) across the CoAo. Clinically significant PGppare those greater than 20 mmHg under resting conditions, in which case the patient is referred for a second intervention to repair the CoAo. In this study, we demonstrate the feasibility of using Physics-Informed Neural Networks (PINNs) to predict PGppin patients with AoCo non-invasively, based on images obtained from cardiac magnetic resonance imaging. We analyzed a group of 3 patients with CoAo under resting and pharmacological stress conditions. We were able to obtain PGppvalues very close to the actual values obtained by diagnostic catheterization, with an absolute error and average percentage error of 0.57 mmHg and 8.29% for the resting condition, and 4.13 mmHg and 8.63% for the pharmacological stress condition. Our method also successfully identified the only patient who presented a clinically significant PGppunder resting conditions, with differences of less than 1 mmHg.

Medical image synthesis represents a critical area of research in clinical decision-making, aiming to overcome the challenges associated with acquiring multiple image modalities for an accurate clinical workflow. This approach proves beneficial in estimating an image of a desired modality from a given source modality among the most common medical imaging contrasts, such as Computed Tomography (CT), Magnetic Resonance Imaging (MRI), and Positron Emission Tomography (PET). However, translating between two image modalities presents difficulties due to the complex and non-linear domain mappings. Deep learning-based generative modelling has exhibited superior performance in synthetic image contrast applications compared to conventional image synthesis methods. This survey comprehensively reviews deep learning-based medical imaging translation from 2018 to 2023 on pseudo-CT, synthetic MR, and synthetic PET. We provide an overview of synthetic contrasts in medical imaging and the most frequently employed deep learning networks for medical image synthesis. Additionally, we conduct a detailed analysis of each synthesis method, focusing on their diverse model designs based on input domains and network architectures. We also analyse novel network architectures, ranging from conventional CNNs to the recent Transformer and Diffusion models. This analysis includes comparing loss functions, available datasets and anatomical regions, and image quality assessments and performance in other downstream tasks. Finally, we discuss the challenges and identify solutions within the literature, suggesting possible future directions. We hope that the insights offered in this survey paper will serve as a valuable roadmap for researchers in the field of medical image synthesis.

The diagnosis of heart failure usually includes a global functional assessment, such as ejection fraction measured by magnetic resonance imaging. However, these metrics have low discriminate power to distinguish different cardiomyopathies, which may not affect the global function of the heart. Quantifying local deformations in the form of cardiac strain can provide helpful information, but it remains a challenge. In this work, we introduce WarpPINN, a physics-informed neural network to perform image registration to obtain local metrics of heart deformation. We apply this method to cine magnetic resonance images to estimate the motion during the cardiac cycle. We inform our neural network of the near-incompressibility of cardiac tissue by penalizing the Jacobian of the deformation field. The loss function has two components: an intensity-based similarity term between the reference and the warped template images, and a regularizer that represents the hyperelastic behavior of the tissue. The architecture of the neural network allows us to easily compute the strain via automatic differentiation to assess cardiac activity. We use Fourier feature mappings to overcome the spectral bias of neural networks, allowing us to capture discontinuities in the strain field. The algorithm is tested on synthetic examples and on a cine SSFP MRI benchmark of 15 healthy volunteers, where it is trained to learn the deformation mapping of each case. We outperform current methodologies in landmark tracking and provide physiological strain estimations in the radial and circumferential directions. WarpPINN provides precise measurements of local cardiac deformations that can be used for a better diagnosis of heart failure and can be used for general image registration tasks. Source code is available at https://github.com/fsahli/WarpPINN.

While MRI allows to encode the motion of tissue in the magnetization's phase, it remains yet a challenge to obtain high fidelity motion images due to wraps in the phase for high encoding efficiencies. Therefore, we propose an optimal multiple motion encoding method (OMME) and exemplify it in Magnetic Resonance Elastography (MRE) data. OMME is formulated as a non-convex least-squares problem for the motion using an arbitrary number of phase-contrast measurements with different motion encoding gradients (MEGs). The mathematical properties of OMME are proved in terms of standard deviation and dynamic range of the motion's estimate for arbitrary MEGs combination which are confirmed using synthetically generated data. OMME's performance is assessed on MRE data from in vivo human brain experiments and compared to dual encoding strategies. The unwrapped images are further used to reconstruct stiffness maps and compared to the ones obtained using conventional unwrapping methods. OMME allowed to successfully combine several MRE phase images with different MEGs, outperforming dual encoding strategies in either motion-to-noise ratio (MNR) or number of successfully reconstructed voxels with good noise stability. This lead to stiffness maps with greater resolution of details than obtained with conventional unwrapping methods. The proposed OMME method allows for a flexible and noise robust increase in the dynamic range and thus provides wrap-free phase images with high MNR. In MRE, the method may be especially suitable when high resolution images with high MNR are needed.

The susceptibility of super paramagnetic iron oxide (SPIO) particles makes them a useful contrast agent for different purposes in MRI. These particles are typically quantified with relaxometry or by measuring the inhomogeneities they produced. These methods rely on the phase, which is unreliable for high concentrations. We present in this study a novel Deep Learning method to quantify the SPIO concentration distribution. We acquired the data with a new sequence called View Line in which the field map information is encoded in the geometry of the image. The novelty of our network is that it uses residual blocks as the bottleneck and multiple decoders to improve the gradient flow in the network. Each decoder predicts a different part of the wavelet decomposition of the concentration map. This decomposition improves the estimation of the concentration, and also it accelerates the convergence of the model. We tested our SPIO concentration reconstruction technique with simulated images and data from actual scans from phantoms. The simulations were done using images from the IXI dataset, and the phantoms consisted of plastic cylinders containing agar with SPIO particles at different concentrations. In both experiments, the model was able to quantify the distribution accurately.

While the clinical gold standard for pressure difference measurements is invasive catheterization, 4D Flow MRI is a promising tool for enabling a non-invasive quantification, by linking highly spatially resolved velocity measurements with pressure differences via the incompressible Navier-Stokes equations. In this work we provide a validation and comparison with phantom and clinical patient data of pressure difference maps estimators. We compare the classical Pressure Poisson Estimator (PPE) and the new Stokes Estimator (STE) against catheter pressure measurements under a variety of stenosis severities and flow intensities. Specifically, we use several 4D Flow data sets of realistic aortic phantoms with different anatomic and hemodynamic severities and two patients with aortic coarctation. The phantom data sets are enriched by subsampling to lower resolutions, modification of the segmentation and addition of synthetic noise, in order to study the sensitivity of the pressure difference estimators to these factors. Overall, the STE method yields more accurate results than the PPE method compared to catheterization data. The superiority of the STE becomes more evident at increasing Reynolds numbers with a better capacity of capturing pressure gradients in strongly convective flow regimes. The results indicate an improved robustness of the STE method with respect to variation in lumen segmentation. However, with heuristic removal of the wall-voxels, the PPE can reach a comparable accuracy for lower Reynolds' numbers.

In MR Fingerprinting (MRF), balanced Steady-State Free Precession (bSSFP) has advantages over unbalanced SSFP because it retains the spin history achieving a higher signal-to-noise ratio (SNR) and scan efficiency. However, bSSFP-MRF is not frequently used because it is sensitive to off-resonance, producing artifacts and blurring, and affecting the parametric map quality. Here we propose a novel Spatial Off-resonance Correction (SOC) approach for reducing these artifacts in bSSFP-MRF with spiral trajectories. SOC-MRF uses each pixel's Point Spread Function to create system matrices that encode both off-resonance and gridding effects. We iteratively compute the inverse of these matrices to reduce the artifacts. We evaluated the proposed method using brain simulations and actual MRF acquisitions of a standardized T1/T2 phantom and five healthy subjects. The results show that the off-resonance distortions in T1/T2 maps were considerably reduced using SOC-MRF. For T2, the Normalized Root Mean Square Error (NRMSE) was reduced from 17.3 to 8.3% (simulations) and from 35.1 to 14.9% (phantom). For T1, the NRMS was reduced from 14.7 to 7.7% (simulations) and from 17.7 to 6.7% (phantom). For in-vivo, the mean and standard deviation in different ROI in white and gray matter were significantly improved. For example, SOC-MRF estimated an average T2 for white matter of 77ms (the ground truth was 74ms) versus 50 ms of MRF. For the same example the standard deviation was reduced from 18 ms to 6ms. The corrections achieved with the proposed SOC-MRF may expand the potential applications of bSSFP-MRF, taking advantage of its better SNR property.

We proposed a novel method called HARP-I, which enhances the estimation of motion from tagged Magnetic Resonance Imaging (MRI). The harmonic phase of the images is unwrapped and treated as noisy measurements of reference coordinates on a deformed domain, obtaining motion with high accuracy using Radial Basis Functions interpolations. Results were compared against Shortest Path HARP Refinement (SP-HR) and Sine-wave Modeling (SinMod), two harmonic image-based techniques for motion estimation from tagged images. HARP-I showed a favorable similarity with both methods under noise-free conditions, whereas a more robust performance was found in the presence of noise. Cardiac strain was better estimated using HARP-I at almost any motion level, giving strain maps with less artifacts. Additionally, HARP-I showed better temporal consistency as a new method was developed to fix phase jumps between frames. In conclusion, HARP-I showed to be a robust method for the estimation of motion and strain under ideal and non-ideal conditions.

Dual-VENC strategies have been proposed to improve the velocity-to-noise ratio in phase-contrast MRI. However, they are based on aliasing-free high-VENC data. The aim of this paper is to propose a dual-VENC velocity estimation method allowing high-VENC aliased data. For this purpose, we reformulate the phase-contrast velocity as a least squares estimator, providing a natural framework for including multiple encoding gradient measurements. By analyzing the mathematical properties of both single- and dual-VENC problems, we can justify theoretically high/low-VENC ratios such that the aliasing velocity can be minimized. The resulting reconstruction algorithm was assessed using three types of data: numerical, experimental, and volunteers. In clinical practice, this method would allow shorter examination times by avoiding tedious adaptation of VENC values by repeated scans.

Velocity measurements from 4D flow MRI are prone to be affected by several imperfections of the MR system. Assuming that blood is incompressible, we propose a novel method for enhancing the velocity field by reducing its divergence. To enhance the velocity data, we added a corrector velocity to each voxel such that the divergence is minimized. The method was validated using an analytical Womersley flow model for different settings of resolution and noise levels. The performance of the proposed method was also assessed in volunteers and patients. Results demonstrated a significant reduction of the divergence depending on the size of the regularization term, obtaining a reduction close to 50% of the mean divergence with negligible modification of flow parameters. Remarkably, we found that the reduction of the divergence, in percentage, was independent of volunteers, resolution or noise.

Several 2D methods have been proposed to estimate WSS and OSI from PC-MRI, neglecting the longitudinal velocity gradients that typically arise in cardiovascular flow, particularly on vessel geometries whose cross section and centerline orientation strongly vary in the axial direction. Thus, the contribution of longitudinal velocity gradients remains understudied. In this work, we propose a 3D finite-element method for the quantification of WSS and OSI from 3D-CINE PC-MRI that accounts for both in-plane and longitudinal velocity gradients. We demonstrate the convergence and robustness of the method on cylindrical geometries using a synthetic phantom based on the Poiseuille flow equation. We also show that, in the presence of noise, the method is both stable and accurate. Using computational fluid dynamics simulations, we show that the proposed 3D method results in more accurate WSS estimates than those obtained from a 2D analysis not considering out-of-plane velocity gradients. Further, we conclude that for irregular geometries the accurate prediction of WSS requires the consideration of longitudinal gradients in the velocity field. Additionally, we compute 3D maps of WSS and OSI for 3D-CINE PC-MRI data sets from an aortic phantom and sixteen healthy volunteers and two patients. The OSI values show a greater dispersion than WSS, which is strongly dependent on the PC-MRI resolution. We envision that the proposed 3D method will improve the estimation of WSS and OSI from 3D-CINE PC-MRI images, allowing for more accurate estimates in vessels with pathologies that induce high longitudinal velocity gradients, such as coarctations and aneurisms.

BACKGROUND: Calcium (Ca2+) propagates within tissues serving as an important information carrier. In particular, cilia beat frequency in oviduct cells is partially regulated by Ca2+ changes. Thus, measuring the calcium density and characterizing the traveling wave plays a key role in understanding biological phenomena. However, current methods to measure propagation velocities and other wave characteristics involve several manual or time-consuming procedures. This limits the amount of information that can be extracted, and the statistical quality of the analysis. RESULTS: Our work provides a framework based on image processing procedures that enables a fast, automatic and robust characterization of data from two-filter fluorescence Ca2+ experiments. We calculate the mean velocity of the wave-front, and use theoretical models to extract meaningful parameters like wave amplitude, decay rate and time of excitation. CONCLUSIONS: Measurements done by different operators showed a high degree of reproducibility. This framework is also extended to a single filter fluorescence experiments, allowing higher sampling rates, and thus an increased accuracy in velocity measurements.

Long acquisition times are still a limitation for many applications of magnetic resonance imaging (MRI), specially in 3-D and dynamic imaging. Several undersampling reconstruction techniques have been proposed to overcome this problem. These techniques are based on acquiring less samples than specified by the Nyquist criterion and estimating the nonacquired data by using some sort of prior information. Most of these reconstruction methods use prior information based on estimations of the pixel intensities of the images and therefore they are prone to introduce spatial or temporal blurring. Instead of using the pixel intensities, we propose to use information that allows us to sort the pixels of an image from darkest to brightest. The set of order relations which sort the pixels of an image has been called intensity order. The intensity order of an image can be estimated from low-resolution images, adjacent slices in volumetric acquisitions, temporal correlation in dynamic sequences or from prior reconstructions. Our technique for reconstruction using intensity order (TRIO) consists of looking for an image that satisfies the intensity order and minimizes the discrepancy between the acquired and reconstructed data. Results show that TRIO can effectively reconstruct 2-D-cine cardiac MR images (under-sampling factor of 4), estimating correctly the temporal evolution of the objects. Furthermore, TRIO is used as a second stage reconstruction after reconstructing with other techniques, keyhole, sliding window and k-t BLAST, to estimate the order information. In all cases the images are improved by TRIO.