Search

With the rise of open data, identifiability of individuals based on 3D renderings obtained from routine structural magnetic resonance imaging (MRI) scans of the head has become a growing privacy concern. To protect subject privacy, several algorithms have been developed to de-identify imaging data using blurring, defacing or refacing. Completely removing facial structures provides the best re-identification protection but can significantly impact post-processing steps, like brain morphometry. As an alternative, refacing methods that replace individual facial structures with generic templates have a lower effect on the geometry and intensity distribution of original scans, and are able to provide more consistent post-processing results by the price of higher re-identification risk and computational complexity. In the current study, we propose a novel method for anonymised face generation for defaced 3D T1-weighted scans based on a 3D conditional generative adversarial network. To evaluate the performance of the proposed de-identification tool, a comparative study was conducted between several existing defacing and refacing tools, with two different segmentation algorithms (FAST and Morphobox). The aim was to evaluate (i) impact on brain morphometry reproducibility, (ii) re-identification risk, (iii) balance between (i) and (ii), and (iv) the processing time. The proposed method takes 9 seconds for face generation and is suitable for recovering consistent post-processing results after defacing., Comment: preprint

Deep learning methods have become the state of the art for undersampled MR reconstruction. Particularly for cases where it is infeasible or impossible for ground truth, fully sampled data to be acquired, self-supervised machine learning methods for reconstruction are becoming increasingly used. However potential issues in the validation of such methods, as well as their generalizability, remain underexplored. In this paper, we investigate important aspects of the validation of self-supervised algorithms for reconstruction of undersampled MR images: quantitative evaluation of prospective reconstructions, potential differences between prospective and retrospective reconstructions, suitability of commonly used quantitative metrics, and generalizability. Two self-supervised algorithms based on self-supervised denoising and the deep image prior were investigated. These methods are compared to a least squares fitting and a compressed sensing reconstruction using in-vivo and phantom data. Their generalizability was tested with prospectively under-sampled data from experimental conditions different to the training. We show that prospective reconstructions can exhibit significant distortion relative to retrospective reconstructions/ground truth. Furthermore, pixel-wise quantitative metrics may not capture differences in perceptual quality accurately, in contrast to a perceptual metric. In addition, all methods showed potential for generalization; however, generalizability is more affected by changes in anatomy/contrast than other changes. We further showed that no-reference image metrics correspond well with human rating of image quality for studying generalizability. Finally, we showed that a well-tuned compressed sensing reconstruction and learned denoising perform similarly on all data., Comment: Accepted for publication at the Journal of Machine Learning for Biomedical Imaging (MELBA) https://www.melba-journal.org/papers/2022:022.html

The current multiple sclerosis (MS) diagnostic criteria lack specificity, and this may lead to misdiagnosis, which remains an issue in present-day clinical practice. In addition, conventional biomarkers only moderately correlate with MS disease progression. Recently, advanced MS lesional imaging biomarkers such as cortical lesions (CL), the central vein sign (CVS), and paramagnetic rim lesions (PRL), visible in specialized magnetic resonance imaging (MRI) sequences, have shown higher specificity in differential diagnosis. Moreover, studies have shown that CL and PRL are potential prognostic biomarkers, the former correlating with cognitive impairments and the latter with early disability progression. As machine learning-based methods have achieved extraordinary performance in the assessment of conventional imaging biomarkers, such as white matter lesion segmentation, several automated or semi-automated methods have been proposed for CL, CVS, and PRL as well. In the present review, we first introduce these advanced MS imaging biomarkers and their imaging methods. Subsequently, we describe the corresponding machine learning-based methods that were used to tackle these clinical questions, putting them into context with respect to the challenges they are still facing, including non-standardized MRI protocols, limited datasets, and moderate inter-rater variability. We conclude by presenting the current limitations that prevent their broader deployment and suggesting future research directions.

Accurate characterization of in utero human brain maturation is critical as it involves complex and interconnected structural and functional processes that may influence health later in life. Magnetic resonance imaging is a powerful tool to investigate equivocal neurological patterns during fetal development. However, the number of acquisitions of satisfactory quality available in this cohort of sensitive subjects remains scarce, thus hindering the validation of advanced image processing techniques. Numerical phantoms can mitigate these limitations by providing a controlled environment with a known ground truth. In this work, we present FaBiAN, an open-source Fetal Brain magnetic resonance Acquisition Numerical phantom that simulates clinical T2-weighted fast spin echo sequences of the fetal brain. This unique tool is based on a general, flexible and realistic setup that includes stochastic fetal movements, thus providing images of the fetal brain throughout maturation comparable to clinical acquisitions. We demonstrate its value to evaluate the robustness and optimize the accuracy of an algorithm for super-resolution fetal brain magnetic resonance imaging from simulated motion-corrupted 2D low-resolution series as compared to a synthetic high-resolution reference volume. We also show that the images generated can complement clinical datasets to support data-intensive deep learning methods for fetal brain tissue segmentation., Comment: 23 pages, 9 figures (including Supplementary Material), 4 tables, 1 supplement. Submitted to Scientific Reports (2021)

Multiple sclerosis (MS) is an inflammatory and neurodegenerative disease characterized by diffuse and focal areas of tissue loss. Conventional MRI techniques such as T1-weighted and T2-weighted scans are generally used in the diagnosis and prognosis of the disease. Yet, these methods are limited by the lack of specificity between lesions, their perilesional area and non-lesional tissue. Alternative MRI techniques exhibit a higher level of sensitivity to focal and diffuse MS pathology than conventional MRI acquisitions. However, they still suffer from limited specificity when considered alone. In this work, we have combined tissue microstructure information derived from multicompartment diffusion MRI and T2 relaxometry models to explore the voxel-based prediction power of a machine learning model in a cohort of MS patients and healthy controls. Our results show that the combination of multi-modal features, together with a boosting enhanced decision-tree based classifier, which combines a set of weak classifiers to form a strong classifier via a voting mechanism, is able to utilise the complementary information for the classification of abnormal tissue., Comment: Accepted at ISBI2021

Relaxometry studies in preterm and at-term newborns have provided insight into brain microstructure, thus opening new avenues for studying normal brain development and supporting diagnosis in equivocal neurological situations. However, such quantitative techniques require long acquisition times and therefore cannot be straightforwardly translated to in utero brain developmental studies. In clinical fetal brain magnetic resonance imaging routine, 2D low-resolution T2-weighted fast spin echo sequences are used to minimize the effects of unpredictable fetal motion during acquisition. As super-resolution techniques make it possible to reconstruct a 3D high-resolution volume of the fetal brain from clinical low-resolution images, their combination with quantitative acquisition schemes could provide fast and accurate T2 measurements. In this context, the present work demonstrates the feasibility of using super-resolution reconstruction from conventional T2-weighted fast spin echo sequences for 3D isotropic T2 mapping. A quantitative magnetic resonance phantom was imaged using a clinical T2-weighted fast spin echo sequence at variable echo time to allow for super-resolution reconstruction at every echo time and subsequent T2 mapping of samples whose relaxometric properties are close to those of fetal brain tissue. We demonstrate that this approach is highly repeatable, accurate and robust when using six echo times (total acquisition time under 9 minutes) as compared to gold-standard single-echo spin echo sequences (several hours for one single 2D slice)., Comment: 11 pages, 4 figures, 1 table, 1 supplement, to appear in Proceedings in Medical Image Computing and Computer Assisted Intervention (MICCAI), Peru, October 2020

Recovering the T2 distribution from multi-echo T2 magnetic resonance (MR) signals is challenging but has high potential as it provides biomarkers characterizing the tissue micro-structure, such as the myelin water fraction (MWF). In this work, we propose to combine machine learning and aspects of parametric (fitting from the MRI signal using biophysical models) and non-parametric (model-free fitting of the T2 distribution from the signal) approaches to T2 relaxometry in brain tissue by using a multi-layer perceptron (MLP) for the distribution reconstruction. For training our network, we construct an extensive synthetic dataset derived from biophysical models in order to constrain the outputs with \textit{a priori} knowledge of \textit{in vivo} distributions. The proposed approach, called Model-Informed Machine Learning (MIML), takes as input the MR signal and directly outputs the associated T2 distribution. We evaluate MIML in comparison to non-parametric and parametric approaches on synthetic data, an ex vivo scan, and high-resolution scans of healthy subjects and a subject with Multiple Sclerosis. In synthetic data, MIML provides more accurate and noise-robust distributions. In real data, MWF maps derived from MIML exhibit the greatest conformity to anatomical scans, have the highest correlation to a histological map of myelin volume, and the best unambiguous lesion visualization and localization, with superior contrast between lesions and normal appearing tissue. In whole-brain analysis, MIML is 22 to 4980 times faster than non-parametric and parametric methods, respectively., Comment: Preprint submitted to Medical Image Analysis (July 14, 2020)

Purpose: To develop SPARCQ (Signal Profile Asymmetries for Rapid Compartment Quantification), a novel approach to quantify fat fraction using asymmetries in the phase-cycled bSSFP profile. Methods: SPARCQ uses phase-cycling to obtain bSSFP frequency profiles, which display asymmetries in the presence of fat and water at certain repetition times. For each voxel, the measured signal profile is decomposed into a weighted sum of simulated profiles via multi-compartment dictionary matching. Each dictionary entry represents a single-compartment bSSFP profile with a specific off-resonance frequency and relaxation time ratio. Using the results of dictionary matching, the fractions of the different off-resonance components are extracted for each voxel, generating quantitative maps of water and fat fraction and banding-artifact-free images for the entire image volume. SPARCQ was validated using simulations, experiments in a water-fat phantom and in knees of healthy volunteers. Experimental results were compared with reference proton density fat fractions obtained with 1H-MRS (phantoms) and with multiecho gradient-echo MRI (phantoms and volunteers). SPARCQ repeatability was evaluated in six scan-rescan experiments. Results: Simulations showed that fat fraction quantification is accurate and robust for signal-to-noise ratios greater than 20. Phantom experiments demonstrated good agreement between SPARCQ and gold standard fat fractions. In volunteers, banding-artifact-free quantitative maps and water-fat-separated images obtained with SPARCQ and ME-GRE demonstrated the expected contrast between fatty and non-fatty tissues. The coefficient of repeatability of SPARCQ fat fraction was 0.0512. Conclusion: SPARCQ demonstrates potential for fat quantification using asymmetries in bSSFP profiles, and may be a promising alternative to conventional fat fraction quantification techniques., Comment: 30 pages, 8 figures, submitted to Magnetic Resonance in Medicine

Fluorine-19 (19F) MRI of injected perfluorocarbon emulsions (PFCs) allows for the non-invasive quantification of inflammation and cell tracking, but suffers from a low signal-to-noise ratio and extended scan time. To address this limitation, we tested the hypothesis that a 19F MRI pulse sequence that combines a specific undersampling regime with signal averaging has increased sensitivity and robustness against motion artifacts compared to a non-averaged fully-sampled dataset, when both are reconstructed with compressed sensing. To this end, numerical simulations and phantom experiments were performed to characterize the point spread function (PSF) of undersampling patterns and the vulnerability to noise of acquisition-reconstruction strategies with paired numbers of x signal averages and acceleration factor x (NAx-AFx). At all investigated noise levels, the DSC of the acquisition-reconstruction strategies strongly depended on the regularization parameters and acceleration factor. In phantoms, motion robustness of an NA8-AF8 undersampling pattern versus NA1-AF1 was evaluated with simulated and real motions. Differences were assessed with Dice similarity coefficients (DSC), and were consistently higher for NA8-AF8 compared to NA1-AF1 strategy, for both simulated and real cyclic motions (P<0.001). Both acquisition-reconstruction strategies were validated in vivo in mice (n=2) injected with perfluoropolyether. These images displayed a sharper delineation of the liver with the NA8-AF8 strategy than with the NA1-AF1 strategy. In conclusion, we validated the hypothesis that in 19F MRI, the combination of undersampling and averaging improves both the sensitivity and the robustness against motion artifacts compared to a non-averaged fully-sampled dataset, when both are reconstructed with compressed sensing.

Purpose: To study the effects of magnetization transfer (MT, in which a semisolid spin pool interacts with the free pool), in the context of magnetic resonance fingerprinting (MRF). Methods: Simulations and phantom experiments were performed to study the impact of MT on the MRF signal and its potential influence on T1 and T2 estimation. Subsequently, an MRF sequence implementing off-resonance MT pulses and a dictionary with an MT dimension by incorporating a two-pool model were used to estimate the fractional pool size in addition to the B1+, T1, and T2 values. The proposed method was evaluated in the human brain. Results: Simulations and phantom experiments showed that an MRF signal obtained from a cross-linked bovine serum sample is influenced by MT. Using a dictionary based on an MT model, a better match between simulations and acquired MR signals can be obtained (NRMSE 1.3% versus 4.7%). Adding off-resonance MT pulses can improve the differentiation of MT from T1 and T2. In-vivo results showed that MT affects the MRF signals from white matter (fractional pool-size ~16%) and gray matter (fractional pool-size ~10%). Furthermore, longer T1 (~1060 ms versus ~860 ms) and T2 values (~47 ms versus ~35 ms) can be observed in white matter if MT is accounted for. Conclusion: Our experiments demonstrated a potential influence of MT on the quantification of T1 and T2 with MRF. A model that encompasses MT effects can improve the accuracy of estimated relaxation parameters and allows quantification of the fractional pool size., Comment: 11 pages, 5 figures, 2 tables, 2 supplements

In this work, we present a comparison of a shallow and a deep learning architecture for the automated segmentation of white matter lesions in MR images of multiple sclerosis patients. In particular, we train and test both methods on early stage disease patients, to verify their performance in challenging conditions, more similar to a clinical setting than what is typically provided in multiple sclerosis segmentation challenges. Furthermore, we evaluate a prototype naive combination of the two methods, which refines the final segmentation. All methods were trained on 32 patients, and the evaluation was performed on a pure test set of 73 cases. Results show low lesion-wise false positives (30%) for the deep learning architecture, whereas the shallow architecture yields the best Dice coefficient (63%) and volume difference (19%). Combining both shallow and deep architectures further improves the lesion-wise metrics (69% and 26% lesion-wise true and false positive rate, respectively)., Comment: Accepted to the MICCAI 2018 Brain Lesion (BrainLes) workshop