Deep learning microstructure estimation of developing brains from diffusion MRI: A newborn and fetal study

Diffusion-weighted magnetic resonance imaging (dMRI) is widely used to assess the brain white matter. Fiber orientation distribution functions (FODs) are a common way of representing the orientation and density of white matter fibers. However, with standard FOD computation methods, accurate estimation requires a large number of measurements that usually cannot be acquired for newborns and fetuses. We propose to overcome this limitation by using a deep learning method to map as few as six diffusion-weighted measurements to the target FOD. To train the model, we use the FODs computed using multi-shell high angular resolution measurements as target. Extensive quantitative evaluations show that the new deep learning method, using significantly fewer measurements, achieves comparable or superior results than standard methods such as Constrained Spherical Deconvolution and two state-of-the-art deep learning methods. For voxels with one and two fibers, respectively, our method shows an agreement rate in terms of the number of fibers of 77.5% and 22.2%, which is 3% and 5.4% higher than other deep learning methods, and an angular error of 10° and 20°, which is 6° and 5° lower than other deep learning methods. To determine baselines for assessing the performance of our method, we compute agreement metrics using densely sampled newborn data. Moreover, we demonstrate the generalizability of the new deep learning method across scanners, acquisition protocols, and anatomy on two clinical external datasets of newborns and fetuses. We validate fetal FODs, successfully estimated for the first time with deep learning, using post-mortem histological data. Our results show the advantage of deep learning in computing the fiber orientation density for the developing brain from in-vivo dMRI measurements that are often very limited due to constrained acquisition times. Our findings also highlight the intrinsic limitations of dMRI for probing the developing brain microstructure.

Robust thalamic nuclei segmentation from T1-weighted MRI using polynomial intensity transformation

Accurate segmentation of thalamic nuclei, crucial for understanding their role in healthy cognition and in pathologies, is challenging to achieve on standard T1-weighted (T1w) magnetic resonance imaging (MRI) due to poor image contrast. White-matter-nulled (WMn) MRI sequences improve intrathalamic contrast but are not part of clinical protocols or extant databases. In this study, we introduce histogram-based polynomial synthesis (HIPS), a fast preprocessing transform step that synthesizes WMn-like image contrast from standard T1w MRI using a polynomial approximation for intensity transformation. HIPS was incorporated into THalamus Optimized Multi-Atlas Segmentation (THOMAS) pipeline, a method developed and optimized for WMn MRI. HIPS-THOMAS was compared to a convolutional neural network (CNN)-based segmentation method and THOMAS modified for the use of T1w images (T1w-THOMAS). The robustness and accuracy of the three methods were tested across different image contrasts (MPRAGE, SPGR, and MP2RAGE), scanner manufacturers (PHILIPS, GE, and Siemens), and field strengths (3 T and 7 T). HIPS-transformed images improved intra-thalamic contrast and thalamic boundaries, and HIPS-THOMAS yielded significantly higher mean Dice coefficients and reduced volume errors compared to both the CNN method and T1w-THOMAS. Finally, all three methods were compared using the frequently travelling human phantom MRI dataset for inter- and intra-scanner variability, with HIPS displaying the least inter-scanner variability and performing comparably with T1w-THOMAS for intra-scanner variability. In conclusion, our findings highlight the efficacy and robustness of HIPS in enhancing thalamic nuclei segmentation from standard T1w MRI.

Assessment of fetal corpus callosum biometry by 3D super-resolution reconstructed T2-weighted magnetic resonance imaging

Objective

To assess the accuracy of corpus callosum (CC) biometry, including sub-segments, using 3D super-resolution fetal brain MRI (SR) compared to 2D or 3D ultrasound (US) and clinical low-resolution T2-weighted MRI (T2WS).

Method

Fetal brain biometry was conducted by two observers on 57 subjects [21-35 weeks of gestational age (GA)], including 11 cases of partial CC agenesis. Measures were performed by a junior observer (obs1) on US, T2WS and SR and by a senior neuroradiologist (obs2) on T2WS and SR. CC biometric regression with GA was established. Statistical analysis assessed agreement within and between modalities and observers.

Results

This study shows robust SR to US concordance across gestation, surpassing T2WS. In obs1, SR aligns with US, except for genu and CC length (CCL), enhancing splenium visibility. In obs2, SR closely corresponds to US, differing in rostrum and CCL. The anterior CC (rostrum and genu) exhibits higher variability. SR's regression aligns better with literature (US) for CCL, splenium and body than T2WS. SR is the method with the least missing values.

Conclusion

SR yields CC biometry akin to US (excluding anterior CC). Thanks to superior 3D visualization and better through plane spatial resolution, SR allows to perform CC biometry more frequently than T2WS.

Robust thalamic nuclei segmentation from T1-weighted MRI using polynomial intensity transformation

Accurate segmentation of thalamic nuclei, crucial for understanding their role in healthy cognition and in pathologies, is challenging to achieve on standard T1-weighted (T1w) magnetic resonance imaging (MRI) due to poor image contrast. White-matter-nulled (WMn) MRI sequences improve intrathalamic contrast but are not part of clinical protocols or extant databases. In this study, we introduce histogram-based polynomial synthesis (HIPS), a fast preprocessing transform step that synthesizes WMn-like image contrast from standard T1w MRI using a polynomial approximation for intensity transformation. HIPS was incorporated into THalamus Optimized Multi-Atlas Segmentation (THOMAS) pipeline, a method developed and optimized for WMn MRI. HIPS-THOMAS was compared to a convolutional neural network (CNN)-based segmentation method and THOMAS modified for T1w images (T1w-THOMAS). The robustness and accuracy of the three methods were tested across different image contrasts (MPRAGE, SPGR, and MP2RAGE), scanner manufacturers (PHILIPS, GE, and Siemens), and field strengths (3T and 7T). HIPS-transformed images improved intra-thalamic contrast and thalamic boundaries, and HIPS-THOMAS yielded significantly higher mean Dice coefficients and reduced volume errors compared to both the CNN method and T1w-THOMAS. Finally, all three methods were compared using the frequently travelling human phantom MRI dataset for inter- and intra-scanner variability, with HIPS displaying the least inter-scanner variability and performing comparably with T1w-THOMAS for intra-scanner variability. In conclusion, our findings highlight the efficacy and robustness of HIPS in enhancing thalamic nuclei segmentation from standard T1w MRI.

CROSS-AGE AND CROSS-SITE DOMAIN SHIFT IMPACTS ON DEEP LEARNING-BASED WHITE MATTER FIBER ESTIMATION IN NEWBORN AND BABY BRAINS

Deep learning models have shown great promise in estimating tissue microstructure from limited diffusion magnetic resonance imaging data. However, these models face domain shift challenges when test and train data are from different scanners and protocols, or when the models are applied to data with inherent variations such as the developing brains of infants and children scanned at various ages. Several techniques have been proposed to address some of these challenges, such as data harmonization or domain adaptation in the adult brain. However, those techniques remain unexplored for the estimation of fiber orientation distribution functions in the rapidly developing brains of infants. In this work, we extensively investigate the age effect and domain shift within and across two different cohorts of 201 newborns and 165 babies using the Method of Moments and fine-tuning strategies. Our results show that reduced variations in the microstructural development of babies in comparison to newborns directly impact the deep learning models' cross-age performance. We also demonstrate that a small number of target domain samples can significantly mitigate domain shift problems.

A Multicenter Longitudinal MRI Study Assessing LeMan-PV Software Accuracy in the Detection of White Matter Lesions in Multiple Sclerosis Patients

BACKGROUND

Detecting new and enlarged lesions in multiple sclerosis (MS) patients is needed to determine their disease activity. LeMan-PV is a software embedded in the scanner reconstruction system of one vendor, which automatically assesses new and enlarged white matter lesions (NELs) in the follow-up of MS patients; however, multicenter validation studies are lacking.

PURPOSE

To assess the accuracy of LeMan-PV for the longitudinal detection NEL white-matter MS lesions in a multicenter clinical setting.

STUDY TYPE

Retrospective, longitudinal.

SUBJECTS

A total of 206 patients with a definitive MS diagnosis and at least two follow-up MRI studies from five centers participating in the Swiss Multiple Sclerosis Cohort study. Mean age at first follow-up = 45.2 years (range: 36.9-52.8 years); 70 males.

FIELD STRENGTH/SEQUENCE

Fluid attenuated inversion recovery (FLAIR) and T1-weighted magnetization prepared rapid gradient echo (T1-MPRAGE) sequences at 1.5 T and 3 T.

ASSESSMENT

The study included 313 MRI pairs of datasets. Data were analyzed with LeMan-PV and compared with a manual "reference standard" provided by a neuroradiologist. A second rater (neurologist) performed the same analysis in a subset of MRI pairs to evaluate the rating-accuracy. The Sensitivity (Se), Specificity (Sp), Accuracy (Acc), F1-score, lesion-wise False-Positive-Rate (aFPR), and other measures were used to assess LeMan-PV performance for the detection of NEL at 1.5 T and 3 T. The performance was also evaluated in the subgroup of 123 MRI pairs at 3 T.

STATISTICAL TESTS

Intraclass correlation coefficient (ICC) and Cohen's kappa (CK) were used to evaluate the agreement between readers.

RESULTS

The interreader agreement was high for detecting new lesions (ICC = 0.97, Pvalue < 10-20 , CK = 0.82, P value = 0) and good (ICC = 0.75, P value < 10-12 , CK = 0.68, P value = 0) for detecting enlarged lesions. Across all centers, scanner field strengths (1.5 T, 3 T), and for NEL, LeMan-PV achieved: Acc = 61%, Se = 65%, Sp = 60%, F1-score = 0.44, aFPR = 1.31. When both follow-ups were acquired at 3 T, LeMan-PV accuracy was higher (Acc = 66%, Se = 66%, Sp = 66%, F1-score = 0.28, aFPR = 3.03).

DATA CONCLUSION

In this multicenter study using clinical data settings acquired at 1.5 T and 3 T, and variations in MRI protocols, LeMan-PV showed similar sensitivity in detecting NEL with respect to other recent 3 T multicentric studies based on neural networks. While LeMan-PV performance is not optimal, its main advantage is that it provides automated clinical decision support integrated into the radiological-routine flow.

EVIDENCE LEVEL

4 TECHNICAL EFFICACY: Stage 2.

How far MS lesion detection and segmentation are integrated into the clinical workflow? A systematic review

INTRODUCTION

Over the past few years, the deep learning community has developed and validated a plethora of tools for lesion detection and segmentation in Multiple Sclerosis (MS). However, there is an important gap between validating models technically and clinically. To this end, a six-step framework necessary for the development, validation, and integration of quantitative tools in the clinic was recently proposed under the name of the Quantitative Neuroradiology Initiative (QNI).

AIMS

Investigate to what extent automatic tools in MS fulfill the QNI framework necessary to integrate automated detection and segmentation into the clinical neuroradiology workflow.

METHODS

Adopting the systematic Cochrane literature review methodology, we screened and summarised published scientific articles that perform automatic MS lesions detection and segmentation. We categorised the retrieved studies based on their degree of fulfillment of QNI's six-steps, which include a tool's technical assessment, clinical validation, and integration.

RESULTS

We found 156 studies; 146/156 (94%) fullfilled the first QNI step, 155/156 (99%) the second, 8/156 (5%) the third, 3/156 (2%) the fourth, 5/156 (3%) the fifth and only one the sixth.

CONCLUSIONS

To date, little has been done to evaluate the clinical performance and the integration in the clinical workflow of available methods for MS lesion detection/segmentation. In addition, the socio-economic effects and the impact on patients' management of such tools remain almost unexplored.

Fetal brain tissue annotation and segmentation challenge results

In-utero fetal MRI is emerging as an important tool in the diagnosis and analysis of the developing human brain. Automatic segmentation of the developing fetal brain is a vital step in the quantitative analysis of prenatal neurodevelopment both in the research and clinical context. However, manual segmentation of cerebral structures is time-consuming and prone to error and inter-observer variability. Therefore, we organized the Fetal Tissue Annotation (FeTA) Challenge in 2021 in order to encourage the development of automatic segmentation algorithms on an international level. The challenge utilized FeTA Dataset, an open dataset of fetal brain MRI reconstructions segmented into seven different tissues (external cerebrospinal fluid, gray matter, white matter, ventricles, cerebellum, brainstem, deep gray matter). 20 international teams participated in this challenge, submitting a total of 21 algorithms for evaluation. In this paper, we provide a detailed analysis of the results from both a technical and clinical perspective. All participants relied on deep learning methods, mainly U-Nets, with some variability present in the network architecture, optimization, and image pre- and post-processing. The majority of teams used existing medical imaging deep learning frameworks. The main differences between the submissions were the fine tuning done during training, and the specific pre- and post-processing steps performed. The challenge results showed that almost all submissions performed similarly. Four of the top five teams used ensemble learning methods. However, one team's algorithm performed significantly superior to the other submissions, and consisted of an asymmetrical U-Net network architecture. This paper provides a first of its kind benchmark for future automatic multi-tissue segmentation algorithms for the developing human brain in utero.

Direct segmentation of brain white matter tracts in diffusion MRI

The brain white matter consists of a set of tracts that connect distinct regions of the brain. Segmentation of these tracts is often needed for clinical and research studies. Diffusion-weighted MRI offers unique contrast to delineate these tracts. However, existing segmentation methods rely on intermediate computations such as tractography or estimation of fiber orientation density. These intermediate computations, in turn, entail complex computations that can result in unnecessary errors. Moreover, these intermediate computations often require dense multi-shell measurements that are unavailable in many clinical and research applications. As a result, current methods suffer from low accuracy and poor generalizability. Here, we propose a new deep learning method that segments these tracts directly from the diffusion MRI data, thereby sidestepping the intermediate computation errors. Our experiments show that this method can achieve segmentation accuracy that is on par with the state of the art methods (mean Dice Similarity Coefficient of 0.826). Compared with the state of the art, our method offers far superior generalizability to undersampled data that are typical of clinical studies and to data obtained with different acquisition protocols. Moreover, we propose a new method for detecting inaccurate segmentations and show that it is more accurate than standard methods that are based on estimation uncertainty quantification. The new methods can serve many critically important clinical and scientific applications that require accurate and reliable non-invasive segmentation of white matter tracts.

Deep learning microstructure estimation of developing brains from diffusion MRI: a newborn and fetal study

Diffusion-weighted magnetic resonance imaging (dMRI) is widely used to assess the brain white matter. Fiber orientation distribution functions (FODs) are a common way of representing the orientation and density of white matter fibers. However, with standard FOD computation methods, accurate estimation of FODs requires a large number of measurements that usually cannot be acquired for newborns and fetuses. We propose to overcome this limitation by using a deep learning method to map as few as six diffusion-weighted measurements to the target FOD. To train the model, we use the FODs computed using multi-shell high angular resolution measurements as target. Extensive quantitative evaluations show that the new deep learning method, using significantly fewer measurements, achieves comparable or superior results to standard methods such as Constrained Spherical Deconvolution. We demonstrate the generalizability of the new deep learning method across scanners, acquisition protocols, and anatomy on two clinical datasets of newborns and fetuses. Additionally, we compute agreement metrics within the HARDI newborn dataset, and validate fetal FODs with post-mortem histological data. The results of this study show the advantage of deep learning in inferring the microstructure of the developing brain from in-vivo dMRI measurements that are often very limited due to subject motion and limited acquisition times, but also highlight the intrinsic limitations of dMRI in the analysis of the developing brain microstructure. These findings, therefore, advocate for the need for improved methods that are tailored to studying the early development of human brain.

Temporo-basal sulcal connections: a manual annotation protocol and an investigation of sexual dimorphism and heritability

The temporo-basal region of the human brain is composed of the collateral, the occipito-temporal, and the rhinal sulci. We manually rated (using a novel protocol) the connections between rhinal/collateral (RS-CS), collateral/occipito-temporal (CS-OTS) and rhinal/occipito-temporal (RS-OTS) sulci, using the MRI of nearly 3400 individuals including around 1000 twins. We reported both the associations between sulcal polymorphisms as well with a wide range of demographics (e.g. age, sex, handedness). Finally, we also estimated the heritability, and the genetic correlation between sulcal connections. We reported the frequency of the sulcal connections in the general population, which were hemisphere dependent. We found a sexual dimorphism of the connections, especially marked in the right hemisphere, with a CS-OTS connection more frequent in females (approximately 35-40% versus 20-25% in males) and an RS-CS connection more common in males (approximately 40-45% versus 25-30% in females). We confirmed associations between sulcal connections and characteristics of incomplete hippocampal inversion (IHI). We estimated the broad sense heritability to be 0.28-0.45 for RS-CS and CS-OTS connections, with hints of dominant contribution for the RS-CS connection. The connections appeared to share some of their genetic causing factors as indicated by strong genetic correlations. Heritability appeared much smaller for the (rarer) RS-OTS connection.

BigBrain-MR: a new digital phantom with anatomically-realistic magnetic resonance properties at 100-µm resolution for magnetic resonance methods development

The benefits, opportunities and growing availability of ultra-high field magnetic resonance imaging (MRI) for humans have prompted an expansion in research and development efforts towards increasingly more advanced high-resolution imaging techniques. To maximize their effectiveness, these efforts need to be supported by powerful computational simulation platforms that can adequately reproduce the biophysical characteristics of MRI, with high spatial resolution. In this work, we have sought to address this need by developing a novel digital phantom with realistic anatomical detail up to 100-µm resolution, including multiple MRI properties that affect image generation. This phantom, termed BigBrain-MR, was generated from the publicly available BigBrain histological dataset and lower-resolution in-vivo 7T-MRI data, using a newly-developed image processing framework that allows mapping the general properties of the latter into the fine anatomical scale of the former. Overall, the mapping framework was found to be effective and robust, yielding a diverse range of realistic "in-vivo-like" MRI contrasts and maps at 100-µm resolution. BigBrain-MR was then tested in three imaging applications (motion effects and interpolation, super-resolution imaging, and parallel imaging reconstruction) to investigate its properties, value and validity as a simulation platform. The results consistently showed that BigBrain-MR can closely approximate the behavior of real in-vivo data, more realistically and with more extensive features than a more classic option such as the Shepp-Logan phantom. Its flexibility in simulating different contrast mechanisms and artifacts may also prove valuable for educational applications. BigBrain-MR is therefore deemed a favorable choice to support methodological development and demonstration in brain MRI, and has been made freely available to the community.

Multimodal Magnetic Resonance Imaging Depicts Widespread and Subregion Specific Anomalies in the Thalamus of Early-Psychosis and Chronic Schizophrenia Patients

BACKGROUND AND HYPOTHESIS

Although the thalamus has a central role in schizophrenia pathophysiology, contributing to sensory, cognitive, and sleep alterations, the nature and dynamics of the alterations occurring within this structure remain largely elusive. Using a multimodal magnetic resonance imaging (MRI) approach, we examined whether anomalies: (1) differ across thalamic subregions/nuclei, (2) are already present in the early phase of psychosis (EP), and (3) worsen in chronic schizophrenia (SCHZ).

STUDY DESIGN

T1-weighted and diffusion-weighted images were analyzed to estimate gray matter concentration (GMC) and microstructural parameters obtained from the spherical mean technique (intra-neurite volume fraction [VFINTRA)], intra-neurite diffusivity [DIFFINTRA], extra-neurite mean diffusivity [MDEXTRA], extra-neurite transversal diffusivity [TDEXTRA]) within 7 thalamic subregions.

RESULTS

Compared to age-matched controls, the thalamus of EP patients displays previously unreported widespread microstructural alterations (VFINTRA decrease, TDEXTRA increase) that are associated with similar alterations in the whole brain white matter, suggesting altered integrity of white matter fiber tracts in the thalamus. In both patient groups, we also observed more localized and heterogenous changes (either GMC decrease, MDEXTRA increase, or DIFFINTRA decrease) in mediodorsal, posterior, and ventral anterior parts of the thalamus in both patient groups, suggesting that the nature of the alterations varies across subregions. GMC and DIFFINTRA in the whole thalamus correlate with global functioning, while DIFFINTRA in the subregion encompassing the medial pulvinar is significantly associated with negative symptoms in SCHZ.

CONCLUSION

Our data reveals both widespread and more localized thalamic anomalies that are already present in the early phase of psychosis.

BMAT: An open-source BIDS managing and analysis tool

Magnetic Resonance Imaging (MRI) is an established technique to study in vivo neurological disorders such as Multiple Sclerosis (MS). To avoid errors on MRI data organization and automated processing, a standard called Brain Imaging Data Structure (BIDS) has been recently proposed. The BIDS standard eases data sharing and processing within or between centers by providing guidelines for their description and organization. However, the transformation from the complex unstructured non-open file data formats coming directly from the MRI scanner to a correct BIDS structure can be cumbersome and time consuming. This hinders a wider adoption of the BIDS format across different study centers. To solve this problem and ease the day-to-day use of BIDS for the neuroimaging scientific community, we present the BIDS Managing and Analysis Tool (BMAT). The BMAT software is a complete and easy-to-use local open-source neuroimaging analysis tool with a graphical user interface (GUI) that uses the BIDS format to organize and process brain MRI data for MS imaging research studies. BMAT provides the possibility to translate data from MRI scanners to the BIDS structure, create and manage BIDS datasets as well as develop and run automated processing pipelines, and is faster than its competitor. BMAT software propose the possibility to download useful analysis apps, especially applied to MS research with lesion segmentation and processing of imaging contrasts for novel disease biomarkers such as the central vein sign and the paramagnetic rim lesions.

Cortical lesions, central vein sign, and paramagnetic rim lesions in multiple sclerosis: Emerging machine learning techniques and future avenues

The current diagnostic criteria for multiple sclerosis (MS) 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, some 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 as well for CL, PRL, and CVS. In the present review, we first introduce these MS biomarkers and their imaging methods. Subsequently, we describe the corresponding machine learning-based methods that were proposed to tackle these clinical questions, putting them into context with respect to the challenges they are 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.

Fluid and White Matter Suppression: New Sensitive 3 T Magnetic Resonance Imaging Contrasts for Cortical Lesion Detection in Multiple Sclerosis

OBJECTIVE

Cortical lesions are common in multiple sclerosis (MS), but their visualization is challenging on conventional magnetic resonance imaging. The uniform image derived from magnetization prepared 2 rapid acquisition gradient echoes (MP2RAGE uni ) detects cortical lesions with a similar rate as the criterion standard sequence, double inversion recovery. Fluid and white matter suppression (FLAWS) provides multiple reconstructed contrasts acquired during a single acquisition. These contrasts include FLAWS minimum image (FLAWS min ), which provides an exquisite sensitivity to the gray matter signal and therefore may facilitate cortical lesion identification, as well as high contrast FLAWS (FLAWS hco ), which gives a contrast that is similar to one of MP2RAGE uni . In this study, we compared the manual detection rate of cortical lesions on MP2RAGE uni , FLAWS min , and FLAWS hco in MS patients. Furthermore, we assessed whether the combined detection rate on FLAWS min and FLAWS hco was superior to MP2RAGE uni for cortical lesions identification. Last, we compared quantitative T1 maps (qT1) provided by both MP2RAGE and FLAWS in MS lesions.

MATERIALS AND METHODS

We included 30 relapsing-remitting MS patients who underwent MP2RAGE and FLAWS magnetic resonance imaging with isotropic spatial resolution of 1 mm at 3 T. Cortical lesions were manually segmented by consensus of 3 trained raters and classified as intracortical or leukocortical lesions on (1) MP2RAGE uniform/flat images, (2) FLAWS min , and (3) FLAWS hco . In addition, segmented lesions on FLAWS min and FLAWS hco were merged to produce a union lesion map (FLAWS min + hco ). Number and volume of all cortical, intracortical, and leukocortical lesions were compared among MP2RAGE uni , FLAWS min , and FLAWS hco using Friedman test and between MP2RAGE uni and FLAWS min + hco using Wilcoxon signed rank test. The FLAWS T1 maps were then compared with the reference MP2RAGE T1 maps using relative differences in percentage. In an exploratory analysis, individual cortical lesion counts of the 3 raters were compared, and interrater variability was quantified using Fleiss ϰ.

RESULTS

In total, 633 segmentations were made on the 3 contrasts, corresponding to 355 cortical lesions. The median number and volume of single cortical, intracortical, and leukocortical lesions were comparable among MP2RAGE uni , FLAWS min , and FLAWS hco . In patients with cortical lesions (22/30), median cumulative lesion volume was larger on FLAWS min (587 μL; IQR, 1405 μL) than on MP2RAGE uni (490 μL; IQR, 990 μL; P = 0.04), whereas there was no difference between FLAWS min and FLAWS hco , or FLAWS hco and MP2RAGE uni . FLAWS min + hco showed significantly greater numbers of cortical (median, 4.5; IQR, 15) and leukocortical (median, 3.5; IQR, 12) lesions than MP2RAGE uni (median, 3; IQR, 10; median, 2.5; IQR, 7; both P < 0.001). Interrater agreement was moderate on MP2RAGE uni (ϰ = 0.582) and FLAWS hco (ϰ = 0.584), but substantial on FLAWS min (ϰ = 0.614). qT1 in lesions was similar between MP2RAGE and FLAWS.

CONCLUSIONS

Cortical lesions identification in FLAWS min and FLAWS hco was comparable to MP2RAGE uni . The combination of FLAWS min and FLAWS hco allowed to identify a higher number of cortical lesions than MP2RAGE uni , whereas qT1 maps did not differ between the 2 acquisition schemes.

A comparative assessment of myelin-sensitive measures in multiple sclerosis patients and healthy subjects

INTRODUCTION

Multiple Sclerosis (MS) is a common neurological disease primarily characterized by myelin damage in lesions and in normal - appearing white and gray matter (NAWM, NAGM). Several quantitative MRI (qMRI) methods are sensitive to myelin characteristics by measuring specific tissue biophysical properties. However, there are currently few studies assessing the relative reproducibility and sensitivity of qMRI measures to MS pathology in vivo in patients.

METHODS

We performed two studies. The first study assessed of the sensitivity of qMRI measures to MS pathology: in this work, we recruited 150 MS and 100 healthy subjects, who underwent brain MRI at 3 T including quantitative T1 mapping (qT1), quantitative susceptibility mapping (QSM), magnetization transfer saturation imaging (MTsat) and myelin water imaging for myelin water fraction (MWF). The sensitivity of qMRIs to MS focal pathology (MS lesions vs peri-plaque white/gray matter (PPWM/PPGM)) was studied lesion-wise; the sensitivity to diffuse normal appearing (NA) pathology was measured using voxel-wise threshold-free cluster enhancement (TFCE) in NAWM and vertex-wise inflated cortex analysis in NAGM. Furthermore, the sensitivity of qMRI to the identification of lesion tissue was investigated using a voxel-wise logistic regression analysis to distinguish MS lesion and PP voxels. The second study assessed the reproducibility of myelin-sensitive qMRI measures in a single scanner. To evaluate the intra-session and inter-session reproducibility of qMRI measures, we have investigated 10 healthy subjects, who underwent two brain 3 T MRIs within the same day (without repositioning), and one after 1-week interval. Five region of interest (ROIs) in white and deep grey matter areas were segmented, and inter- and intra- session reproducibility was studied using the intra-class correlation coefficient (ICC). Further, we also investigated the voxel-wise reproducibility of qMRI measures in NAWM and NAGM.

RESULTS

qT1 and QSM showed the highest sensitivity to distinguish MS focal WM and cortical pathology from peri-plaque WM (P < 0.0001), although QSM also showed the highest variance when applied to lesions. MWF and MTsat exhibited the highest sensitivity to NAWM pathology (P < 0.01). On the other hand, qT1 appeared to be the most sensitive measure to NAGM pathology (P < 0.01). All myelin-sensitive qMRI measures exhibited high inter/intra sessional ICCs in various WM and deep GM ROIs, in NAWM and in NAGM (ICC 0.82 ± 0.12).

CONCLUSION

This work shows that the applied qT1, MWF, MTsat and QSM are highly reproducible and exhibit differential sensitivity to focal and diffuse WM and GM pathology in MS patients.

A multi-scale probabilistic atlas of the human connectome

The human brain is a complex system that can be efficiently represented as a network of structural connectivity. Many imaging studies would benefit from such network information, which is not always available. In this work, we present a whole-brain multi-scale structural connectome atlas. This tool has been derived from a cohort of 66 healthy subjects imaged with optimal technology in the setting of the Human Connectome Project. From these data we created, using extensively validated diffusion-data processing, tractography and gray-matter parcellation tools, a multi-scale probabilistic atlas of the human connectome. In addition, we provide user-friendly and accessible code to match this atlas to individual brain imaging data to extract connection-specific quantitative information. This can be used to associate individual imaging findings, such as focal white-matter lesions or regional alterations, to specific connections and brain circuits. Accordingly, network-level consequences of regional changes can be analyzed even in absence of diffusion and tractography data. This method is expected to broaden the accessibility and lower the yield for connectome research.

Multiple sclerosis cortical lesion detection with deep learning at ultra-high-field MRI

Manually segmenting multiple sclerosis (MS) cortical lesions (CLs) is extremely time consuming, and past studies have shown only moderate inter-rater reliability. To accelerate this task, we developed a deep-learning-based framework (CLAIMS: Cortical Lesion AI-Based Assessment in Multiple Sclerosis) for the automated detection and classification of MS CLs with 7 T MRI. Two 7 T datasets, acquired at different sites, were considered. The first consisted of 60 scans that include 0.5 mm isotropic MP2RAGE acquired four times (MP2RAGE×4), 0.7 mm MP2RAGE, 0.5 mm T₂ *-weighted GRE, and 0.5 mm T₂ *-weighted EPI. The second dataset consisted of 20 scans including only 0.75 × 0.75 × 0.9 mm³ MP2RAGE. CLAIMS was first evaluated using sixfold cross-validation with single and multi-contrast 0.5 mm MRI input. Second, the performance of the model was tested on 0.7 mm MP2RAGE images after training with either 0.5 mm MP2RAGE×4, 0.7 mm MP2RAGE, or alternating the two. Third, its generalizability was evaluated on the second external dataset and compared with a state-of-the-art technique based on partial volume estimation and topological constraints (MSLAST). CLAIMS trained only with MP2RAGE×4 achieved results comparable to those of the multi-contrast model, reaching a CL true positive rate of 74% with a false positive rate of 30%. Detection rate was excellent for leukocortical and subpial lesions (83%, and 70%, respectively), whereas it reached 53% for intracortical lesions. The correlation between disability measures and CL count was similar for manual and CLAIMS lesion counts. Applying a domain-scanner adaptation approach and testing CLAIMS on the second dataset, the performance was superior to MSLAST when considering a minimum lesion volume of 6 μL (lesion-wise detection rate of 71% versus 48%). The proposed framework outperforms previous state-of-the-art methods for automated CL detection across scanners and protocols. In the future, CLAIMS may be useful to support clinical decisions at 7 T MRI, especially in the field of diagnosis and differential diagnosis of MS patients.

A new advanced MRI biomarker for remyelinated lesions in Multiple Sclerosis

Objectives

Neuropathological studies have shown that multiple sclerosis (MS) lesions are heterogeneous in terms of myelin/axon damage and repair as well as iron content. However, it remains a challenge to identify specific chronic lesion types, especially remyelinated lesions, in vivo in patients with MS.

Methods

We performed 3 studies: (1) a cross‐sectional study in a prospective cohort of 115 patients with MS and 76 healthy controls, who underwent 3 T magnetic resonance imaging (MRI) for quantitative susceptibility mapping (QSM), myelin water fraction (MWF), and neurite density index (NDI) maps. White matter (WM) lesions in QSM were classified into 5 QSM lesion types (iso‐intense, hypo‐intense, hyperintense, lesions with hypo‐intense rims, and lesions with paramagnetic rim legions [PRLs]); (2) a longitudinal study of 40 patients with MS to study the evolution of lesions over 2 years; (3) a postmortem histopathology‐QSM validation study in 3 brains of patients with MS to assess the accuracy of QSM classification to identify neuropathological lesion types in 63 WM lesions.

Results

At baseline, hypo‐ and isointense lesions showed higher mean MWF and NDI values compared to other QSM lesion types (p < 0.0001). Further, at 2‐year follow‐up, hypo‐/iso‐intense lesions showed an increase in MWF. Postmortem analyses revealed that QSM highly accurately identifies (1) fully remyelinated areas as hypo‐/iso‐intense (sensitivity = 88.89% and specificity = 100%), (2) chronic inactive lesions as hyperintense (sensitivity = 71.43% and specificity = 92.00%), and (3) chronic active/smoldering lesions as PRLs (sensitivity = 92.86% and specificity = 86.36%).

Interpretation

These results provide the first evidence that it is possible to distinguish chronic MS lesions in a clinical setting, hereby supporting with new biomarkers to develop and assess remyelinating treatments. ANN NEUROL 2022;92:486–502

Through-Plane Super-Resolution With Autoencoders in Diffusion Magnetic Resonance Imaging of the Developing Human Brain

Fetal brain diffusion magnetic resonance images (MRI) are often acquired with a lower through-plane than in-plane resolution. This anisotropy is often overcome by classical upsampling methods such as linear or cubic interpolation. In this work, we employ an unsupervised learning algorithm using an autoencoder neural network for single-image through-plane super-resolution by leveraging a large amount of data. Our framework, which can also be used for slice outliers replacement, overperformed conventional interpolations quantitatively and qualitatively on pre-term newborns of the developing Human Connectome Project. The evaluation was performed on both the original diffusion-weighted signal and the estimated diffusion tensor maps. A byproduct of our autoencoder was its ability to act as a denoiser. The network was able to generalize fetal data with different levels of motions and we qualitatively showed its consistency, hence supporting the relevance of pre-term datasets to improve the processing of fetal brain images.

An automatic multi-tissue human fetal brain segmentation benchmark using the Fetal Tissue Annotation Dataset

It is critical to quantitatively analyse the developing human fetal brain in order to fully understand neurodevelopment in both normal fetuses and those with congenital disorders. To facilitate this analysis, automatic multi-tissue fetal brain segmentation algorithms are needed, which in turn requires open datasets of segmented fetal brains. Here we introduce a publicly available dataset of 50 manually segmented pathological and non-pathological fetal magnetic resonance brain volume reconstructions across a range of gestational ages (20 to 33 weeks) into 7 different tissue categories (external cerebrospinal fluid, grey matter, white matter, ventricles, cerebellum, deep grey matter, brainstem/spinal cord). In addition, we quantitatively evaluate the accuracy of several automatic multi-tissue segmentation algorithms of the developing human fetal brain. Four research groups participated, submitting a total of 10 algorithms, demonstrating the benefits the dataset for the development of automatic algorithms.

Chronic White Matter Inflammation and Serum Neurofilament Levels in Multiple Sclerosis

OBJECTIVE

To assess whether chronic white matter inflammation in patients with multiple sclerosis (MS) as detected in vivo by paramagnetic rim MRI lesions (PRLs) is associated with higher serum neurofilament light chain (sNfL) levels, a marker of neuroaxonal damage.

METHODS

In 118 patients with MS with no gadolinium-enhancing lesions or recent relapses, we analyzed 3D-submillimeter phase MRI and sNfL levels. Histopathologic evaluation was performed in 25 MS lesions from 20 additional autopsy MS cases.

RESULTS

In univariable analyses, participants with ≥2 PRLs (n = 43) compared to those with ≤1 PRL (n = 75) had higher age-adjusted sNfL percentiles (median, 91 and 68; p < 0.001) and higher Multiple Sclerosis Severity Scale scores (MSSS median, 4.3 and 2.4; p = 0.003). In multivariable analyses, sNfL percentile levels were higher in PRLs ≥2 cases (β_add, 16.3; 95% confidence interval [CI], 4.6-28.0; p < 0.01), whereas disease-modifying treatment (DMT), Expanded Disability Status Scale (EDSS) score, and T2 lesion load did not affect sNfL. In a similar model, sNfL percentile levels were highest in cases with ≥4 PRLs (n = 30; β_add, 30.4; 95% CI, 15.6-45.2; p < 0.01). Subsequent multivariable analysis revealed that PRLs ≥2 cases also had higher MSSS (β_add, 1.1; 95% CI, 0.3-1.9; p < 0.01), whereas MSSS was not affected by DMT or T2 lesion load. On histopathology, both chronic active and smoldering lesions exhibited more severe acute axonal damage at the lesion edge than in the lesion center (edge vs center: p = 0.004 and p = 0.0002, respectively).

CONCLUSION

Chronic white matter inflammation was associated with increased levels of sNfL and disease severity in nonacute MS, suggesting that PRL contribute to clinically relevant, inflammation-driven neurodegeneration.

GAMER-MRI in Multiple Sclerosis Identifies the Diffusion-Based Microstructural Measures That Are Most Sensitive to Focal Damage: A Deep-Learning-Based Analysis and Clinico-Biological Validation

Conventional magnetic resonance imaging (cMRI) in multiple sclerosis (MS) patients provides measures of focal brain damage and activity, which are fundamental for disease diagnosis, prognosis, and the evaluation of response to therapy. However, cMRI is insensitive to the damage to the microenvironment of the brain tissue and the heterogeneity of MS lesions. In contrast, the damaged tissue can be characterized by mathematical models on multishell diffusion imaging data, which measure different compartmental water diffusion. In this work, we obtained 12 diffusion measures from eight diffusion models, and we applied a deep-learning attention-based convolutional neural network (CNN) (GAMER-MRI) to select the most discriminating measures in the classification of MS lesions and the perilesional tissue by attention weights. Furthermore, we provided clinical and biological validation of the chosen metrics-and of their most discriminative combinations-by correlating their respective mean values in MS patients with the corresponding Expanded Disability Status Scale (EDSS) and the serum level of neurofilament light chain (sNfL), which are measures of disability and neuroaxonal damage. Our results show that the neurite density index from neurite orientation and dispersion density imaging (NODDI), the measures of the intra-axonal and isotropic compartments from microstructural Bayesian approach, and the measure of the intra-axonal compartment from the spherical mean technique NODDI were the most discriminating (respective attention weights were 0.12, 0.12, 0.15, and 0.13). In addition, the combination of the neurite density index from NODDI and the measures for the intra-axonal and isotropic compartments from the microstructural Bayesian approach exhibited a stronger correlation with EDSS and sNfL than the individual measures. This work demonstrates that the proposed method might be useful to select the microstructural measures that are most discriminative of focal tissue damage and that may also be combined to a unique contrast to achieve stronger correlations to clinical disability and neuroaxonal damage.

Myelin and axon pathology in multiple sclerosis assessed by myelin water and multi-shell diffusion imaging

Damage to the myelin sheath and the neuroaxonal unit is a cardinal feature of multiple sclerosis; however, a detailed characterization of the interaction between myelin and axon damage in vivo remains challenging. We applied myelin water and multi-shell diffusion imaging to quantify the relative damage to myelin and axons (i) among different lesion types; (ii) in normal-appearing tissue; and (iii) across multiple sclerosis clinical subtypes and healthy controls. We also assessed the relation of focal myelin/axon damage with disability and serum neurofilament light chain as a global biological measure of neuroaxonal damage. Ninety-one multiple sclerosis patients (62 relapsing-remitting, 29 progressive) and 72 healthy controls were enrolled in the study. Differences in myelin water fraction and neurite density index were substantial when lesions were compared to healthy control subjects and normal-appearing multiple sclerosis tissue: both white matter and cortical lesions exhibited a decreased myelin water fraction and neurite density index compared with healthy (P < 0.0001) and peri-plaque white matter (P < 0.0001). Periventricular lesions showed decreased myelin water fraction and neurite density index compared with lesions in the juxtacortical region (P < 0.0001 and P < 0.05). Similarly, lesions with paramagnetic rims showed decreased myelin water fraction and neurite density index relative to lesions without a rim (P < 0.0001). Also, in 75% of white matter lesions, the reduction in neurite density index was higher than the reduction in the myelin water fraction. Besides, normal-appearing white and grey matter revealed diffuse reduction of myelin water fraction and neurite density index in multiple sclerosis compared to healthy controls (P < 0.01). Further, a more extensive reduction in myelin water fraction and neurite density index in normal-appearing cortex was observed in progressive versus relapsing-remitting participants. Neurite density index in white matter lesions correlated with disability in patients with clinical deficits (P < 0.01, beta = -10.00); and neurite density index and myelin water fraction in white matter lesions were associated to serum neurofilament light chain in the entire patient cohort (P < 0.01, beta = -3.60 and P < 0.01, beta = 0.13, respectively). These findings suggest that (i) myelin and axon pathology in multiple sclerosis is extensive in both lesions and normal-appearing tissue; (ii) particular types of lesions exhibit more damage to myelin and axons than others; (iii) progressive patients differ from relapsing-remitting patients because of more extensive axon/myelin damage in the cortex; and (iv) myelin and axon pathology in lesions is related to disability in patients with clinical deficits and global measures of neuroaxonal damage.

MPRAGE to MP2RAGE UNI translation via generative adversarial network improves the automatic tissue and lesion segmentation in multiple sclerosis patients

BACKGROUND AND OBJECTIVE

Compared to the conventional magnetization-prepared rapid gradient-echo imaging (MPRAGE) MRI sequence, the specialized magnetization prepared 2 rapid acquisition gradient echoes (MP2RAGE) shows a higher brain tissue and lesion contrast in multiple sclerosis (MS) patients. The goal of this work is to retrospectively generate realistic-looking MP2RAGE uniform images (UNI) from already acquired MPRAGE images in order to improve the automatic lesion and tissue segmentation.

METHODS

For this task we propose a generative adversarial network (GAN). Multi-contrast MRI data of 12 healthy controls and 44 patients diagnosed with MS was retrospectively analyzed. Imaging was acquired at 3T using a SIEMENS scanner with MPRAGE, MP2RAGE, FLAIR, and DIR sequences. We train the GAN with both healthy controls and MS patients to generate synthetic MP2RAGE UNI images. These images were then compared to the real MP2RAGE UNI (considered as ground truth) analyzing the output of automatic brain tissue and lesion segmentation tools. Reference-based metrics as well as the lesion-wise true and false positives, Dice coefficient, and volume difference were considered for the evaluation. Statistical differences were assessed with the Wilcoxon signed-rank test.

RESULTS

The synthetic MP2RAGE UNI significantly improves the lesion and tissue segmentation masks in terms of Dice coefficient and volume difference (p-values < 0.001) compared to the MPRAGE. For the segmentation metrics analyzed no statistically significant differences are found between the synthetic and acquired MP2RAGE UNI.

CONCLUSION

Synthesized MP2RAGE UNI images are visually realistic and improve the output of automatic segmentation tools.

Ophthalmic Magnetic Resonance Imaging: Where Are We (Heading To)?

Magnetic resonance imaging of the eye and orbit (MReye) is a cross-domain research field, combining (bio)physics, (bio)engineering, physiology, data sciences and ophthalmology. A growing number of reports document technical innovations of MReye and promote their application in preclinical research and clinical science. Realizing the progress and promises, this review outlines current trends in MReye. Examples of MReye strategies and their clinical relevance are demonstrated. Frontier applications in ocular oncology, refractive surgery, ocular muscle disorders and orbital inflammation are presented and their implications for explorations into ophthalmic diseases are provided. Substantial progress in anatomically detailed, high-spatial resolution MReye of the eye, orbit and optic nerve is demonstrated. Recent developments in MReye of ocular tumors are explored, and its value for personalized eye models derived from machine learning in the treatment planning of uveal melanoma and evaluation of retinoblastoma is highlighted. The potential of MReye for monitoring drug distribution and for improving treatment management and the assessment of individual responses is discussed. To open a window into the eye and into (patho)physiological processes that in the past have been largely inaccessible, advances in MReye at ultrahigh magnetic field strengths are discussed. A concluding section ventures a glance beyond the horizon and explores future directions of MReye across multiple scales, including in vivo electrolyte mapping of sodium and other nuclei. This review underscores the need for the (bio)medical imaging and ophthalmic communities to expand efforts to find solutions to the remaining unsolved problems and technical obstacles of MReye, with the objective to transfer methodological advancements driven by MR physics into genuine clinical value.

White Matter Alterations Between Brain Network Hubs Underlie Processing Speed Impairment in Patients With Schizophrenia

Processing speed (PS) impairment is one of the most severe and common cognitive deficits in schizophrenia. Previous studies have reported correlations between PS and white matter diffusion properties, including fractional anisotropy (FA), in several fiber bundles in schizophrenia, suggesting that white matter alterations could underpin decreased PS. In schizophrenia, white matter alterations are most prevalent within inter-hub connections of the rich club. However, the spatial and topological characteristics of this association between PS and FA have not been investigated in patients. In this context, we tested whether structural connections comprising the rich club network would underlie PS impairment in 298 patients with schizophrenia or schizoaffective disorder and 190 healthy controls from the Australian Schizophrenia Research Bank. PS, measured using the digit symbol coding task, was largely (Cohen’s d = 1.33) and significantly (P < .001) reduced in the patient group when compared with healthy controls. Significant associations between PS and FA were widespread in the patient group, involving all cerebral lobes. FA was not associated with other cognitive measures of phonological fluency and verbal working memory in patients, suggesting specificity to PS. A topological analysis revealed that despite being spatially widespread, associations between PS and FA were over-represented among connections forming the rich club network. These findings highlight the need to consider brain network topology when investigating high-order cognitive functions that may be spatially distributed among several brain regions. They also reinforce the evidence that brain hubs and their interconnections may be particularly vulnerable parts of the brain in schizophrenia.

Fetal Brain Biometric Measurements on 3D Super-Resolution Reconstructed T2-Weighted MRI: An Intra- and Inter-observer Agreement Study

We present the comparison of two-dimensional (2D) fetal brain biometry on magnetic resonance (MR) images using orthogonal 2D T2-weighted sequences (T2WSs) vs. one 3D super-resolution (SR) reconstructed volume and evaluation of the level of confidence and concordance between an experienced pediatric radiologist (obs1) and a junior radiologist (obs2). Twenty-five normal fetal brain MRI scans (18-34 weeks of gestation) including orthogonal 3-mm-thick T2WSs were analyzed retrospectively. One 3D SR volume was reconstructed per subject based on multiple series of T2WSs. The two observers performed 11 2D biometric measurements (specifying their level of confidence) on T2WS and SR volumes. Measurements were compared using the paired Wilcoxon rank sum test between observers for each dataset (T2WS and SR) and between T2WS and SR for each observer. Bland-Altman plots were used to assess the agreement between each pair of measurements. Measurements were made with low confidence in three subjects by obs1 and in 11 subjects by obs2 (mostly concerning the length of the corpus callosum on T2WS). Inter-rater intra-dataset comparisons showed no significant difference (p > 0.05), except for brain axial biparietal diameter (BIP) on T2WS and for brain and skull coronal BIP and coronal transverse cerebellar diameter (DTC) on SR. None of them remained significant after correction for multiple comparisons. Inter-dataset intra-rater comparisons showed statistical differences in brain axial and coronal BIP for both observers, skull coronal BIP for obs1, and axial and coronal DTC for obs2. After correction for multiple comparisons, only axial brain BIP remained significantly different, but differences were small (2.95 ± 1.73 mm). SR allows similar fetal brain biometry as compared to using the conventional T2WS while improving the level of confidence in the measurements and using a single reconstructed volume.

Model-informed machine learning for multi-component T2 relaxometry

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 a priori knowledge of 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 a Gaussian Mixture Fitting (parametric) and Regularized Non-Negative Least Squares algorithms (non-parametric) 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 the non-parametric and parametric methods, respectively.

GAMER MRI: Gated-attention mechanism ranking of multi-contrast MRI in brain pathology

INTRODUCTION

During the last decade, a multitude of novel quantitative and semiquantitative MRI techniques have provided new information about the pathophysiology of neurological diseases. Yet, selection of the most relevant contrasts for a given pathology remains challenging. In this work, we developed and validated a method, Gated-Attention MEchanism Ranking of multi-contrast MRI in brain pathology (GAMER MRI), to rank the relative importance of MR measures in the classification of well understood ischemic stroke lesions. Subsequently, we applied this method to the classification of multiple sclerosis (MS) lesions, where the relative importance of MR measures is less understood.

METHODS

GAMER MRI was developed based on the gated attention mechanism, which computes attention weights (AWs) as proxies of importance of hidden features in the classification. In the first two experiments, we used Trace-weighted (Trace), apparent diffusion coefficient (ADC), Fluid-Attenuated Inversion Recovery (FLAIR), and T1-weighted (T1w) images acquired in 904 acute/subacute ischemic stroke patients and in 6,230 healthy controls and patients with other brain pathologies to assess if GAMER MRI could produce clinically meaningful importance orders in two different classification scenarios. In the first experiment, GAMER MRI with a pretrained convolutional neural network (CNN) was used in conjunction with Trace, ADC, and FLAIR to distinguish patients with ischemic stroke from those with other pathologies and healthy controls. In the second experiment, GAMER MRI with a patch-based CNN used Trace, ADC and T1w to differentiate acute ischemic stroke lesions from healthy tissue. The last experiment explored the performance of patch-based CNN with GAMER MRI in ranking the importance of quantitative MRI measures to distinguish two groups of lesions with different pathological characteristics and unknown quantitative MR features. Specifically, GAMER MRI was applied to assess the relative importance of the myelin water fraction (MWF), quantitative susceptibility mapping (QSM), T1 relaxometry map (qT1), and neurite density index (NDI) in distinguishing 750 juxtacortical lesions from 242 periventricular lesions in 47 MS patients. Pair-wise permutation t-tests were used to evaluate the differences between the AWs obtained for each quantitative measure.

RESULTS

In the first experiment, we achieved a mean test AUC of 0.881 and the obtained AWs of FLAIR and the sum of AWs of Trace and ADC were 0.11 and 0.89, respectively, as expected based on previous knowledge. In the second experiment, we achieved a mean test F1 score of 0.895 and a mean AW of Trace = 0.49, of ADC = 0.28, and of T1w = 0.23, thereby confirming the findings of the first experiment. In the third experiment, MS lesion classification achieved test balanced accuracy = 0.777, sensitivity = 0.739, and specificity = 0.814. The mean AWs of T1map, MWF, NDI, and QSM were 0.29, 0.26, 0.24, and 0.22 (p < 0.001), respectively.

CONCLUSIONS

This work demonstrates that the proposed GAMER MRI might be a useful method to assess the relative importance of MRI measures in neurological diseases with focal pathology. Moreover, the obtained AWs may in fact help to choose the best combination of MR contrasts for a specific classification problem.

RimNet: A deep 3D multimodal MRI architecture for paramagnetic rim lesion assessment in multiple sclerosis

OBJECTIVES

In multiple sclerosis (MS), the presence of a paramagnetic rim at the edge of non-gadolinium-enhancing lesions indicates perilesional chronic inflammation. Patients featuring a higher paramagnetic rim lesion burden tend to have more aggressive disease. The objective of this study was to develop and evaluate a convolutional neural network (CNN) architecture (RimNet) for automated detection of paramagnetic rim lesions in MS employing multiple magnetic resonance (MR) imaging contrasts.

MATERIALS AND METHODS

Imaging data were acquired at 3 Tesla on three different scanners from two different centers, totaling 124 MS patients, and studied retrospectively. Paramagnetic rim lesion detection was independently assessed by two expert raters on T2*-phase images, yielding 462 rim-positive (rim+) and 4857 rim-negative (rim-) lesions. RimNet was designed using 3D patches centered on candidate lesions in 3D-EPI phase and 3D FLAIR as input to two network branches. The interconnection of branches at both the first network blocks and the last fully connected layers favors the extraction of low and high-level multimodal features, respectively. RimNet's performance was quantitatively evaluated against experts' evaluation from both lesion-wise and patient-wise perspectives. For the latter, patients were categorized based on a clinically relevant threshold of 4 rim+ lesions per patient. The individual prediction capabilities of the images were also explored and compared (DeLong test) by testing a CNN trained with one image as input (unimodal).

RESULTS

The unimodal exploration showed the superior performance of 3D-EPI phase and 3D-EPI magnitude images in the rim+/- classification task (AUC = 0.913 and 0.901), compared to the 3D FLAIR (AUC = 0.855, Ps < 0.0001). The proposed multimodal RimNet prototype clearly outperformed the best unimodal approach (AUC = 0.943, P < 0.0001). The sensitivity and specificity achieved by RimNet (70.6% and 94.9%, respectively) are comparable to those of experts at the lesion level. In the patient-wise analysis, RimNet performed with an accuracy of 89.5% and a Dice coefficient (or F1 score) of 83.5%.

CONCLUSIONS

The proposed prototype showed promising performance, supporting the usage of RimNet for speeding up and standardizing the paramagnetic rim lesions analysis in MS.

Evolution of Cortical and White Matter Lesion Load in Early-Stage Multiple Sclerosis: Correlation With Neuroaxonal Damage and Clinical Changes

Introduction: Changes in cortical and white matter lesion (CL, WML) load are pivotal metrics to diagnose and monitor multiple sclerosis patients. Yet, the relationship between (i) changes in CL/WML load and disease progression and between (ii) changes in CL/WML load and neurodegeneration at early MS stages is not yet established. In this work, we have assessed the hypothesis that the combined CL and WML load as well as their 2-years evolution are surrogate markers of neurodegeneration and clinical progression at early MS stages. To achieve this goal, we have studied a group of RRMS patients and have investigated the impact of both CL and WML load on neuroaxonal damage as measured by serum neurofilament light chain (sNfL). Next, we have explored whether changes in CL/WML load over 2 years in the same cohort of early-MS are related to motor and cognitive changes.

Methods: Thirty-two RRMS patients (<5 years disease duration) underwent: (i) 3T MRI for CL/WML detection and clinical assessment at baseline and 2-years follow-up; and (ii) baseline blood test for sNfL. The correlation between the number and volume of CL/WML and sNfL was assessed by using the Spearman's rank correlation coefficient and a generalized linear model (GLM). A GLM was also used to assess the relationship between (i) the number/volume of new, enlarged, resolved, shrunken, stable lesions and (ii) the difference in clinical scores between two time-points.

Results: At baseline, sNfL levels correlated with both total CL count/volume (ρ = 0.6/0.7, Corr-P <0.017/Corr-P < 0.001) and with total WML count/volume (ρ = 0.6/0.6, Corr-P < 0.01 for both). Baseline sNfL levels also correlated with new WML count/volume (ρ = 0.6/0.5, Corr-P < 0.01/Corr-P < 0.05) but not with new CL. Longitudinal changes in CL and WML count and volume were significantly associated with (i) sustained attention, auditory information, processing speed and flexibility (p < 0.01), (ii) verbal memory (p < 0.01); (iii) verbal fluency (p < 0.05); and (iv) hand-motor function (p < 0.05).

Discussion: Changes in cortical and white matter focal damage in early MS patients correlate with global neuroaxonal damage and is associated to cognitive performances.

Partial-volume modeling reveals reduced gray matter in specific thalamic nuclei early in the time course of psychosis and chronic schizophrenia

The structural complexity of the thalamus, due to its mixed composition of gray and white matter, make it challenging to disjoint and quantify each tissue contribution to the thalamic anatomy. This work promotes the use of partial‐volume‐based over probabilistic‐based tissue segmentation approaches to better capture thalamic gray matter differences between patients at different stages of psychosis (early and chronic) and healthy controls. The study was performed on a cohort of 23 patients with schizophrenia, 41 with early psychosis and 69 age and sex‐matched healthy subjects. Six tissue segmentation approaches were employed to obtain the gray matter concentration/probability images. The statistical tests were applied at three different anatomical scales: whole thalamus, thalamic subregions and voxel‐wise. The results suggest that the partial volume model estimation of gray matter is more sensitive to detect atrophies within the thalamus of patients with psychosis. However all the methods detected gray matter deficit in the pulvinar, particularly in early stages of psychosis. This study demonstrates also that the gray matter decrease varies nonlinearly with age and between nuclei. While a gray matter loss was found in the pulvinar of patients in both stages of psychosis, reduced gray matter in the mediodorsal was only observed in early psychosis subjects. Finally, our analyses point to alterations in a sub‐region comprising the lateral posterior and ventral posterior nuclei. The obtained results reinforce the hypothesis that thalamic gray matter assessment is more reliable when the tissues segmentation method takes into account the partial volume effect.

CVSnet: A machine learning approach for automated central vein sign assessment in multiple sclerosis

The central vein sign (CVS) is an efficient imaging biomarker for multiple sclerosis (MS) diagnosis, but its application in clinical routine is limited by inter-rater variability and the expenditure of time associated with manual assessment. We describe a deep learning-based prototype for automated assessment of the CVS in white matter MS lesions using data from three different imaging centers. We retrospectively analyzed data from 3 T magnetic resonance images acquired on four scanners from two different vendors, including adults with MS (n = 42), MS mimics (n = 33, encompassing 12 distinct neurological diseases mimicking MS) and uncertain diagnosis (n = 5). Brain white matter lesions were manually segmented on FLAIR* images. Perivenular assessment was performed according to consensus guidelines and used as ground truth, yielding 539 CVS-positive (CVS⁺ ) and 448 CVS-negative (CVS^- ) lesions. A 3D convolutional neural network ("CVSnet") was designed and trained on 47 datasets, keeping 33 for testing. FLAIR* lesion patches of CVS⁺ /CVS^- lesions were used for training and validation (n = 375/298) and for testing (n = 164/150). Performance was evaluated lesion-wise and subject-wise and compared with a state-of-the-art vesselness filtering approach through McNemar's test. The proposed CVSnet approached human performance, with lesion-wise median balanced accuracy of 81%, and subject-wise balanced accuracy of 89% on the validation set, and 91% on the test set. The process of CVS assessment, in previously manually segmented lesions, was ~ 600-fold faster using the proposed CVSnet compared with human visual assessment (test set: 4 seconds vs. 40 minutes). On the validation and test sets, the lesion-wise performance outperformed the vesselness filter method (P < 0.001). The proposed deep learning prototype shows promising performance in differentiating MS from its mimics. Our approach was evaluated using data from different hospitals, enabling larger multicenter trials to evaluate the benefit of introducing the CVS marker into MS diagnostic criteria.

Improved susceptibility-weighted imaging for high contrast and resolution thalamic nuclei mapping at 7T

PURPOSE

The thalamus is an important brain structure and neurosurgical target, but its constituting nuclei are challenging to image non-invasively. Recently, susceptibility-weighted imaging (SWI) at ultra-high field has shown promising capabilities for thalamic nuclei mapping. In this work, several methodological improvements were explored to enhance SWI quality and contrast, and specifically its ability for thalamic imaging.

METHODS

High-resolution SWI was performed at 7T in healthy participants, and the following techniques were applied: (a) monitoring and retrospective correction of head motion and B₀ perturbations using integrated MR navigators, (b) segmentation and removal of venous vessels on the SWI data using vessel enhancement filtering, and (c) contrast enhancement by tuning the parameters of the SWI phase-magnitude combination. The resulting improvements were evaluated with quantitative metrics of image quality, and by comparison to anatomo-histological thalamic atlases.

RESULTS

Even with sub-millimeter motion and natural breathing, motion and field correction produced clear improvements in both magnitude and phase data quality (76% and 41%, respectively). The improvements were stronger in cases of larger motion/field deviations, mitigating the dependence of image quality on subject performance. Optimizing the SWI phase-magnitude combination yielded substantial improvements in image contrast, particularly in the thalamus, well beyond previously reported SWI results. The atlas comparisons provided compelling evidence of anatomical correspondence between SWI features and several thalamic nuclei, for example, the ventral intermediate nucleus. Vein detection performed favorably inside the thalamus, and vein removal further improved visualization.

CONCLUSION

Altogether, the proposed developments substantially improve high-resolution SWI, particularly for thalamic nuclei imaging.

Quantification in Musculoskeletal Imaging Using Computational Analysis and Machine Learning: Segmentation and Radiomics

Although still limited in clinical practice, quantitative analysis is expected to increase the value of musculoskeletal (MSK) imaging. Segmentation aims at isolating the tissues and/or regions of interest in the image and is crucial to the extraction of quantitative features such as size, signal intensity, or image texture. These features may serve to support the diagnosis and monitoring of disease. Radiomics refers to the process of extracting large amounts of features from radiologic images and combining them with clinical, biological, genetic, or any other type of complementary data to build diagnostic, prognostic, or predictive models. The advent of machine learning offers promising prospects for automatic segmentation and integration of large amounts of data. We present commonly used segmentation methods and describe the radiomics pipeline, highlighting the challenges to overcome for adoption in clinical practice. We provide some examples of applications from the MSK literature.

Pretherapeutic resting-state fMRI profiles are associated with MR signature volumes after stereotactic radiosurgical thalamotomy for essential tremor

OBJECTIVEEssential tremor (ET) is the most common movement disorder. Drug-resistant ET can benefit from standard stereotactic deep brain stimulation or radiofrequency thalamotomy or, alternatively, minimally invasive techniques, including stereotactic radiosurgery (SRS) and high-intensity focused ultrasound, at the level of the ventral intermediate nucleus (Vim). The aim of the present study was to evaluate potential correlations between pretherapeutic interconnectivity (IC), as depicted on resting-state functional MRI (rs-fMRI), and MR signature volume at 1 year after Vim SRS for tremor, to be able to potentially identify hypo- and hyperresponders based only on pretherapeutic neuroimaging data.METHODSSeventeen consecutive patients with ET were included, who benefitted from left unilateral SRS thalamotomy (SRS-T) between September 2014 and August 2015. Standard tremor assessment and rs-fMRI were acquired pretherapeutically and 1 year after SRS-T. A healthy control group was also included (n = 12). Group-level independent component analysis (ICA; only n = 17 for pretherapeutic rs-fMRI) was applied. The mean MR signature volume was 0.125 ml (median 0.063 ml, range 0.002-0.600 ml). The authors correlated baseline IC with 1-year MR signatures within all networks. A 2-sample t-test at the level of each component was first performed in two groups: group 1 (n = 8, volume < 0.063 ml) and group 2 (n = 9, volume ≥ 0.063 ml). These groups did not statistically differ by age, duration of symptoms, baseline ADL score, ADL point decrease at 1 year, time to tremor arrest, or baseline tremor score on the treated hand (TSTH; p > 0.05). An ANOVA was then performed on each component, using individual subject-level maps and continuous values of 1-year MR signatures, correlated with pretherapeutic IC.RESULTSUsing 2-sample t-tests, two networks were found to be statistically significant: network 3, including the brainstem, motor cerebellum, bilateral thalamus, and left supplementary motor area (SMA) (pFWE = 0.004, cluster size = 94), interconnected with the red nucleus (MNI -2, -22, -32); and network 9, including the brainstem, posterior insula, bilateral thalamus, and left SMA (pFWE = 0.002, cluster size = 106), interconnected with the left SMA (MNI 24, -28, 44). Higher pretherapeutic IC was associated with higher MR volumes, in a network including the anterior default-mode network and bilateral thalamus (ANOVA, pFWE = 0.004, cluster size = 73), interconnected with cerebellar lobule V (MNI -12, -70, -22). Moreover, in the same network, radiological hyporesponders presented with negative IC values.CONCLUSIONSThese findings have clinical implications for predicting MR signature volumes after SRS-T. Here, using pretherapeutic MRI and data processing without prior hypothesis, the authors showed that pretherapeutic network interconnectivity strength predicts 1-year MR signature volumes following SRS-T.

Normalization of aberrant pretherapeutic dynamic functional connectivity of extrastriate visual system in patients who underwent thalamotomy with stereotactic radiosurgery for essential tremor: a resting-state functional MRI study

OBJECTIVE

The tremor circuitry has commonly been hypothesized to be driven by one or multiple pacemakers within the cerebello-thalamo-cortical pathway, including the cerebellum, contralateral motor thalamus, and primary motor cortex. However, previous studies, using multiple methodologies, have advocated that tremor could be influenced by changes within the right extrastriate cortex, at both the structural and functional level. The purpose of this work was to evaluate the role of the extrastriate cortex in tremor generation and further arrest after left unilateral stereotactic radiosurgery thalamotomy (SRS-T).

METHODS

The authors considered 12 healthy controls (HCs, group 1); 15 patients with essential tremor (ET, right-sided, drug-resistant; group 2) before left unilateral SRS-T; and the same 15 patients (group 3) 1 year after the intervention, to account for delayed effects. Blood oxygenation level-dependent functional MRI during resting state was used to characterize the dynamic interactions of the right extrastriate cortex, comparing HC subjects against patients with ET before and 1 year after SRS-T. In particular, the authors applied coactivation pattern analysis to extract recurring whole-brain spatial patterns of brain activity over time.

RESULTS

The authors found 3 different sets of coactivating regions within the right extrastriate cortex in HCs and patients with pretherapeutic ET, reminiscent of the "cerebello-visuo-motor," "thalamo-visuo-motor" (including the targeted thalamus), and "basal ganglia and extrastriate" networks. The occurrence of the first pattern was decreased in pretherapeutic ET compared to HCs, whereas the other two patterns showed increased occurrences. This suggests a misbalance between the more prominent cerebellar circuitry and the thalamo-visuo-motor and basal ganglia networks. Multiple regression analysis showed that pretherapeutic standard tremor scores negatively correlated with the increased occurrence of the thalamo-visuo-motor network, suggesting a compensatory pathophysiological trait. Clinical improvement after SRS-T was related to changes in occurrences of the basal ganglia and extrastriate cortex circuitry, which returned to HC values after the intervention, suggesting that the dynamics of the extrastriate cortex had a role in tremor generation and further arrest after the intervention.

CONCLUSIONS

The data in this study point to a broader implication of the visual system in tremor generation, and not only through visual feedback, given its connections to the dorsal visual stream pathway and the cerebello-thalamo-cortical circuitry, with which its dynamic balance seems to be a crucial feature for reduced tremor. Furthermore, SRS-T seems to bring abnormal pretherapeutic connectivity of the extrastriate cortex to levels comparable to those of HC subjects.

Comparison of MRI-based automated segmentation methods and functional neurosurgery targeting with direct visualization of the Ventro-intermediate thalamic nucleus at 7T

The ventro-intermediate nucleus (Vim), as part of the motor thalamic nuclei, is a commonly used target in functional stereotactic neurosurgery for treatment of drug-resistant tremor. As it cannot be directly visualized on routinely used magnetic resonance imaging (MRI), its clinical targeting is performed using indirect methods. Recent literature suggests that the Vim can be directly visualized on susceptibility-weighted imaging (SWI) acquired at 7 T. Our work aims to assess the distinguishable Vim on 7 T SWI in both healthy-population and patients and, using it as a reference, to compare it with: (1) The clinical targeting, (2) The automated parcellation of thalamic subparts based on 3 T diffusion MRI (dMRI), and (3) The multi-atlas segmentation techniques. In 95.2% of the data, the manual outline was adjacent to the inferior lateral border of the dMRI-based motor-nuclei group, while in 77.8% of the involved cases, its ventral part enclosed the Guiot points. Moreover, the late MRI signature in the patients was always observed in the anterior part of the manual delineation and it overlapped with the multi-atlas outline. Overall, our study provides new insight on Vim discrimination through MRI and imply novel strategies for its automated segmentation, thereby opening new perspectives for standardizing the clinical targeting.

In-vivo probabilistic atlas of human thalamic nuclei based on diffusion- weighted magnetic resonance imaging

The thalamic nuclei are involved in many neurodegenerative diseases and therefore, their identification is of key importance in numerous clinical treatments. Automated segmentation of thalamic subparts is currently achieved by exploring diffusion-weighted magnetic resonance imaging (DW-MRI), but in absence of such data, atlas-based segmentation can be used as an alternative. Currently, there is a limited number of available digital atlases of the thalamus. Moreover, all atlases are created using a few subjects only, thus are prone to errors due to the inter-subject variability of the thalamic morphology. In this work, we present a probabilistic atlas of anatomical subparts of the thalamus built upon a relatively large dataset where the individual thalamic parcellation was done by employing a recently proposed automatic diffusion-based clustering method. Our analyses, comparing the segmentation performance between the atlas-based and the clustering method, demonstrate the ability of the provided atlas to substitute the automated diffusion-based subdivision in the individual space when the DW-MRI is not available.