PARE: A framework for removal of confounding effects from any distance-based dimension reduction method.

Dimension reduction tools preserving similarity and graph structure such as t-SNE and UMAP can capture complex biological patterns in high-dimensional data. However, these tools typically are not designed to separate effects of interest from unwanted effects due to confounders. We introduce the partial embedding (PARE) framework, which enables removal of confounders from any distance-based dimension reduction method. We then develop partial t-SNE and partial UMAP and apply these methods to genomic and neuroimaging data. For lower-dimensional visualization, our results show that the PARE framework can remove batch effects in single-cell sequencing data as well as separate clinical and technical variability in neuroimaging measures. We demonstrate that the PARE framework extends dimension reduction methods to highlight biological patterns of interest while effectively removing confounding effects.

Trans-Synaptic Degeneration Following Acute Optic Neuritis in Multiple Sclerosis.

OBJECTIVE: To explore longitudinal changes in brain volumetric measures and retinal layer thicknesses following acute optic neuritis (AON) in people with multiple sclerosis (PwMS), to investigate the process of trans-synaptic degeneration, and determine its clinical relevance. METHODS: PwMS were recruited within 40 days of AON onset (n = 49), and underwent baseline retinal optical coherence tomography and brain magnetic resonance imaging followed by longitudinal tracking for up to 5 years. A comparator cohort of PwMS without a recent episode of AON were similarly tracked (n = 73). Mixed-effects linear regression models were used. RESULTS: Accelerated atrophy of the occipital gray matter (GM), calcarine GM, and thalamus was seen in the AON cohort, as compared with the non-AON cohort (-0.76% vs -0.22% per year [p = 0.01] for occipital GM, -1.83% vs -0.32% per year [p = 0.008] for calcarine GM, -1.17% vs -0.67% per year [p = 0.02] for thalamus), whereas rates of whole-brain, cortical GM, non-occipital cortical GM atrophy, and T2 lesion accumulation did not differ significantly between the cohorts. In the AON cohort, greater AON-induced reduction in ganglion cell+inner plexiform layer thickness over the first year was associated with faster rates of whole-brain (r = 0.32, p = 0.04), white matter (r = 0.32, p = 0.04), and thalamic (r = 0.36, p = 0.02) atrophy over the study period. Significant relationships were identified between faster atrophy of the subcortical GM and thalamus, with worse visual function outcomes after AON. INTERPRETATION: These results provide in-vivo evidence for anterograde trans-synaptic degeneration following AON in PwMS, and suggest that trans-synaptic degeneration may be related to clinically-relevant visual outcomes. ANN NEUROL 2023;93:76-87.

Deep gray matter substructure volumes and depressive symptoms in a large multiple sclerosis cohort.

BACKGROUND: Consistent findings on underlying brain features or specific structural atrophy patterns contributing to depression in multiple sclerosis (MS) are limited. OBJECTIVE: To investigate how deep gray matter (DGM) features predict depressive symptom trajectories in MS patients. METHODS: We used data from the MS Partners Advancing Technology and Health Solutions (MS PATHS) network in which standardized patient information and outcomes are collected. We performed whole-brain segmentation using SLANT-CRUISE. We assessed if DGM structures were associated with elevated depressive symptoms over follow-up and with depressive symptom phenotypes. RESULTS: We included 3844 participants (average age: 46.05 ± 11.83 years; 72.7% female) of whom 1905 (49.5%) experienced ⩾1 periods of elevated depressive symptoms over 2.6 ± 0.9 years mean follow-up. Higher caudate, putamen, accumbens, ventral diencephalon, thalamus, and amygdala volumes were associated with lower odds of elevated depressive symptoms over follow-up (odds ratio (OR) range per 1 SD (standard deviation) increase in volume: 0.88-0.94). For example, a 1 SD increase in accumbens or caudate volume was associated with 12% or 10% respective lower odds of having a period of elevated depressive symptoms over follow-up (for accumbens: OR: 0.88; 95% confidence interval (CI): 0.83-0.93; p < 0.001; for caudate: OR: 0.90; 95% CI: 0.85-0.96; p = 0.003). CONCLUSION: Lower DGM volumes were associated with depressive symptom trajectories in MS.

Unsupervised MR harmonization by learning disentangled representations using information bottleneck theory.

In magnetic resonance (MR) imaging, a lack of standardization in acquisition often causes pulse sequence-based contrast variations in MR images from site to site, which impedes consistent measurements in automatic analyses. In this paper, we propose an unsupervised MR image harmonization approach, CALAMITI (Contrast Anatomy Learning and Analysis for MR Intensity Translation and Integration), which aims to alleviate contrast variations in multi-site MR imaging. Designed using information bottleneck theory, CALAMITI learns a globally disentangled latent space containing both anatomical and contrast information, which permits harmonization. In contrast to supervised harmonization methods, our approach does not need a sample population to be imaged across sites. Unlike traditional unsupervised harmonization approaches which often suffer from geometry shifts, CALAMITI better preserves anatomy by design. The proposed method is also able to adapt to a new testing site with a straightforward fine-tuning process. Experiments on MR images acquired from ten sites show that CALAMITI achieves superior performance compared with other harmonization approaches.

Frequency and phase correction of J-difference edited MR spectra using deep learning.

PURPOSE: To investigate whether a deep learning-based (DL) approach can be used for frequency-and-phase correction (FPC) of MEGA-edited MRS data. METHODS: Two neural networks (1 for frequency, 1 for phase) consisting of fully connected layers were trained and validated using simulated MEGA-edited MRS data. This DL-FPC was subsequently tested and compared to a conventional approach (spectral registration [SR]) and to a model-based SR implementation (mSR) using in vivo MEGA-edited MRS datasets. Additional artificial offsets were added to these datasets to further investigate performance. RESULTS: The validation showed that DL-based FPC was capable of correcting within 0.03 Hz of frequency and 0.4°of phase offset for unseen simulated data. DL-based FPC performed similarly to SR for the unmanipulated in vivo test datasets. When additional offsets were added to these datasets, the networks still performed well. However, although SR accurately corrected for smaller offsets, it often failed for larger offsets. The mSR algorithm performed well for larger offsets, which was because the model was generated from the in vivo datasets. In addition, the computation times were much shorter using DL-based FPC or mSR compared to SR for heavily distorted spectra. CONCLUSION: These results represent a proof of principle for the use of DL for preprocessing MRS data.

Effect of disease-modifying therapies on subcortical gray matter atrophy in multiple sclerosis.

BACKGROUND: The effects of disease-modifying therapies (DMTs) on region-specific brain atrophy in multiple sclerosis (MS) are unclear. OBJECTIVE: To determine the effects of higher versus lower efficacy DMTs on rates of brain substructure atrophy in MS. METHODS: A non-randomized, observational cohort of people with MS followed with annual brain magnetic resonance imaging (MRI) was evaluated retrospectively. Whole brain, subcortical gray matter (GM), cortical GM, and cerebral white matter (WM) volume fractions were obtained. DMTs were categorized as higher (DMT-H: natalizumab and rituximab) or lower (DMT-L: interferon-beta and glatiramer acetate) efficacy. Follow-up epochs were analyzed if participants had been on a DMT for ⩾6 months prior to baseline and had at least one follow-up MRI while on DMTs in the same category. RESULTS: A total of 86 DMT epochs (DMT-H: n = 32; DMT-L: n = 54) from 78 participants fulfilled the study inclusion criteria. Mean follow-up was 2.4 years. Annualized rates of thalamic (-0.15% vs -0.81%; p = 0.001) and putaminal (-0.27% vs -0.73%; p = 0.001) atrophy were slower during DMT-H compared to DMT-L epochs. These results remained significant in multivariate analyses including demographics, clinical characteristics, and T2 lesion volume. CONCLUSION: DMT-H treatment may be associated with slower rates of subcortical GM atrophy, especially of the thalamus and putamen. Thalamic and putaminal volumes are promising imaging biomarkers in MS.

Brain and retinal atrophy in African-Americans versus Caucasian-Americans with multiple sclerosis: a longitudinal study.

On average, African Americans with multiple sclerosis demonstrate higher inflammatory disease activity, faster disability accumulation, greater visual dysfunction, more pronounced brain tissue damage and higher lesion volume loads compared to Caucasian Americans with multiple sclerosis. Neurodegeneration is an important component of multiple sclerosis, which in part accounts for the clinical heterogeneity of the disease. Brain atrophy appears to be widespread, although it is becoming increasingly recognized that regional substructure atrophy may be of greater clinical relevance. Patient race (within the limitations of self-identified ancestry) is regarded as an important contributing factor. However, there is a paucity of studies examining differences in neurodegeneration and brain substructure volumes over time in African Americans relative to Caucasian American patients. Optical coherence tomography is a non-invasive and reliable tool for measuring structural retinal changes. Recent studies support its utility for tracking neurodegeneration and disease progression in vivo in multiple sclerosis. Relative to Caucasian Americans, African American patients have been found to have greater retinal structural injury in the inner retinal layers. Increased thickness of the inner nuclear layer and the presence of microcystoid macular pathology at baseline predict clinical and radiological inflammatory activity, although whether race plays a role in these changes has not been investigated. Similarly, assessment of outer retinal changes according to race in multiple sclerosis remains incompletely characterized. Twenty-two African Americans and 60 matched Caucasian Americans with multiple sclerosis were evaluated with brain MRI, and 116 African Americans and 116 matched Caucasian Americans with multiple sclerosis were monitored with optical coherence tomography over a mean duration of 4.5 years. Mixed-effects linear regression models were used in statistical analyses. Grey matter (-0.9%/year versus -0.5%: P =0.02), white matter (-0.7%/year versus -0.3%: P =0.04) and nuclear thalamic (-1.5%/year versus -0.7%/year: P =0.02) atrophy rates were approximately twice as fast in African Americans. African Americans also exhibited higher proportions of microcystoid macular pathology (12.1% versus 0.9%, P =0.001). Retinal nerve fibre layer (-1.1% versus -0.8%: P =0.02) and ganglion cell+ inner plexiform layer (-0.7%/year versus -0.4%/year: P =0.01) atrophy rates were faster in African versus Caucasian Americans. African Americans on average exhibited more rapid neurodegeneration than Caucasian Americans and had significantly faster brain and retinal tissue loss. These results corroborate the more rapid clinical progression reported to occur, in general, in African Americans with multiple sclerosis and support the need for future studies involving African Americans in order to identify individual differences in treatment responses in multiple sclerosis.

Slowly eroding lesions in multiple sclerosis.

BACKGROUND: At autopsy, 20%-40% of chronic multiple sclerosis (MS) lesions are labeled "slowly expanding" and feature myelin phagocytosis at the lesion edge. As pathological lesion classification relies on a single, terminal time point, the rate of lesion expansion cannot be directly measured. OBJECTIVE: To study long-term volume changes in individual MS lesions. METHODS: Volumes of individual lesions on proton density magnetic resonance imaging (MRI) acquired between 1992 and 2015 were measured in 22 individuals (one lesion per person). After correction for acquisition protocol, a mixed model evaluated lesion volume changes. RESULTS: The mean (standard deviation) lesion volume at baseline was 142 (82) mL, falling to 74 (51) mL after 16 (3) years. All lesions shrank over time. Change in lesion volume did not correlate with change in supratentorial brain volume ( p = 0.33). In simulations, the results could be explained by a process of slow radial expansion superimposed on substantially more rapid resorption of damaged tissue. CONCLUSION: We noted sustained radiological contraction of MS lesions, a surprising result given that fresh myelin breakdown products within chronic active lesions are observed relatively frequently at autopsy. Therefore, the primary pathological process in chronic lesions, even those described as "slowly expanding," is likely to be tissue loss.

Chemical exchange saturation transfer of the cervical spinal cord at 7 T.

High-magnetic-field (7 T) chemical exchange saturation transfer (CEST) MRI provides information on the tissue biochemical environment. Multiple sclerosis (MS) affects the entire central nervous system, including the spinal cord. Optimal CEST saturation parameters found via simulation were implemented for CEST MRI in 10 healthy controls and 10 patients with MS, and the results were examined using traditional asymmetry analysis and a Lorentzian fitting method. In addition, T1 - and T2 *-weighted images were acquired for lesion localization and the transmitted B1 (+) field was evaluated to guide imaging parameters. Distinct spectral features for all tissue types studied were found both up- and downfield from the water resonance. The z spectra in healthy subjects had the expected z spectral shape with CEST effects apparent from 2.0 to 4.5 ppm. The z spectra from patients with MS demonstrated deviations from this expected normal shape, indicating this method's sensitivity to known pathology as well as to tissues appearing normal on conventional MRI. Examination of the calculated CESTasym revealed increased asymmetry around the amide proton resonance (Δω = 3.5 ppm), but it was apparent that this measure is complicated by detail in the CEST spectrum upfield from water, which is expected to result from the nuclear Overhauser effect. The z spectra upfield (negative ppm range) were also distinct between healthy and diseased tissue, and could not be ignored, particularly when considering the conventional asymmetry analysis used to quantify the CEST effect. For all frequencies greater than +1 ppm, the Lorentzian differences (and z spectra) for lesions and normal-appearing white matter were distinct from those for healthy white matter. The increased frequency separation and signal-to-noise ratio, in concert with prolonged T1 at 7 T, resulted in signal enhancements necessary to detect subtle tissue changes not possible at lower field strengths. This study presents CEST imaging metrics that may be sensitive to the extensive and temporally varying biochemical neuropathology of MS in the spinal cord. Copyright © 2016 John Wiley & Sons, Ltd.

Optimization of 7-T chemical exchange saturation transfer parameters for validation of glycosaminoglycan and amide proton transfer of fibroglandular breast tissue.

PURPOSE: To (a) implement simulation-optimized chemical exchange saturation transfer (CEST) measurements sensitive to amide proton transfer (APT) and glycosaminoglycan (GAG) hydroxyl proton transfer effects in the human breast at 7 T and (b) determine the reliability of these techniques for evaluation of fibroglandular tissue in the healthy breast as a benchmark for future studies of pathologic findings. MATERIALS AND METHODS: All human studies were institutional review board approved, were HIPAA compliant, and included informed consent. The CEST parameters of saturation duration (25 msec) and amplitude (1 µT) were chosen on the basis of simulation-driven optimization for APT contrast enhancement with the CEST effect quantified by using residuals of a Lorentzian fit. Optimized parameters were implemented at 7 T in 10 healthy women in two separate examinations to evaluate the reliability of CEST magnetic resonance (MR) imaging measurements in the breast. CEST z-spectra were acquired over saturation offset frequencies ranging between ±40 ppm by using a quadrature unilateral breast coil. The imaging-repeat imaging reliability was assessed in terms of the intraclass correlation coefficient, which indicates the ratio of between-subject variation to total variation. RESULTS: Simulations were performed of the Bloch equations with chemical exchange-guided selection of optimal values for pulse duration and amplitude, 25 msec and 1 µT, respectively. Reliability was evaluated by using intraclass correlation coefficients (95% confidence intervals), with acceptable results: 0.963 (95% confidence interval: 0.852, 0.991) and 0.903 (95% confidence interval: 0.609, 0.976) for APT and GAG, respectively. CONCLUSION: Simulations were used to derive optimal CEST preparation parameters to elicit maximal CEST contrast enhancement in healthy fibroglandular breast tissue due to APT at 7 T. By using these parameters, reproducible values were obtained for both the amide and hydroxyl protons from CEST MR imaging at 7 T and are feasible in the human breast.

Improved diffusion tensor imaging of the optic nerve using multishot two-dimensional navigated acquisitions.

PURPOSE: A diffusion-weighted multishot echo-planar imaging approach combined with SENSE and a two-dimensional (2D) navigated motion correction was investigated as an alternative to conventional single-shot counterpart to obtain optic nerve images at higher spatial resolution with reduced artifacts. METHODS: Fifteen healthy subjects were enrolled in the study. Six of these subjects underwent a repeated acquisition at least 2 weeks after the initial scan session to address reproducibility. Both single-shot and multishot diffusion tensor imaging studies of the human optic nerve were performed with matched scan time. Effect of subject motions were corrected using 2D phase navigator during multishot image reconstruction. Tensor-derived indices from proposed multishot were compared against conventional single-shot approach. Image resolution difference, right-left optic nerve asymmetry, and test-retest reproducibility were also assessed. RESULTS: In vivo results of acquired multishot images and quantitative maps of diffusion properties of the optic nerve showed significantly reduced image artifacts (e.g., distortions and blurring), and the derived diffusion indices were comparable to those from other studies. Single-shot scans presented larger variability between right and left optic nerves than multishot scans. Multishot scans also presented smaller variations across scans at different time points when compared with single-shot counterparts. CONCLUSION: The multishot technique has considerable potential for providing improved information on optic nerve pathology and may also be translated to higher fields.

Image harmonization improves consistency of intra-rater delineations of MS lesions in heterogeneous MRI.

Clinical magnetic resonance images (MRIs) lack a standard intensity scale due to differences in scanner hardware and the pulse sequences used to acquire the images. When MRIs are used for quantification, as in the evaluation of white matter lesions (WMLs) in multiple sclerosis, this lack of intensity standardization becomes a critical problem affecting both the staging and tracking of the disease and its treatment. This paper presents a study of harmonization on WML segmentation consistency, which is evaluated using an object detection classification scheme that incorporates manual delineations from both the original and harmonized MRIs. A cohort of ten people scanned on two different imaging platforms was studied. An expert rater, blinded to the image source, manually delineated WMLs on images from both scanners before and after harmonization. It was found that there is closer agreement in both global and per-lesion WML volume and spatial distribution after harmonization, demonstrating the importance of image harmonization prior to the creation of manual delineations. These results could lead to better truth models in both the development and evaluation of automated lesion segmentation algorithms.

Development of Medical Imaging Data Standardization for Imaging-Based Observational Research: OMOP Common Data Model Extension.

The rapid growth of artificial intelligence (AI) and deep learning techniques require access to large inter-institutional cohorts of data to enable the development of robust models, e.g., targeting the identification of disease biomarkers and quantifying disease progression and treatment efficacy. The Observational Medical Outcomes Partnership Common Data Model (OMOP CDM) has been designed to accommodate a harmonized representation of observational healthcare data. This study proposes the Medical Imaging CDM (MI-CDM) extension, adding two new tables and two vocabularies to the OMOP CDM to address the structural and semantic requirements to support imaging research. The tables provide the capabilities of linking DICOM data sources as well as tracking the provenance of imaging features derived from those images. The implementation of the extension enables phenotype definitions using imaging features and expanding standardized computable imaging biomarkers. This proposal offers a comprehensive and unified approach for conducting imaging research and outcome studies utilizing imaging features.

Bile acid metabolites predict multiple sclerosis progression and supplementation is safe in progressive disease.

Background: Bile acid metabolism is altered in multiple sclerosis (MS) and tauroursodeoxycholic acid (TUDCA) supplementation ameliorated disease in mouse models of MS. Methods: Global metabolomics was performed in an observational cohort of people with MS followed by pathway analysis to examine relationships between baseline metabolite levels and subsequent brain and retinal atrophy. A double-blind, placebo-controlled trial, was completed in people with progressive MS (PMS), randomized to receive either TUDCA (2g daily) or placebo for 16 weeks. Participants were followed with serial clinical and laboratory assessments. Primary outcomes were safety and tolerability of TUDCA, and exploratory outcomes included changes in clinical, laboratory and gut microbiome parameters. Results: In the observational cohort, higher primary bile acid levels at baseline predicted slower whole brain, brain substructure and specific retinal layer atrophy. In the clinical trial, 47 participants were included in our analyses (21 in placebo arm, 26 in TUDCA arm). Adverse events did not significantly differ between arms (p=0.77). The TUDCA arm demonstrated increased serum levels of multiple bile acids. No significant differences were noted in clinical or fluid biomarker outcomes. Central memory CD4+ and Th1/17 cells decreased, while CD4+ naïve cells increased in the TUDCA arm compared to placebo. Changes in the composition and function of gut microbiota were also noted in the TUDCA arm compared to placebo. Conclusion: Bile acid metabolism in MS is linked with brain and retinal atrophy. TUDCA supplementation in PMS is safe, tolerable and has measurable biological effects that warrant further evaluation in larger trials with a longer treatment duration.

Synaptic Protein Loss in Extracellular Vesicles Reflects Brain and Retinal Atrophy in People With Multiple Sclerosis.

OBJECTIVES: To assess whether the rate of change in synaptic proteins isolated from neuronally enriched extracellular vesicles (NEVs) is associated with brain and retinal atrophy in people with multiple sclerosis (MS). METHODS: People with MS were followed with serial blood draws, MRI (MRI), and optical coherence tomography (OCT) scans. NEVs were immunocaptured from plasma, and synaptopodin and synaptophysin proteins were measured using ELISA. Subject-specific rates of change in synaptic proteins, as well as brain and retinal atrophy, were determined and correlated. RESULTS: A total of 50 people with MS were included, 46 of whom had MRI and 45 had OCT serially. The rate of change in NEV synaptopodin was associated with whole brain (rho = 0.31; p = 0.04), cortical gray matter (rho = 0.34; p = 0.03), peripapillary retinal nerve fiber layer (rho = 0.37; p = 0.01), and ganglion cell/inner plexiform layer (rho = 0.41; p = 0.006) atrophy. The rate of change in NEV synaptophysin was also correlated with whole brain (rho = 0.31; p = 0.04) and cortical gray matter (rho = 0.31; p = 0.049) atrophy. DISCUSSION: NEV-derived synaptic proteins likely reflect neurodegeneration and may provide additional circulating biomarkers for disease progression in MS.

Amide proton transfer imaging of the breast at 3 T: establishing reproducibility and possible feasibility assessing chemotherapy response.

Chemical exchange saturation transfer imaging can generate contrast that is sensitive to amide protons associated with proteins and peptides (termed amide proton transfer, APT). In breast cancer, APT contrast may report on underlying changes in microstructural tissue composition. However, to date, there have been no developments or applications of APT chemical exchange saturation transfer to breast cancer. As a result, the aims of this study were to (i) experimentally explore optimal scan parameters for breast chemical exchange saturation transfer near the amide resonance at 3 T, (ii) establish the reliability of APT imaging of healthy fibroglandular tissue, and (iii) demonstrate preliminary results on APT changes in locally advanced breast cancer observed during the course of neoadjuvant chemotherapy. Chemical exchange saturation transfer measurements were experimentally optimized on cross-linked bovine serum albumin phantoms, and the reliability of APT imaging was assessed in 10 women with no history of breast disease. The mean difference between test-retest APT values was not significantly different from zero, and the individual difference values were not dependent on the average APT value. The 95% confidence interval limits were ±0.70% (α = 0.05), and the repeatability was 1.91. APT measurements were also performed in three women before and after one cycle of chemotherapy. Following therapy, APT increased in the one patient with progressive disease and decreased in the two patients with a partial or complete response. Together, these results suggest that APT imaging may report on treatment response in these patients.

HACA3: A unified approach for multi-site MR image harmonization.

The lack of standardization and consistency of acquisition is a prominent issue in magnetic resonance (MR) imaging. This often causes undesired contrast variations in the acquired images due to differences in hardware and acquisition parameters. In recent years, image synthesis-based MR harmonization with disentanglement has been proposed to compensate for the undesired contrast variations. The general idea is to disentangle anatomy and contrast information from MR images to achieve cross-site harmonization. Despite the success of existing methods, we argue that major improvements can be made from three aspects. First, most existing methods are built upon the assumption that multi-contrast MR images of the same subject share the same anatomy. This assumption is questionable, since different MR contrasts are specialized to highlight different anatomical features. Second, these methods often require a fixed set of MR contrasts for training (e.g., both T1-weighted and T2-weighted images), limiting their applicability. Lastly, existing methods are generally sensitive to imaging artifacts. In this paper, we present Harmonization with Attention-based Contrast, Anatomy, and Artifact Awareness (HACA3), a novel approach to address these three issues. HACA3 incorporates an anatomy fusion module that accounts for the inherent anatomical differences between MR contrasts. Furthermore, HACA3 can be trained and applied to any combination of MR contrasts and is robust to imaging artifacts. HACA3 is developed and evaluated on diverse MR datasets acquired from 21 sites with varying field strengths, scanner platforms, and acquisition protocols. Experiments show that HACA3 achieves state-of-the-art harmonization performance under multiple image quality metrics. We also demonstrate the versatility and potential clinical impact of HACA3 on downstream tasks including white matter lesion segmentation for people with multiple sclerosis and longitudinal volumetric analyses for normal aging subjects. Code is available at https://github.com/lianruizuo/haca3.

Intersite brain MRI volumetric biases persist even in a harmonized multisubject study of multiple sclerosis.

BACKGROUND AND PURPOSE: Multicenter study designs involving a variety of MRI scanners have become increasingly common. However, these present the issue of biases in image-based measures due to scanner or site differences. To assess these biases, we imaged 11 volunteers with multiple sclerosis (MS) with scan and rescan data at four sites. METHODS: Images were acquired on Siemens or Philips scanners at 3 Tesla. Automated white matter lesion detection and whole-brain, gray and white matter, and thalamic volumetry were performed, as well as expert manual delineations of T1 magnetization-prepared rapid acquisition gradient echo and T2 fluid-attenuated inversion recovery lesions. Random-effect and permutation-based nonparametric modeling was performed to assess differences in estimated volumes within and across sites. RESULTS: Random-effect modeling demonstrated model assumption violations for most comparisons of interest. Nonparametric modeling indicated that site explained >50% of the variation for most estimated volumes. This expanded to >75% when data from both Siemens and Philips scanners were included. Permutation tests revealed significant differences between average inter- and intrasite differences in most estimated brain volumes (P < .05). The automatic activation of spine coil elements during some acquisitions resulted in a shading artifact in these images. Permutation tests revealed significant differences between thalamic volume measurements from acquisitions with and without this artifact. CONCLUSION: Differences in brain volumetry persisted across MR scanners despite protocol harmonization. These differences were not well explained by variance component modeling; however, statistical innovations for mitigating intersite differences show promise in reducing biases in multicenter studies of MS.

Bi-directional Synthesis of Pre- and Post-contrast MRI via Guided Feature Disentanglement.

Magnetic resonance imaging (MRI) with gadolinium contrast is widely used for tissue enhancement and better identification of active lesions and tumors. Recent studies have shown that gadolinium deposition can accumulate in tissues including the brain, which raises safety concerns. Prior works have tried to synthesize post-contrast T1-weighted MRIs from pre-contrast MRIs to avoid the use of gadolinium. However, contrast and image representations are often entangled during the synthesis process, resulting in synthetic post-contrast MRIs with undesirable contrast enhancements. Moreover, the synthesis of pre-contrast MRIs from post-contrast MRIs which can be useful for volumetric analysis is rarely investigated in the literature. To tackle pre- and post- contrast MRI synthesis, we propose a BI-directional Contrast Enhancement Prediction and Synthesis (BICEPS) network that enables disentanglement of contrast and image representations via a bi-directional image-to-image translation(I2I)model. Our proposed model can perform both pre-to-post and post-to-pre contrast synthesis, and provides an interpretable synthesis process by predicting contrast enhancement maps from the learned contrast embedding. Extensive experiments on a multiple sclerosis dataset demonstrate the feasibility of applying our bidirectional synthesis and show that BICEPS outperforms current methods.

Autoencoder based self-supervised test-time adaptation for medical image analysis.

Deep neural networks have been successfully applied to medical image analysis tasks like segmentation and synthesis. However, even if a network is trained on a large dataset from the source domain, its performance on unseen test domains is not guaranteed. The performance drop on data obtained differently from the network's training data is a major problem (known as domain shift) in deploying deep learning in clinical practice. Existing work focuses on retraining the model with data from the test domain, or harmonizing the test domain's data to the network training data. A common practice is to distribute a carefully-trained model to multiple users (e.g., clinical centers), and then each user uses the model to process their own data, which may have a domain shift (e.g., varying imaging parameters and machines). However, the lack of availability of the source training data and the cost of training a new model often prevents the use of known methods to solve user-specific domain shifts. Here, we ask whether we can design a model that, once distributed to users, can quickly adapt itself to each new site without expensive retraining or access to the source training data? In this paper, we propose a model that can adapt based on a single test subject during inference. The model consists of three parts, which are all neural networks: a task model (T) which performs the image analysis task like segmentation; a set of autoencoders (AEs); and a set of adaptors (As). The task model and autoencoders are trained on the source dataset and can be computationally expensive. In the deployment stage, the adaptors are trained to transform the test image and its features to minimize the domain shift as measured by the autoencoders' reconstruction loss. Only the adaptors are optimized during the testing stage with a single test subject thus is computationally efficient. The method was validated on both retinal optical coherence tomography (OCT) image segmentation and magnetic resonance imaging (MRI) T1-weighted to T2-weighted image synthesis. Our method, with its short optimization time for the adaptors (10 iterations on a single test subject) and its additional required disk space for the autoencoders (around 15 MB), can achieve significant performance improvement. Our code is publicly available at: https://github.com/YufanHe/self-domain-adapted-network.