Correcting synthetic MRI contrast-weighted images using deep learning.

Synthetic magnetic resonance imaging (MRI) offers a scanning paradigm where a fast multi-contrast sequence can be used to estimate underlying quantitative tissue parameter maps, which are then used to synthesize any desirable clinical contrast by retrospectively changing scan parameters in silico. Two benefits of this approach are the reduced exam time and the ability to generate arbitrary contrasts offline. However, synthetically generated contrasts are known to deviate from the contrast of experimental scans. The reason for contrast mismatch is the necessary exclusion of some unmodeled physical effects such as partial voluming, diffusion, flow, susceptibility, magnetization transfer, and more. The inclusion of these effects in signal encoding would improve the synthetic images, but would make the quantitative imaging protocol impractical due to long scan times. Therefore, in this work, we propose a novel deep learning approach that generates a multiplicative correction term to capture unmodeled effects and correct the synthetic contrast images to better match experimental contrasts for arbitrary scan parameters. The physics inspired deep learning model implicitly accounts for some unmodeled physical effects occurring during the scan. As a proof of principle, we validate our approach on synthesizing arbitrary inversion recovery fast spin-echo scans using a commercially available 2D multi-contrast sequence. We observe that the proposed correction visually and numerically reduces the mismatch with experimentally collected contrasts compared to conventional synthetic MRI. Finally, we show results of a preliminary reader study and find that the proposed method statistically significantly improves in contrast and SNR as compared to synthetic MR images.

Indirect impact of the COVID-19 pandemic on the care and outcomes of people with MS: A combined survey and insurance claims study.

BACKGROUND: In the context of the COVID-19 pandemic, people with multiple sclerosis (pwMS) have been particularly vulnerable to adverse outcomes due to increased risk of severe infection and/or widespread disruptions in care. The CopeMS study led by The University of Texas at Austin and the MS Association of America investigates the long-term impact of the COVID-19 pandemic on healthcare access, disease modifying therapy (DMT) utilization and outcomes of pwMS. METHODS: This retrospective cohort analysis used Optum's de-identified Clinformatics® Data Mart Database (CDM), a large de-identified administrative healthcare claims database to identify pwMS who were continuously enrolled from 01/01/2019 to 12/31/2020 and assessed changes in the utilization of DMTs and healthcare services during the COVID-19 pandemic compared to the year prior. Additionally, a national survey of pwMS and healthcare providers (HCPs) was conducted to further understand the indirect impact of the pandemic on healthcare resource utilization (HCRU), outcomes and prescription patterns. RESULTS: Out of 529 pwMS in our national survey, over 47 % reported that their overall health and neurologic symptoms had deteriorated during the COVID-19 pandemic, with increased anxiety, and inability to maintain exercise habits as leading perceived causes for worsening. Survey respondents reported widespread disruption of MS-related services during the pandemic. In the Optum database, we identified 39,209 pwMS validating inclusion criteria. We observed a decrease in the utilization of MS-related services in 2020 compared to 2019. Significantly fewer pwMS had visits with their neurologist, primary care provider, physical or occupational therapist despite an increased utilization of telemedicine services. Fewer pwMS had magnetic resonance imaging (MRI) studies of the brain or spinal cord during the pandemic. Only 22.2 % of HCPs surveyed agreed that the perceived risk of more severe COVID-19 infection on a specific DMT influenced their therapeutic decisions. In the Optum database, individuals with an established diagnosis of MS prior to 2019 saw decreases in utilization of platform and moderate efficacy DMTs. In this group, those over the age of 55 saw a decrease in utilization of B-cell therapies (rate ratio 0.79, CI 0.75-0.83), whereas individuals under the age of 55 saw an increase in utilization of B-cell therapies (rate ratio 1.10, CI 1.03-1.17). We did not see any difference in rates of starting DMTs in persons diagnosed in 2019 prior to the pandemic and those diagnosed in 2020. Compared to 2019, B-cell therapies were prescribed more frequently in pwMS diagnosed in 2020 who were younger than 55 or commercially insured (rate ratio 1.35, CI 1.11-1.63). CONCLUSION: The COVID-19 pandemic was associated with perceived worsening of neurological symptoms in pwMS. Despite the expansion of telemedicine, we observed decreased access to healthcare services important to the comprehensive care of pwMS. Additionally, we observed changes in DMT utilization in pwMS during the pandemic, particularly in older adults with an established diagnosis of MS.

Cerebral blood flow is associated with markers of neurodegeneration in Huntington's disease.

INTRODUCTION: The ultimate cause of neuronal death in Huntington's disease (HD) is still uncertain. Apart from impairment in systems handling abnormal proteins, other mechanisms might contribute to neurodegeneration and progression of HD. Decreased cerebral blood flow (CBF) has been described in other neurodegenerative disorders and may play a role in HD. OBJECTIVES: To investigate CBF changes in HD gene carriers. METHODS: A group of 39 HD gene carriers (18 premanifest and 21 manifest HD) and 16 controls underwent a comprehensive clinical evaluation and a brain magnetic resonance imaging protocol that included pseudo-continuous arterial spin labeling to quantify CBF. Regions of interest (ROI) analyses were performed to compare CBF in controls vs premanifest HD vs manifest HD. Correlation analyses were performed to ascertain the relationship between CBF and clinical and biomarkers data. RESULTS: We found a decrease in CBF in bilateral caudate and putamen of patients with manifest HD in comparison with controls. CBF of premanifest HD carriers in the same ROIs was midway between controls and the HD patients, with differences not reaching statistical significance. Lower CBF in caudate and putamen was associated with worse motor symptoms, functionality, and cognitive performance. CBF was also associated with markers of neurodegeneration: higher CBF in caudate and putamen significantly correlated to higher volumes in the same ROI and to lower levels of neurofilament light chain. CONCLUSION: As CBF changes in caudate and putamen nuclei were associated with markers of neurodegeneration and with clinical outcomes, decreased CBF and oxygen supply could emerge as a relevant mechanism contributing to degeneration in HD.

Enrollment of Non-White Participants and Reporting of Race and Ethnicity in Phase III Trials of Multiple Sclerosis DMTs: A Systematic Review.

BACKGROUND AND OBJECTIVES: Black and Hispanic people with multiple sclerosis (MS) (pMS) have been found to have different disease courses or worse outcomes associated with MS compared to White pMS. They are also more likely to be negatively affected by social determinants of health, further worsening disparities in outcomes. As these disparities may affect treatment response, non-White pMS must be included in trials for greater generalizability of research and therefore more inclusive treatment plans. In this study, we aimed to evaluate how representation of non-White groups in phase III trials of approved disease-modifying therapies (DMTs) has evolved over time and how race and ethnicity are reported in medical journals and on manufacturer websites. METHODS: We conducted a systematic review of the PubMed database from 1995 to June 2020 to identify manufacturer-sponsored phase III trials for Food and Drug Administration-approved MS DMTs. We explored how race and ethnicity were reported in trial publications. Using studies where information was available, we analyzed the representation of non-White pMS over time and compared to multinational census data. We reviewed patient- and health care provider (HCP)-facing websites of available DMTs to assess the dissemination of information on racial and ethnic representation in trials. RESULTS: A total of 44 phase III trial publications were reviewed, representing 45 trials, among which 17 (37.8%) did not report race or ethnicity, 14 (31.1%) reported race and ethnicity as proportion of White participants only, and 14 (31.1%) reported 2 or more races/ethnicities. When compared to multinational census data, non-White pMS were significantly underrepresented in MS trials. Due to lack of data, trends in representation of other races and ethnicities could not be assessed. No patient- or HCP-facing DMT websites reported data on race and ethnicity in pivotal trials. Study results are available on our study dashboard. CONCLUSION: Race and ethnicity are underreported in MS DMT trial publications and race and ethnic representation are omitted from manufacturer websites. When available, data show that non-White pMS are significantly underrepresented in MS trials. The availability of this information is crucial for patients, together with their HCPs, to make informed decisions about their care.

Microglia Activation in Basal Ganglia Is a Late Event in Huntington Disease Pathophysiology.

OBJECTIVE: To define the role played by microglia in different stages of Huntington disease (HD), we used the TSPO radioligand [11C]-ER176 and PET to evaluate microglial activation in relation to neurodegeneration and in relation to the clinical features seen at premanifest and manifest stages of the disease. METHODS: This is a cross-sectional study in which 18 subjects (6 controls, 6 premanifest, and 6 manifest HD gene carriers) underwent a [11C]-ER176 PET scan and an MRI for anatomic localization. Segmentation of regions of interest (ROIs) was performed, and group differences in [11C]-ER176 binding (used to evaluate the extent of microglial activation) were assessed by the standardized uptake value ratio (SUVR). Microglial activation was correlated with ROIs volumes, disease burden, and the scores obtained in the clinical scales. As an exploratory aim, we evaluated the dynamic functions of microglia in vitro, by using induced microglia-like (iMG) cells from peripheral blood monocytes. RESULTS: Individuals with manifest HD present higher [11C]-ER176 SUVR in both globi pallidi and putamina in comparison with controls. No differences were observed when we compared premanifest HD with controls or with manifest HD. We also found a significant correlation between increased microglial activation and cumulative disease burden, and with reduced volumes. iMG from controls, premanifest HD, and manifest HD patients showed similar phagocytic capacity. CONCLUSIONS: Altogether, our data demonstrate that microglial activation is involved in HD pathophysiology and is associated with disease progression.

Automatic detection and segmentation of brain metastases on multimodal MR images with a deep convolutional neural network.

Stereotactic treatments are today the reference techniques for the irradiation of brain metastases in radiotherapy. The dose per fraction is very high, and delivered in small volumes (diameter <1 cm). As part of these treatments, effective detection and precise segmentation of lesions are imperative. Many methods based on deep-learning approaches have been developed for the automatic segmentation of gliomas, but very little for that of brain metastases. We adapted an existing 3D convolutional neural network (DeepMedic) to detect and segment brain metastases on MRI. At first, we sought to adapt the network parameters to brain metastases. We then explored the single or combined use of different MRI modalities, by evaluating network performance in terms of detection and segmentation. We also studied the interest of increasing the database with virtual patients or of using an additional database in which the active parts of the metastases are separated from the necrotic parts. Our results indicated that a deep network approach is promising for the detection and the segmentation of brain metastases on multimodal MRI.