Highlights • Deep learning improved thalamus segmentation in multiple sclerosis brain scans.• Atlas based methods overestimated thalamus volume despite spatial overlap.• Voxel overlap and volume accuracy diverged across segmentation tools.• Quantitative magnetic resonance maps modestly improved disability associations.

Multiple sclerosis (MS) is a chronic inflammatory neurodegenerative disorder that typically affects young adults and is primarily characterized by demyelinating lesions in the central nervous system (CNS). According to the Revised McDonald Criteria, the clinical diagnosis of MS can be established based on a combination of clinical observations, the presence of focal lesions in at least two distinct CNS areas on magnetic resonance imaging (MRI) and the detection of specific oligoclonal bands in the cerebrospinal fluid. Conventional MRI remains a cornerstone of MS diagnosis and disease monitoring, providing high-resolution assessments of lesion burden and brain atrophy. In addition, advanced MRI methods are increasingly applied in research settings to probe myelin integrity, iron deposition, and biochemical changes, with the potential to complement established diagnostic workflows in the future. Despite remarkable advances in the management of MS over the past two decades, complex differential diagnoses and the lack of effective imaging tools for therapy monitoring remain major obstacles, thus channeling the development of innovative molecular imaging probes that can be harnessed in clinical practice. Indeed, positron emission tomography (PET) has a significant potential to advance the contemporary diagnosis and management of MS. Given the solid body of evidence implicating myelin dysfunction in the pathophysiology of MS, myelin-targeted imaging probes have been developed, and are currently under clinical evaluation for MS diagnosis and therapy monitoring. In parallel, ligands for the 18 kDa translocator protein (TSPO) and the cannabinoid receptor type 2 (CB2R) have been employed to capture neuroinflammatory processes by visualizing microglial activation, while other tracers allow the assessment of synaptic integrity across various disease stages of MS. Further, PET probes have been employed to delineate the role of activated microglia and facilitate the assessment of synaptic dysfunction across all disease stages of MS. This review discusses the challenges and opportunities of translational molecular imaging by highlighting key molecular concepts that are currently leveraged for diagnostic imaging, patient stratification, therapy monitoring and drug development in MS. Moreover, we shed light on potential future developments that hold promise to advance our understanding of MS pathophysiology, with the ultimate goal to provide the best possible patient care for every individual MS patient.

Background and Objectives Progression independent of relapse activity (PIRA) is associated with worse outcomes in people with multiple sclerosis (pwMS). Although previous research has linked PIRA to accelerated brain and spinal cord atrophy and compartmentalized chronic inflammation, the role of white matter (WM) tract degeneration remains unclear. This study aimed to explore the relationship between PIRA and the integrity of major WM tracts using diffusion tensor imaging (DTI). Methods A cohort of 258 pwMS was stratified based on the presence or absence of PIRA over a 4-year follow-up period. At the end of follow-up, DTI metrics were compared between groups using propensity score–weighted linear regression models to account for potential confounders. Results PwMS with ≥1 PIRA event (n = 39) exhibited significant reductions in fractional anisotropy and increases in radial, axial, and mean diffusivity within the corpus callosum and motor tracts (false discovery rate–adjusted p ≤ 0.04) compared with those without PIRA, indicating more pronounced WM damage. Discussion Our findings highlight an association between PIRA and microstructural damage in key WM tracts. The observed DTI changes likely reflect processes such as Wallerian degeneration and contribute to the growing evidence linking PIRA to neurodegeneration.

Remyelination of cortical lesions in people with multiple sclerosis (pwMS) has been shown to be extensive. In this work, we aimed to assess whether postmortem quantitative MRI (qMRI) can help detect those areas. We imaged six fixed whole brains of deceased pwMS by 3T‐MRI using magnetization transfer ratio (MTR, 570 μm isotropic), myelin water fraction (MWF, 1000 μm isotropic), quantitative T1 (qT1, 670 μm isotropic), quantitative susceptibility mapping (QSM, 330 μm isotropic) and radial diffusivity (RD, 1300 or 1400 μm isotropic) maps. Immunohistochemistry for myelin proteins was performed in 129 tissue blocks including the cortex and enabled the detection of cortical demyelination (DM), cortical remyelination (RM), and normal‐appearing cortex (NAC). We identified 25 DM, 25 RM, and for each of these areas, a corresponding NAC near the lesion. Wilcoxon paired tests showed that: (a) qT1 and RD were higher and QSM lower in DM versus NAC (all p < 0.001), whereas RD was higher and QSM lower in RM versus NAC (p = 0.048 and p < 0.01 respectively); (b) mean qT1 in RM did not differ from mean qT1 in NAC (p = 0.074); (c) MWF and MTR were not different between DM and RM. We compared the delta between DM versus NAC (∆DM) and the delta between RM versus NAC (∆RM) using a Mann–Whitney test, in which RM showed a partial recovery of qT1 only (∆qT1 DM > ∆qT1 RM, p = 0.045). Mixed‐effect models confirmed the findings obtained using univariate analyses. qT1 and QSM, but not RD, correlated with MBP intensity (r = −0.28, p < 0.01 and r = 0.29, p < 0.01 respectively). A Bonferroni correction was performed for multiple testing. Our data show that qT1 is altered in demyelinated but not in remyelinated cortical areas, while QSM and RD are affected by any cortical abnormalities. Accordingly, qT1 might be considered a potential imaging biomarker of cortical RM.

We report biomarker treatment effects in the GRADUATE I and II phase 3 studies of gantenerumab in early Alzheimer's disease (AD).

Traditional Diffusion Tensor Imaging (DTI) metrics are affected by crossing fibers and lesions. Most of the previous tractometry works use the single diffusion tensor, which leads to limited sensitivity and challenging interpretation of the results in crossing fiber regions. In this work, we propose a tractometry pipeline that combines white matter tractography with multi-tensor fixel-based metrics. These multi-tensors are estimated using the stable, accurate and robust to noise Multi-Resolution Discrete Search method (MRDS). The spatial coherence of the multi-tensor field estimated with MRDS, which includes up to three anisotropic and one isotropic tensors, is tractography-regularized using the Track Orientation Density Imaging method. Our end-to-end tractometry pipeline goes from raw data to track-specific multi-tensor-metrics tract profiles that are robust to noise and crossing fibers. A comprehensive evaluation conducted in a phantom simulating healthy and damaged tissue with the standard model, as well as in a healthy cohort of 20 individuals scanned along 5 time points, demonstrates the advantages of using multi-tensor metrics over traditional single-tensor metrics in tractometry. Qualitative assessment in a cohort of patients with relapsing-remitting multiple sclerosis reveals that the pipeline effectively detects white matter anomalies in the presence of crossing fibers and lesions.