Zhou, Yuxiang

Background: The potential long-term effects of Severe Acute Respiratory Syndrome Coronavirus (SARS-CoV-2) infection on the brain structure have not yet been fully elucidated. Even though existing studies have reported structural changes in the post-COVID-19 period, the results remain highly inconsistent and controversial. As such, identifying an imaging biomarker for post-COVID brains is still under investigation. This review aims to comprehensively summarize the structural MRI (sMRI) studies that focus on volumetric brain changes at least two weeks following COVID-19 infection. Methods: A systematic literature search was conducted on PubMed, SCOPUS, Web of Science, EMBASE, and Google Scholar up to 9 September 2025. Studies that utilized sMRI to assess volumetric brain changes post-COVID at greater than two weeks following infection were included. Exclusion criteria encompassed research involving pediatric or adolescent populations and imaging modalities other than sMRI. Preprints, reviews, case reports, case series and post-mortem studies were also excluded. Results: Forty-one studies satisfied the inclusion criteria and consisted of 2895 patients and 1729 healthy controls. Despite the wide variability in image acquisition protocols, data processing methods, and comorbidities between studies, multiple studies reported statistically significant volumetric reductions in the hippocampus, amygdala, thalamus, basal ganglia, nucleus accumbens and the cerebellum months to years after infection, especially in older hospitalized patients with severe COVID-19. Conclusions: The emerging literature reports long-term volume changes across various brain regions in individuals previously infected with COVID-19; however, the evidence is inconsistent. Specific imaging biomarkers following exposure to SARS-CoV-2 infection and the underlying mechanisms of these changes are yet to be identified. Future studies with harmonized imaging protocols, longitudinal designs, and integrated biomarker and clinical data are needed to define robust biomarkers and elucidate the pathophysiology of these findings.

Abstract BACKGROUND Glioblastoma (GBM) is a molecularly and phenotypically heterogeneous tumor, with regional invasion patterns driving treatment resistance and universal recurrence. In infiltrative regions, GBM cells adopt at least two predominant invasive phenotypes: Neuronal-enriched (NEU), with synaptic and neurodevelopment programs, and Glycolytic/Plurimetabolic (GPM), with heightened immune and metabolic activity. These subtypes differ in therapeutic vulnerability and microenvironmental interactions; features not captured by conventional contrast-enhanced (T1+C) MRI due to minimal blood-brain barrier (BBB) disruption. Our group sought to develop predictive models to distinguish variability of NEU/GPM subtypes within infiltrative GBM, utilizing graph convolutional networks (GCN), advanced multiparametric MRI (mpMRI), and spatially-matched image-localized biopsies with transcriptomic profiling. METHODS We analyzed 18 mpMRI features —including conventional, diffusion-weighted, and Dynamic Susceptibility Contrast (DSC) perfusion MRI— of 101 image-localized biopsies from MRI non-enhancing infiltrative margins of 64 GBM patient tumors. Features were hierarchically clustered per modality; each cluster was represented as a node in a Neo4j graph. Edges connected nodes sharing overlapping samples. Nodes were annotated with NEU or GPM subtype proportion. GCNs were trained to classify subtype-enriched nodes using graph topology and node features. RESULTS GCNs achieved high predictive performance on held-out test data (AUC = 0.95 for NEU, 0.88 for GPM; 70:30 train-test split) with similar 5-fold cross-validation performance and generalized well to an independent cohort (AUC > 0.85, n = 45). Importantly, performance was robust without relying on T1+C imaging, enabling prediction beyond BBB-disrupted regions. GCNs outperformed traditional classifiers (logistic regression, random forest). CONCLUSION Our results demonstrate that GCN-modeling of mpMRI features enables non-invasive prediction of invasive GBM subtypes at the tumor margin. Unlike convolutional neural networks (CNN), which rely on spatial continuity, GCNs leverage relational structure between heterogeneous imaging features, making them ideal for modeling sparse, non-grid data. This may support non-invasive regional vulnerability detection, surgical targeting, and improved response assessment.

AbstractINTRODUCTIONResting‐state functional magnetic resonance imaging (fMRI) graph theory may help detect subtle functional connectivity changes affecting memory prior to impairment.METHODSCognitively normal apolipoprotein E (APOE) ε4 carriers/noncarriers underwent longitudinal cognitive assessment and one‐time MRI. The relationship of left/right hippocampal connectivity and memory trajectory were compared between carriers/noncarriers.RESULTSSteepness of verbal memory decline correlated with decreased connectivity in the left hippocampus, only among APOE ε4 carriers. Right hippocampal metrics were not correlated with memory and there were no significant correlations in the noncarriers. Verbal memory decline correlated with left hippocampal volume loss for both carriers and noncarriers, with no other significant volumetric findings.DISCUSSIONFindings support early hippocampal dysfunction in intact carriers, the AD disconnection hypothesis, and left hippocampal dysfunction earlier than the right. Combining lateralized graph theoretical metrics with a sensitive measure of memory trajectory allowed for detection of early‐stage changes in APOE ε4 carriers before symptoms of mild cognitive impairment are present.Highlights Graph theory connectivity detects preclinical hippocampal changes in APOE ε4 carriers. The AD disconnection hypothesis was supported in unimpaired APOE ε4 carriers. Hippocampal dysfunction starts asymmetrically on the left.

AbstractIntroductionMultiple positron emission tomography (PET) tracers are available for amyloid imaging, posing a significant challenge to consensus interpretation and quantitative analysis. We accordingly developed and validated a deep learning model as a harmonization strategy.MethodA Residual Inception Encoder‐Decoder Neural Network was developed to harmonize images between amyloid PET image pairs made with Pittsburgh Compound‐B and florbetapir tracers. The model was trained using a dataset with 92 subjects with 10‐fold cross validation and its generalizability was further examined using an independent external dataset of 46 subjects.ResultsSignificantly stronger between‐tracer correlations (P < .001) were observed after harmonization for both global amyloid burden indices and voxel‐wise measurements in the training cohort and the external testing cohort.DiscussionWe proposed and validated a novel encoder‐decoder based deep model to harmonize amyloid PET imaging data from different tracers. Further investigation is ongoing to improve the model and apply to additional tracers.