Background/Objectives: Physical exercise reduces breast cancer (BC) risk and improves survival, yet the biological mechanisms remain incompletely understood. Exercise may modulate systemic immunity and local immune cell infiltration in the tumor microenvironment. In this systematic review and meta-analysis, we examined the effects of exercise on immune cells and immune-related markers in patients with BC. Methods: This study followed PRISMA guidelines and was prospectively registered in PROSPERO (CRD420251082444). Four databases (PubMed, Web of Science, Scopus, and Cochrane Library) were searched from inception through December 2025. Randomized controlled trials evaluating exercise interventions in patients with BC or BC survivors and reporting immune cell outcomes were included. Meta-analyses were performed on studies reporting natural killer cells, natural killer cell activity, T-cell subpopulations, and B cells. Results: A total of 18 studies involving 911 participants (539 in exercise intervention groups) were included in the systematic review, with eight studies included in meta-analyses. Exercise interventions did not show significant effects on circulating natural killer cell counts, natural killer cell activity, T-cell subpopulations (CD3⁺, CD4⁺, and CD8⁺), or B-cell levels when compared to control groups. Conclusions: Exercise does not appear to induce consistent changes in resting circulating immune cell populations in patients with BC or BC survivors, indicating that exercise is immunologically safe while potentially exerting effects beyond circulating cell counts. Further large-scale research is required.

Alterations in the gut microbiota (GM) are increasingly implicated in Alzheimer's disease (AD) pathogenesis through the microbiota-gut-brain axis (MGBA), influencing neuroinflammation, amyloid-β deposition, and tau aggregation. The inherent complexity of microbiota data necessitates artificial intelligence (AI) and machine learning (ML) to identify novel biomarkers and therapeutic targets. In this narrative review, we examined AI applications in AD-related GM research, including metataxonomic and metagenomic approaches. AI methods facilitate microbial signature identification and enterotype characterization associated with AD progression, enable multi-omics integration clarifying molecular pathways linking GM dysbiosis to neurodegeneration, support predictive modeling for early AD detection, and inform personalized MGBA-targeted interventions. Employed techniques include classical ML (e.g., random forests) for taxonomic classification, deep learning (e.g., neural networks) for complex pattern recognition, and explainable AI for identifying key bacterial taxa and metabolic functions. Significant challenges persist regarding methodological standardization, larger longitudinal cohorts, and ethical considerations in clinical applications. Future priorities include developing interpretable models, integrating longitudinal multi-omics data to establish causality, and validating AI-derived insights for precision AD prevention and treatment. AI uniquely enables high-dimensional, non-linear analyses that advance understanding of microbiome-brain interactions, though clinical translation requires rigorous validation and interdisciplinary collaboration.

Background and Objectives: The concept of the gut–muscle–brain axis encompasses the intricate, multidirectional interactions between the gut microbiota (GM), physical exercise (PE), and the central nervous system. Within this framework, gut dysbiosis has been implicated in the pathogenesis of Alzheimer’s disease (AD). Given that cognitive functions in AD appear to benefit from PE, it is plausible to hypothesize that these improvements may be partially mediated by PE-induced alterations in GM taxonomy. Therefore, the objective of this study is to evaluate the potential effects of PE in the GM and their implications for AD. Methods: A systematic review was conducted in PubMed, Web of Science and Scopus following the PRISMA guidelines up to July 2025 for preclinical controlled trials that assessed the effects of PE on the GM of AD animal models. A random-effects model meta-analysis was performed to estimate the pooled effect of PE on GM frequency or composition. This study received no external funding. Results: Eight studies were included in the systematic review (sample size, n = 126), of which two could be meta-analyzed. We found that PE significantly reduced Actinobacteria abundance (MD = −0.005%; 95% CI, −0.008 to −0.002; p = 0.001) with no statistically significant evidence of heterogeneity (I² = 89.60%, Q = 0.102, p = 0.950) or publication bias observed (Begg’s test, p = 0.296), but no significant effects were found for other phylums or genera. Conclusions: PE appears capable of modulating the GM of animal models with AD in a selective and heterogeneous manner. Further studies are needed to clarify the mechanisms by which this is possible and to determinate its impact on the pathogenesis of the disease.