Comment on “Association of peripheral immune markers with brain age and dementia risk estimated using deep learning method”

Dear Editor,

We read with great interest the recent large-scale UK Biobank study by Huang et al^[1], which employs brain age as a quantitative marker of accelerated brain aging to investigate its associations with peripheral immunity and dementia. This work represents a timely attempt to bridge systemic immunity, neurodegeneration, and dementia risk, highlighting the potential of peripheral immune markers as modifiable factors in delaying brain aging. The study adds valuable evidence to the field; however, several issues warrant further discussion to refine interpretation and guide future translational efforts.

First, the study population consisted exclusively of individuals with established dementia. While informative, this approach limits the assessment of predictive value in earlier stages. Immune dysregulation and accelerated brain aging likely emerge years before the onset of clinically overt dementia^[2]. Indeed, longitudinal cohorts suggest that alterations in innate and adaptive immunity responses may precede cognitive impairment by more than a decade^[3]. Extending analyses to prodromal stages such as mild cognitive impairment or preclinical Alzheimer’s disease (AD) would help determine whether immune–brain age coupling could function as an early warning biomarker for dementia prevention.

Second, dementia is a heterogeneous syndrome, and the underlying mechanisms of AD, vascular dementia, dementia with Lewy bodies, and frontotemporal dementia differ considerably. Collapsing these categories may mask subtype-specific patterns. For example, AD is characterized by temporoparietal cortical atrophy and amyloid-driven neuroinflammation, while frontotemporal dementia typically exhibits focal degeneration in frontal or anterior temporal regions with distinct microglial signatures. Identifying immune–neuroimaging profiles that discriminate among dementia subtypes could enhance diagnostic precision and shed light on disease-specific pathophysiology.

Third, the peripheral immune markers, including the neutrophil-to-lymphocyte ratio, platelet-to-lymphocyte ratio, lymphocyte-to-monocyte ratio, and systemic immune-inflammation index, are readily available but non-specific. They are strongly influenced by acute infection, metabolic comorbidities, lifestyle factors, such as smoking or obesity, and medications, including corticosteroids and statins. While these markers offer pragmatic utility in large-scale biobank datasets, they do not capture neuroimmune specificity. The incorporation of central nervous system-related biomarkers, such as the neurofilament light chain, glial fibrillary acidic protein, phosphorylated tau, and amyloid-β, would provide stronger mechanistic plausibility and help link peripheral immune dysregulation to central neurodegenerative processes^[4].

Finally, several methodological refinements could further strengthen the study. The analyses assumed linear relationships between immune ratios and dementia, but threshold or saturation effects are biologically plausible. Nonlinear approaches, such as restricted cubic spline modeling or generalized additive models, could reveal inflection points and clinically meaningful cutoffs. In addition, reliance on bulk leukocyte counts limits mechanistic inference, as it obscures the functional heterogeneity of immune subsets. High-resolution immunophenotyping, which captured CD4⁺ versus CD8⁺ T cells, regulatory T cells, monocyte subtypes, B cells, and natural killer cells, would clarify which specific immune compartments drive accelerated brain aging. We have been assured that the article complies with TITAN Guidelines 2025-governing declaration and use of AI^[5].

In conclusion, Huang et al provide valuable evidence linking peripheral immunity with accelerated brain aging and dementia risk. Future studies incorporating preclinical cohorts, dementia subtype differentiation, more specific inflammatory and neurodegeneration-related biomarkers, and higher-resolution immune profiling will be essential to establish the predictive and mechanistic value of immune–brain age interactions.

Ethical approval

Not applicable.

Consent

Not applicable.

Sources of funding

The authors received no financial support for this research and/or authorship of this article.

Author contributions

T.B.: writing – original draft and conceptualization; P.Z.: writing – review and editing.

Conflicts of interest disclosure

The authors declare that they have no competing interests.

Research registration unique identifying number (UIN)

Not applicable.

Guarantor

Peng Zhang.

Provenance and peer review

Not commissioned.

Data availability statement

Not applicable.