Peripheral serum biomarkers and MRI-derived accelerated brain aging: a cross-sectional study using deep learning-based brain age estimation

Brain age gap is an emerging neuroimaging marker that captures interindividual variation in brain structural aging. Identifying accessible peripheral biomarkers associated with accelerated brain aging may improve early risk recognition and provide insight into systemic biological processes linked to brain health. This study investigated the associations between routine serum laboratory biomarkers and MRI-derived brain age gap in a health examination cohort.

We included 852 participants with available structural MRI and routine laboratory data. Brain age was estimated from structural MRI using a deep learning model, and corrected brain age gap (CorrectedBAG) was derived to reduce age-related prediction bias. We evaluated 25 serum laboratory biomarkers spanning metabolic, immune-inflammatory, hepatic, renal, and bone/nutrition-related domains.

Multivariable linear regression was used to examine associations with CorrectedBAG, and restricted cubic spline models were applied to assess potential nonlinear dose-response relationships. Binary logistic regression was further used to evaluate accelerated brain aging defined by brain age gap thresholds. Among the 25 biomarkers examined, serum alkaline phosphatase (ALP) and neutrophil-to-lymphocyte ratio (NLR) showed independent positive associations with CorrectedBAG.

Higher ALP was associated with greater CorrectedBAG (β = 0.4073, 95% CI 0.1534–0.6613; P = 0.0017), as was higher NLR (β = 0.2558, 95% CI 0.0164–0.4952; P = 0.0365). In logistic regression analyses, ALP was consistently associated with greater odds of accelerated brain aging defined as BAG > 0 years (OR = 1.011, 95% CI 1.003–1.019; P = 0.010) and BAG > 3 years (OR = 1.013, 95% CI 1.004–1.023; P = 0.006). NLR was also associated with BAG > 3 years (OR = 1.333; P = 0.005).

Restricted cubic spline analyses supported approximately linear associations of ALP and NLR with CorrectedBAG. In this Asian health examination cohort, higher serum ALP and NLR were associated with greater MRI-derived brain age gap. These findings suggest that routinely available serum biomarkers may capture peripheral biological processes related to accelerated brain structural aging.

Further studies are warranted to clarify the temporal relationships and biological pathways underlying these associations. We are indebted to the staff of the Health Management Center and Imaging Center, Beijing Tiantan Hospital, Capital Medical University. This research was supported by Clinical and Translational Medicine Research Program. from Chinese Academy of Medical Sciences (2022-I2M-C&T-B-116), the Science Innovation.