DTI-Associated Blood Biomarkers in a Human Perioperative Brain Perturbation Model

Circulating biomarkers of brain perturbation are rarely evaluated against anatomically localized in vivo imaging measures, limiting biological interpretation. We tested whether perioperative changes in candidate biomarkers were associated with diffusion tensor imaging (DTI)-defined microstructural changes in a controlled human perioperative model. In a paired cohort of 75 patients undergoing left-sided encephaloduroarteriosynangiosis, we obtained matched pre- and postoperative DTI and blood samples. Six biomarkers (ubiquitin C-terminal hydrolase-L1 [UCH-L1], glial fibrillary acidic protein [GFAP], S100 calcium-binding protein B [S100B], neurofilament light chain, neuron-specific enolase, and protein tau) were quantified. Analyses included paired comparisons, region-of-interest-based cross-modal correlations, sparse canonical correlation analysis (sCCA) with bootstrap stability assessment, and patient-grouped regularized logistic regression for state discrimination. DTI showed perioperative changes in bilateral temporal-limbic regions and connected white matter, with the largest effects in left hippocampal and amygdalar territories on the operated side. UCH-L1, GFAP, and S100B showed the clearest paired increases and dominated internal perioperative-state discrimination. Blood–DTI associations were heterogeneous across biomarkers and states, and paired-change associations were not confined to these three markers. An observed three-biomarker panel discriminated perioperative state with an area under the receiver operating characteristic curve of 0.914 in the test set. Several circulating biomarkers covaried with regional DTI measures in this exploratory perioperative cohort. UCH-L1, GFAP, and S100B were the most perioperatively responsive and state-discriminative markers, but the DTI-association pattern was not specific to this trio. The findings support a structure-associated framework for evaluating blood biomarkers against regional imaging changes; they do not validate these biomarkers for traumatic brain injury, and relevance to traumatic brain injury (TBI) remains hypothesis-generating.