Linked alterations in structure and autoimmunity biomarkers in remote mild-to-moderate TBI: A multi-modal brain imaging study.

Background: There is significant heterogeneity in the recovery course from mild-to-moderate traumatic brain injury (mmTBI), with many individuals reporting cognitive symptoms during the chronic phase. Although blood-based biomarkers have been identified as a marker of injury severity in the acute phase, the relevance of candidate biomarkers in chronic mmTBI is less clear. Establishing links between blood-based biomarkers, neuroimaging, and cognitive performance is necessary to differentiate subphenotypes in chronic TBI and improve prognostic models.

Methods: Sixty Veterans and non-Veterans with mmTBI completed cognitive testing (WAIS-IV), MRI, and blood collection for blood-based CNS biomarker assessment for cross-sectional comparison. A data fusion technique (Linked Independent Component Analysis [LICA]) was used to simultaneously model structural variability across T1-weighted and diffusion-weighted MRI modalities. Blood serum samples were assayed using an ultrasensitive immunoassay using digital array technology to measure GFAP, NFL, total tau, and UCH-L1 protein levels. A correlation matrix was used to identify which LICA-derived MRI components were associated with (1) a blood-based biomarker, (2) a WAIS-IV index score and (3) clinical characteristics of TBI, using an effect-size cut-off.

Results: LICA-derived Component 4 was associated with a biomarker (UCH-L1), reduced processing speed, and the total number of TBIs. This component was characterized by increased mean diffusivity along the ventral surface of the frontal lobe and decreased fractional anisotropy in bilateral corticospinal tracts and cerebral peduncles.

Conclusions: Our findings contribute to the larger body of literature examining the utility of biomarkers for chronic TBI and underscore the importance of examining heterogeneity within this population.