MRI study maps two sequences of brain barrier leaks and thinning in cognitive impairment

Researchers have used a data-driven modeling approach to sort people with cognitive impairment into distinct groups based on the order in which two brain changes appear: leaks in the blood-brain barrier and thinning of the cortex. The retrospective multi-center study, published in Fluids and Barriers of the CNS, suggests that these two processes do not follow a single fixed sequence in everyone, but instead define at least two different disease patterns with different clinical profiles.

The team studied 216 participants spanning normal cognition, mild cognitive impairment (MCI), and dementia, recruited from two centers. Blood-brain barrier permeability was quantified as Ktrans from dynamic contrast-enhanced MRI, a technique that tracks contrast agent moving from blood into tissue, while cortical thickness was measured from structural MRI across six bilateral cortical regions. Because cross-sectional studies of barrier disruption and atrophy have produced conflicting results, likely reflecting disease heterogeneity, the researchers applied the Subtype and Stage Inference (SuStaIn) algorithm to infer subtypes and the relative ordering of imaging abnormalities from cross-sectional data.

SuStaIn identified three groups. Subtype 0, comprising 99 participants, appeared normal-appearing on these measures. Subtype 1, with 58 participants, showed an atrophy-first pattern, while Subtype 2, with 59 participants, showed a BBB-first pattern in which elevated permeability preceded cortical thinning in the model-inferred sequence. The mean age of the cohort was 70.1 years and 64.4% of participants were female.

The clinical characteristics of the two modeled subtypes differed markedly. Subtype 1 had the highest dementia prevalence at 46.6%, the poorest cognitive performance, and the greatest cortical thinning. It also carried a higher frequency of APOE ε4 carriers than Subtype 2, at 20 of 50 participants (40.0%) versus 10 of 50 (20.0%), a difference the authors report as statistically significant (p = 0.029). Subtype 2, by contrast, had the highest MCI prevalence at 50.8% and the highest Ktrans values, indicating the greatest blood-brain barrier permeability. Notably, Subtype 2 showed a lower white matter hyperintensity burden than both other groups, and cerebrospinal fluid total tau was elevated in both modeled subtypes.

The authors emphasize that SuStaIn inferred these sequences from cross-sectional snapshots, so the ordering reflects a statistical model of disease progression rather than directly observed change within individuals. They state that longitudinal studies are needed to establish whether the atrophy-first and BBB-first patterns represent true temporal sequences occurring in individual patients. The trial registrations cited (KCT0008510, KCT0008699, and KCT0003181) were all retrospectively registered.

Subject of Research: Data-driven ordering of blood-brain barrier permeability changes and cortical thinning in cognitive impairment

Article Title: SuStaIn-inferred sequences of blood-brain barrier permeability and cortical thinning in cognitive impairment: a data-driven DCE-MRI study

Article References: Park, G., Moon, W.-J., Jeon, H. J., Kwon, K. J., Noh, M.-Y., Kim, H.-J., Han, S.-H., Kim, S. H., Toga, A. W., Moon, Y., & Kim, H. (2026). SuStaIn-inferred sequences of blood-brain barrier permeability and cortical thinning in cognitive impairment: a data-driven DCE-MRI study. Fluids and Barriers of the CNS. https://doi.org/10.1186/s12987-026-00896-1

Image Credits: AI Generated

DOI: 10.1186/s12987-026-00896-1

Keywords: blood-brain barrier, cortical thinning, SuStaIn, DCE-MRI, mild cognitive impairment, dementia, Ktrans, APOE ε4, CSF tau, white matter hyperintensities, disease subtypes, neurodegeneration