Disentangling MRI-derived blood–brain barrier leakage into vascular permeability and surface area

Abstract

In neurovascular and neurodegenerative diseases, blood–brain barrier (BBB) disruption can be subtle and diffuse, and MRI-derived measures are confounded by microvascular architecture. Therefore, we aimed to disentangle the MRI-derived blood–brain barrier leakage rate (Ki) into intrinsic vascular permeability (P) and surface area (S). In this cross-sectional study, 45 patients with cSVD and 26 elderly controls underwent dynamic contrast-enhanced (DCE) MRI to determine Ki and vessel architecture imaging (VAI) to estimate S. These measures were used to calculate P. Compared to normal-appearing white matter (NAWM), Ki was higher in WMH (β = 0.297, p < 0.001) and deep gray matter (DGM; β = 0.210, p = 0.007). P was higher in WMH (β = 0.410, p < 0.001) and similar between NAWM and DGM. S was smaller in WMH (β = −0.467, p < 0.001) and larger in DGM (β = 0.540, p < 0.001). Ki (β = 0.289, p = 0.007) and P (β = 0.249, p = 0.021) were only higher in patients than controls in concentric shells surrounding WMH. S was smaller in WMH of patients than controls (β = −0.085, p = 0.049). With age, Ki (β = 0.260, p = 0.008) and P increased (β = 0.273, p < 0.001), while S decreased (β = −0.079, p = 0.025). Disentangling P and S from the BBB leakage rate provides two biologically more specific and interpretable measures, which have counteracting effects on the BBB leakage in aging and disease.