DCE-MRI study on regional and Fazekas-stratified heterogeneity of blood-brain barrier leakage in white matter hyperintensities and related risk factors.

BACKGROUND: Blood-brain barrier (BBB) disruption is a key pathological mechanism of cerebral small vessel disease (CSVD). White matter hyperintensities (WMH), a major imaging feature of this disease, show unclear regional and severity-related differences in BBB leakage, along with uncertain independent risk factors. Conventional understanding that leakage increases steadily with lesion severity has not been validated with refined stratification. This study aimed to characterize the heterogeneous patterns of BBB leakage in periventricular (PWMH) and deep WMH (DWMH) stratified by Fazekas grade and identify their independent risk factors using quantitative dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) parameters. METHODS: This single-center retrospective study included 75 patients with sporadic CSVD who underwent 3.0-tesla (3.0T) DCE-MRI. WMH were classified into PWMH and DWMH locations and graded by the Fazekas scale (0/1, 2, 3). The volume transfer constant (Ktrans) and plasma volume fraction (Vp) were derived from the Extended Tofts two-compartment model to assess BBB permeability and local perfusion, respectively. Group comparisons were performed using Mann-Whitney U tests, correlations were evaluated with Spearman rank analysis, and independent predictors of Ktrans were identified via stepwise multiple linear regression, with a significance threshold of α=0.05 (two-sided). RESULTS: Ktrans was significantly higher in Fazekas grade 2 PWMH than in grade 0/1 (P=0.0014) and significantly lower in grade 3 (P<0.0001). For DWMH, Ktrans was significantly higher in grade 2 (P=0.037) and significantly lower in grade 3 (P=0.0002) relative to grade 0/1. PWMH showed significantly higher Ktrans than DWMH at identical grade 0/1 and grade 2 (both P<0.05), but not at grade 3 (P>0.05). Vp declined progressively with increasing Fazekas grade: in PWMH, Vp was significantly lower in both grade 2 (P=0.027) and grade 3 (P<0.0001) versus grade 0/1; in DWMH, only grade 3 showed significantly lower Vp than grade 0/1 (P=0.021). Spearman correlation analysis revealed that Ktrans was positively correlated with Vp in most subgroups (PWMH grade 0/1: r=0.444, P=0.014; DWMH grade 0/1: r=0.288, P=0.043), except for PWMH grade 3 (r=0.274, P=0.367). Multiple linear regression (Fazekas grades 0-2) showed that age [β=-0.526, 95% confidence interval (CI): -0.944 to -0.108, P=0.015], Vp (β=0.870, 95% CI: 0.031-1.709, P=0.043), and Fazekas grade 2 classification (β=20.970, 95% CI: 12.01-29.94, P<0.0001) were independent predictors of Ktrans in PWMH. For DWMH, type 2 diabetes mellitus (β=9.770, 95% CI: 3.870-15.670, P=0.002) and Fazekas grade 2 classification (β=6.845, 95% CI: 0.977-12.710, P=0.023) were independent positive predictors of Ktrans. CONCLUSIONS: BBB leakage in WMH presents clear regional and severity-dependent heterogeneity. Moderate lesions represent the active stage of barrier disruption. Reduced leakage values in severe lesions may mainly stem from diminished local perfusion restricting contrast agent delivery instead of restored barrier function. Periventricular and deep lesions are regulated by distinct risk factors. These results support refined risk evaluation and targeted intervention for CSVD.