Abstract / Summary
White matter hyperintensities on brain magnetic resonance imaging are established markers of cerebral small vessel disease and typically associate inversely with brain volume in the general population, reflecting irreversible tissue loss. In Cerebral Autosomal Dominant Arteriopathy with Subcortical Infarcts and Leukoencephalopathy, a monogenic small vessel disease caused by NOTCH3 cysteine-altering mutations, white matter hyperintensities may instead reflect tissue water accumulation and perivascular space dilation rather than net structural damage. Whether this distinct pathophysiological relationship is detectable at population level among carriers of NOTCH3 cysteine-altering variants without clinically manifest disease remains unknown. We conducted a cross-sectional neuroimaging study using UK Biobank data acquired between 2014 and 2023. Among 44,943 participants with complete brain magnetic resonance imaging data (mean age 64 years +/- 8 years), 96 carried heterozygous NOTCH3 cysteine-altering variants within epidermal growth factor-like repeat domains 7 to 34 and 44,847 were non-carriers. White matter hyperintensity volume was automatically quantified and normalized to total intracranial volume. Brain parenchymal fraction was defined as combined gray and white matter volumes divided by total intracranial volume. Lacunes were manually assessed in all carriers. Linear regression and generalized additive models examined the association between white matter hyperintensity volume and brain parenchymal fraction stratified by genotype, adjusting for age, sex, lacune count, and vascular risk factors. NOTCH3 variant carriers showed markedly higher normalized white matter hyperintensity volume than non-carriers (0.83% versus 0.36%; P<10^-8) but only marginally higher brain parenchymal fraction (81.1% versus 80.5%; P=0.03). In non-carriers, white matter hyperintensity volume was inversely associated with brain parenchymal fraction (beta=-0.116% per 1% increase; P=3.10^-6). In contrast, NOTCH3 variant carriers demonstrated a positive association (beta=+0.51%; P=0.019), strengthened after adjusting for lacune count (beta=+0.61%; P=0.006). A significant genotype-specific interaction confirmed this divergence (P=0.036), persisting across age groups and after vascular risk factor adjustment. NOTCH3 cysteine-altering variant carriers identified at population level exhibit a fundamentally distinct white matter hyperintensity-brain volume relationship compared with non-carriers, consistent with lesions reflecting tissue water accumulation rather than irreversible structural loss at this early, predominantly preclinical disease stage. These findings establish that genetic context modifies neuroimaging biomarker interpretation in cerebral small vessel disease and underscore the need for genotype-aware approaches when evaluating white matter hyperintensities as markers of disease severity or progression.