Multimodal Magnetic Resonance Imaging Signatures of White Matter Hyperintensities: Mechanistic Insights Into Pathobiological Heterogeneity
10.5853/jos.2025.04168
- Author:
Jinyong CHUNG
;
Hyerin OH
;
Dong-Seok GWAK
;
Dong-Eog KIM
- Publication Type:Review
- From:Journal of Stroke
2026;28(1):1-28
- CountryRepublic of Korea
- Language:English
-
Abstract:
White matter hyperintensity (WMH), a common magnetic resonance imaging (MRI) marker ofcerebral small-vessel disease, is associated with chronic cerebral ischemia; however, themechanistic heterogeneity of WMH remains poorly defined. This review integrates multimodalMRI findings into a mechanism-oriented framework spanning four axes: WMH versus normalappearingwhite matter (NAWM), periventricular versus deep location, lesion core versus perilesionalpenumbra, and longitudinal evolution. Periventricular WMHs are associated with blood–brainbarrier dysfunction, interstitial fluid accumulation, and venous remodeling, whereas deep WMHsare more closely associated with impaired glymphatic/perivascular clearance and enlargedperivascular spaces, and demyelination/macromolecular compromise varying by context. Theperilesional penumbra emerges as a critical transition zone, showing distance-dependent gradientsof microstructural rarefaction, extracellular fluid expansion, perfusion deficits, and reducedvascular reactivity that extend beyond fluid-attenuated inversion recovery-defined borders andrelate to subsequent lesion growth. Longitudinal data further indicate that abnormalities indiffusion, perfusion, and vascular reserve within NAWM precede new WMHs, nominating imagingbiomarkers of progression risk. This framework supports risk stratification beyond total lesionburden, links therapeutic opportunities to mechanism (e.g., blood–brain barrier integrity, glymphaticclearance, and cerebrovascular reactivity), and motivates biologically interpretable readouts forpatient selection and treatment monitoring. Looking forward, standardized spatial classification(including fine-grained, distance-informed parcellations), harmonized penumbra definitions, andintegration of multimodal MRI with pathology will be essential to validate mechanism-specificsubtypes and translate them into scalable, clinically usable endpoints.
