Perivascular diffusion and white matter free water improve identification of cognitive impairment in cerebral small vessel disease

Abstract

Inroduction Cerebral small vessel disease contributes substantially to cognitive decline in older adults, but conventional structural magnetic resonance imaging markers mainly reflect accumulated lesion burden and do not directly characterize diffusion-derived alterations in brain fluid homeostasis. We investigated whether combining diffusion imaging along the perivascular space and white matter free-water imaging improves identification of Montreal Cognitive Assessment (MoCA)-defined cognitive impairment in elderly patients with cerebral small vessel disease. Methods This prospective study included 100 patients with cerebral small vessel disease and 50 healthy controls. According to MoCA scores, patients were classified as cognitively normal (n = 42) or cognitively impaired (n = 58). All participants underwent 3.0-T magnetic resonance imaging, including conventional structural imaging and diffusion tensor imaging. Conventional structural magnetic resonance imaging markers were evaluated on anatomical imaging, whereas diffusion-derived metrics were used to assess perivascular diffusion and free-water burden. Group comparisons, correlation analyses, multivariate logistic regression, and receiver operating characteristic analyses were performed. Results Analysis along the perivascular space decreased progressively from healthy controls to cognitively normal patients and cognitively impaired patients, whereas white matter free water increased in the opposite direction (all p < 0.001). Lower ALPS-Index values and higher white matter free-water values were both associated with lower MoCA scores. The clinical information model showed moderate discrimination for MoCA-defined cognitive impairment (area under the curve = 0.764, 95% confidence interval: 0.669–0.859). After incorporating age, sex, education, hypertension, and diabetes as covariates, the traditional cerebral small vessel disease marker model achieved an area under the curve of 0.819 (95% confidence interval: 0.731–0.908), whereas the glymphatic-edema signature combining ALPS-Index and white matter free water achieved better discriminative performance (area under the curve = 0.928, 95% confidence interval: 0.875–0.981). The combined imaging model incorporating conventional structural markers, ALPS-Index, and white matter free water achieved the highest discriminative performance (area under the curve = 0.950, 95% confidence interval: 0.906–0.993). Discussion Perivascular diffusion imaging and white matter free water capture complementary aspects of altered perivascular diffusion and extracellular free-water accumulation in cerebral small vessel disease and improve identification of MoCA-defined cognitive impairment beyond conventional structural magnetic resonance imaging markers.