MRI-derived global small vessel disease burden serves as markers of hippocampal sclerosis and clinical stage across the probable Alzheimer's disease continuum

Ting-Bin Chen^1*Ya-Che Chen²Hsin-Chieh Chen^3*

• ¹Kuang Tien General Hospital, Taichung City, Taiwan
• ²Department of Radiology, Tri-Service General Hospital, National Defense Medical University, Taipei city, Taiwan
• ³Department of Medical Equipment Development and Application, Hungkuang University, Taichung City, Taiwan

Abstract Introduction Cerebral small vessel disease (SVD) contributes to cognitive decline and hippocampal sclerosis (HS), yet its role across the Alzheimer's disease (AD) continuum remains incompletely understood. We aimed to determine whether composite MRI-based SVD scores serve as markers of cognitive impairment and HS in cognitively unimpaired (CU) individuals, patients with mild cognitive impairment (MCI), and those with probable AD dementia. Methods We retrospectively analyzed 200 participants (24 CUs, 34 MCI, 142 AD) from the dementia registry at Kuang Tien General Hospital (January 2024–June 2025). SVD burden was quantified using three composite imaging scores: global SVD, cerebral amyloid angiopathy (CAA)-SVD, and hypertensive arteriopathy (HA)-SVD. Associations with cognitive performance, clinical staging, and HS were examined using multivariable regression models. Results Global SVD and CAA-SVD scores correlated with Cognitive Abilities Screening Instrument (CASI) total and domain subscores, Clinical Dementia Rating (CDR) global score, and CDR sum of boxes (CDR-SB). Notably, only the global SVD score remained independently associated with both CDR-SB and HS after adjustment for relevant confounders. Discussion MRI-derived global SVD burden, reflecting the combined effects of CAA and HA, is strongly associated with cognition, clinical staging, and HS across the probable AD continuum, supporting the global SVD score as a clinically useful biomarker of vascular contributions. Because MCI/AD diagnoses were based on clinical criteria without confirmation using cerebrospinal fluid or positive positron emission tomography biomarkers, potential misclassification may exist; findings should be interpreted with caution.