Purpose: Alzheimer’s disease (AD) is a progressive neurodegenerative disorder leading to cognitive decline and high social and economic burdens. Early diagnosis and intervention are crucial for slowing disease progression. This study evaluates the role of white matter hyperintensity (WMH) load in predicting the conversion of amnestic mild cognitive impairment (aMCI) to AD, integrating voxel-based and volumetric gray matter analysis with neuropsychological testing.

Methods: A single-center longitudinal study enrolled patients with AD, aMCI, and cognitively normal controls, matched by age (45 years or older) and sex. Participants underwent neurological assessments, cognitive tests, and brain MRI at baseline and after two years. MRI data were processed using TRACE4AD and Quantib ND, AI-based tools for assessing AD progression and brain atrophy. Trace4AD (DeepTrace Technologies) predicts the probability of AD progression within 24 months, while Quantib ND semi-automatically segmented WMH lesions.

Results: At follow-up for 6 subjects on 35 enrolled subjects, all risk predictions were confirmed. High-risk subjects had a greater WMH volume (4.7±1.0 mm³) than low-risk subjects (3.2±1.6 mm³; p=0.073). Bland-Altman analysis highlighted discrepancies between automatic and semi-automatic WMH segmentation, underscoring the need for improved reproducibility.

Conclusions: These findings suggest that WMH load, combined with AI-based volumetric and cognitive assessments, may enhance early AD prediction. Further refinements in segmentation techniques could improve diagnostic accuracy, facilitating earlier interventions for high-risk patients.