Integrated DTI-ALPS and Radiomics for Alzheimer's Disease Diagnosis and Cognitive Function Evaluation

Motivation: Early and aggressive non-invasive diagnosis of Alzheimer's disease (AD) is still hard and critical. Goal(s): To evaluate alterations in glymphatic system (GS) in AD and explore the effectiveness of combining diffusion tensor image analysis along the perivascular space (DTI-ALPS) and radiomics for diagnosing AD. Approach: We developed ALPS-radiomics models to assess classification performance in diagnosing AD. Results: The ALPS index in patients with AD was lower than that in the Healthy Controls (HCs), and the former was correlated with MMSE score. Classification models combining ALPS index and radiomics showed good prediction performance for AD diagnosis in two independent cohorts. Impact: DTI-ALPS shows promise as a valuable tool for accessing GS alterations in AD. Integrated ALPS index and radiomics can improve diagnostic capabilities and assess cognitive function in patients with AD.