Amyloid status imputed from a multimodal classifier including structural MRI distinguishes progressors from...

Area of Science:

• Neurology
• Neuroimaging
• Biomarkers

Background:

• Mild Alzheimer's disease (AD) patients lacking significant amyloid-beta (Aβ) pathology can confound clinical trial results due to unpredictable progression.
• Accurate stratification of participants is crucial for measuring treatment efficacy in AD clinical trials.

Purpose of the Study:

• To evaluate the utility of a prediction model using structural magnetic resonance imaging (MRI) to impute Aβ status in mild AD patients.
• To assess the clinical trajectories of mild AD patients stratified by imputed Aβ status over 18 months.

Main Methods:

• A prediction model combining structural MRI, demographics, and clinical data was developed.
• Aβ status was imputed for a placebo-treated mild AD sub-cohort (N=385) from global phase 3 trials.
• Clinical trajectories were analyzed over 18 months using a mixed-model repeated measures framework, stratified by imputed Aβ status.

Main Results:

• The imputed Aβ-positive cohort exhibited a more rapid decline in cognitive measures (ADAS-Cog14, MMSE) and functional measures (ADCS-iADL) compared to the undifferentiated population.
• This suggests that Aβ pathology influences the rate of clinical decline in mild AD.

Conclusions:

• Imputing Aβ status from MRI scans can serve as a valuable screening tool for mild AD subjects in global clinical trials, particularly when direct amyloid measurement is unavailable.
• This approach can improve participant stratification and potentially enhance the sensitivity of clinical trials to detect treatment effects.