Background:
Combining optimized cognitive (Alzheimer's Disease Assessment Scale-
Cognitive subscale, ADAS-Cog) and atrophy markers of Alzheimer's disease for tracking progression in
clinical trials may provide greater sensitivity than currently used methods, which have yielded negative
results in multiple recent trials. Furthermore, it is critical to clarify the relationship among the subcomponents
yielded by cognitive and imaging testing, to address the symptomatic and anatomical variability
of Alzheimer's disease.
Method:
Using latent variable analysis, we thoroughly investigated the relationship between cognitive
impairment, as assessed on the ADAS-Cog, and cerebral atrophy. A biomarker was developed for Alzheimer's
clinical trials that combines cognitive and atrophy markers.
Results:
Atrophy within specific brain regions was found to be closely related with impairment in cognitive
domains of memory, language, and praxis. The proposed biomarker showed significantly better sensitivity
in tracking progression of cognitive impairment than the ADAS-Cog in simulated trials and a
real world problem. The biomarker also improved the selection of MCI patients (78.8±4.9% specificity
at 80% sensitivity) that will evolve to Alzheimer's disease for clinical trials.
Conclusion:
The proposed biomarker provides a boost to the efficacy of clinical trials focused in the
mild cognitive impairment (MCI) stage by significantly improving the sensitivity to detect treatment
effects and improving the selection of MCI patients that will evolve to Alzheimer’s disease.
A machine learning enrichment strategy for presymptomatic cohorts in Alzheimer’s disease clinical trials