Validating Digital Cognitive Biomarkers to Advance Alzheimer's Drug Development

PROJECT SUMMARY / ABSTRACT
The goal of this study is to evaluate the utility of non-invasive and cost-effective digital cognitive
biomarkers for concurrent prediction of amyloid positivity in pre-clinical stages of Alzheimer’s disease
(AD). As AD research has shifted its focus to earlier stages of the disease course, overcoming the
economic and logistical barriers of identifying cognitively normal subjects with accumulating AD
biomarkers (e.g., amyloid and tau) is of paramount importance. The current gold standard method of
identifying cognitively normal subjects with accumulating AD pathology includes invasive and costly
biomarker imaging or lumbar punctures, which result in high screen failure rates for biomarker positivity
and unnecessarily long lead times for clinical trial enrollment.
In our preliminary study, we used Hierarchical Bayesian Cognitive Processing (HBCP) models to analyze
baseline item response data from wordlist memory (WLM) tests, and we generated digital biomarkers
that distinguished between amyloid positive and amyloid negative groups. This study was conducted
using data from the Alzheimer’s Disease Neuroimaging Initiative (ADNI), including ADAS-Cog WLM tests
and amyloid measurement by PET and cerebrospinal fluid (CSF).
In the proposed study, we will replicate previous results and further evaluate the utility of HBCP model-
generated digital cognitive biomarkers (DCBs) for concurrent prediction of amyloid positivity in pre-clinical
stages of AD. For this study, we will use CSF and PiB-PET biomarker data plus Auditory-Verbal Learning
Test (AVLT) WLM item response data from the ADNI database. Replicating the preliminary results and
further refining those DCBs will enable a pragmatic and cost-effective approach to identifying cognitively
normal but amyloid positive subjects who are in the pre-clinical stages of AD. While physical biomarkers
will remain as industry standards for the foreseeable future, DCBs could play an important
complementary role in the screening process. This would significantly expedite clinical trial enrollment
and bring new AD therapies more quickly to market, while also enabling a scalable approach to identifying
patients who might benefit from disease-modifying therapies once approved.PROJECT NARRATIVE
As the Alzheimer’s disease (AD) research field has shifted its focus to earlier stage intervention,
overcoming the economic and logistical barriers of identifying of cognitively normal subjects with
accumulating AD biomarkers (e.g., amyloid and tau) is of paramount importance. The gold standard
methods for identifying cognitively normal subjects with accumulating AD pathology include costly and
invasive biomarker imaging or lumbar punctures, which result in long lead times for study enrollment and
high screen failure rates for biomarker positivity. We propose to evaluate the utility of non-invasive and
cost efficient digital cognitive biomarkers, to concurrently predict amyloid positivity in pre-clinical stages
of AD.