Background
The current project sought to create combined biomarker-cognitive profile to detect mild Alzheimer’s disease.
Methods
Data was analyzed from 266 participants (129 AD cases [Early AD n=93; Very Early AD n=36]; 137 controls) enrolled in the Texas Alzheimer’s Research and Care Consortium (TARCC). Non-fasting serum samples were collected from each participant and assayed via a multi-plex biomarker assay platform using electrochmiluminescence (ECL). Logistic Regression was utilized to detect early AD using two serum biomarkers (TNFα and IL7), demographic information (age) and one neuropsychological measure (Clock-4 point) as predictor variable. Disease severity was determined via Clinical Dementia Rating scale global scores.
Results
In the total sample (all levels of CDR scores), the combination of biomarkers, cognitive test score, and demographics yielded the obtained sensitivity (SN) of 0.94, specificity (SP) of 0.90 and an overall accuracy of 0.92. When examining early AD cases (i.e. CDR=0.5-1), the biomarker-cognitive profile yielded SN of 0.94, SP of 0.85 and an overall accuracy of 0.91. When restricted to very early AD cases (i.e CDR=0.5), the biomarker-cognitive profile yielded SN of 0.97, SP of 0.72 with an overall accuracy of 0.91.
Conclusions
The combination of demographics + 2 biomarkers + 1 cognitive test created a biomarker-cognitive profile that was highly accurate in detecting AD presence, even in the very early stages. This work demonstrates the complementary nature of each modality (blood biomarkers + neuropsychological assessment) and supports our previously proposed concept for Molecular Neuropsychology.