[S38.002] High Accuracy of a Microarray-Based Blood Test for Alzheimer's Disease

Lucas Restrepo,1Phillip Stafford,2Stephen Johnston2
1Los Angeles, CA, USA, 2Tempe, AZ, USA

OBJECTIVE: To determine the accuracy of a novel blood test for the diagnosis of Alzheimer's Disease (AD).
BACKGROUND: There are no specific tests for AD, the most common neurodegenerative illness and a major cause of disability. The most promising AD biomarkers, namely, PET imaging with Florbetapir and Cerebrospinal Fluid (CSF) profiling of tau protein and AB 1-42, are cumbersome and expensive. Considering that disease-modifying therapies are being developed, diagnostic precision has become a necessity. Here we expand our experience with a blood test called immunosignature (IS), in which a random-sequence peptide microarray is used as the solid phase of an immunoassay to generate antibody binding profiles.
DESIGN/METHODS: Our platform consists of 10,000 random-sequence 17-mers printed on glass slides. Plasma samples are incubated with the array and bound IgG is detected with fluorescent-labeled secondary antibodies. Patients with clinical diagnosis of AD (n=44) and cognitively normal healthy age-matched controls (n=53) from AD Neuroimaging Initiative (ADNI) were tested. Top peptides bound by each group were identified using t tests and ANOVA. Clustering techniques (i.e., self-organizing maps) were used to identify antibody binding patterns.
RESULTS: We found an "AD-like" antibody signature in 47 samples and a “normal” signature in 49 samples. Immunoreactivity gradation was noted, ranging from yellow (AD) to blue (normal). The test correctly classified all 44 AD samples, but misclassified 3 controls as AD. The test had an overall accuracy >95%. Sensitivity was 1 (95% CI: 0.89-1), specificity 0.94 (95% CI: 0.83-0.98), Positive Predictive Value 0.94 (95% CI: 0.81-0.98), and Negative Predictive Value 1 (95% CI: 0.91-1).
CONCLUSIONS: We provide evidence of a highly accurate blood test for AD in a well-curated set of blood samples, justifying testing in a larger patient cohort. If these results are confirmed, IS could be utilized to screen dementia patients prior to more expensive or invasive tests such as amyloid imaging scan and CSF analysis.
Study Supported by: Grant from Cure Alzheimer's Fund (awarded to LR)
Category - Aging, Dementia, and Cognitive and Behavioral Neurology: Early Diagnosis of Dementia

Wednesday, April 30, 2014 4:10 PM

S38: Platform Blitz: Aging, Dementia, and Cognitive and Behavioral Neurology: Plasma and Other Biomarkers (4:00 PM-5:45 PM)


Close Window