Using prediction analysis for microarrays (PAM), peripheral blood mononuclear cells (BMC) gene expression proves useful as predictive markers of sporadic AD. We obtained 90% accuracy for correct AD classification by PAM (36 of 40; four NEC classified as AD) using BMC gene expression. Investigating the contribution of these genes to the pathology of AD and their usefulness as prognostic biomarkers holds promise in the prevention of this devastating disease. Overall design: Peripheral blood mononuclear cells (BMC) were obtained from normal elderly control (NEC) and Alzheimer disease (AD) subjects. Targets from biological replicates of NEC (n=22) and AD (n=18) were generated and the expression profiles were determined using the NIA Human MGC cDNA microarray. Disease classification was preformed by the PAM software provided by Stanford University, using a subset of differentially expressed genes in AD vs. NEC.