Alzheimer's disease (AD) has been linked to a state of cerebral and systemic inflammation. The objective of the present study was to determine whether singular markers or a set of inflammatory biomarkers in peripheral blood allow discrimination between AD patients and healthy controls at the individual level. Using bead based multiplexed sandwich immunoassays, 25 inflammatory biomarkers were measured in 164 serum samples from individuals with early AD and age-matched cognitively healthy elderly controls. The data set was randomly split into a training set for feature selection and classification training and a test set for class prediction of blinded samples (1 : 1 ratio) to evaluate the chosen predictors and parameters. Multivariate data analysis was performed with use of a support vector machine (SVM). After selection of sTNF-R1 as most discriminative parameter in the training set, the application of SVM to the independent test dataset resulted in a 90.0% correct classification for individual AD and control subjects. We identified sTNF-R1 from a marker set consisting of 25 inflammatory biomarkers, which allowed SVM-based discrimination of AD patients from healthy controls on a single-subject classification level comparably well as biomarker panels with a clinically relevant accuracy and validity. Although larger sample populations will be needed to confirm this diagnostic accuracy, our study suggests that sTNF-R1 in serum-either as singular marker or incorporated into a biomarker panel-could be a powerful new biomarker for detection of AD. In addition, selective inhibition of TNF-R1 function may represent a new therapeutic approach in AD.