Objective: Blood-based biomarkers offer a time and costefficient method for screening cognitive dysfunction and blood-based biomarkers can be combined with select cognitive instruments to detect early AD and distinguish MCI from AD. This study examined the ability of combined proteomic and cognitive data to detect amnestic (aMCI) and non-amnestic MCI (non-aMCI). Method: Stored serum samples were analyzed from 219 participants (88 normal controls and 73 MCI (58 aMCI, 15 non-aMCI) from the University of Texas Southwestern Medical Center Alzheimer's Disease Center. Analyses of CRP, SAA, ICAM, VCAM, A2M, B2M, FVII, TNC, CA125, Eotaxin3, IL5, IL6, IL7, IL10, IL18, TARC, TNFa, FABP, I309, PPY and THPO were conducted via electrochemiluminescence using the Meso Scale Discovery platform. Logistic regression models evaluated diagnosis (NC vs. aMCI, NC vs. non-AMCI) as the outcome variable, and serum biomarkers and animal fluency as the predictor variables; age, gender, education and APOE4 genotype were included in the models. Results: Age, gender, education, APOE4, and animal fluency achieved a sensitivity (SN) of 0.25 and specificity (SP) of 0.98 for non-aMCI, and a SN = 0.63 and SP = 0.77 for aMCI. Inclusion of serum proteomic data into the model resulted in SN and SP = 1.0 for non-aMCI and SN = 0.75 and SP = 0.87 for aMCI. Conclusion(s): Results suggest that combined serum proteomic data and select cognitive testing can accurately classify aMCI and non-aMCI. The long-term goal of this work is the creation of a point-of-care procedure to screen for early cognitive loss with appropriate referrals then made for neuropsychological evaluation.