A Novel Panel of Plasma Proteins Predicts Progression in Prodromal Alzheimer’s Disease

Article type: Research Article

Authors: Araújo, Daniella Castro^{a; d; e} | Veloso, Adriano Alonso^a | Gomes, Karina Braga^b | de Souza, Leonardo Cruz^c | Ziviani, Nivio^{a; d} | Caramelli, Paulo^{c; *} | for the Alzheimer’s Disease Neuroimaging Initiative¹

Affiliations: [a] Computer Science Department, Universidade Federal de Minas Gerais, Belo Horizonte, MG, Brazil | [b] School of Pharmacy, Universidade Federal de Minas Gerais, Belo Horizonte, MG, Brazil | [c] School of Medicine, Universidade Federal de Minas Gerais, Belo Horizonte, MG, Brazil | [d] Kunumi, Belo Horizonte, MG, Brazil | [e] Huna, São Paulo, SP, Brazil

Correspondence: [*] Correspondence to: Prof. Paulo Caramelli, Faculdade de Medicina da UFMG, Av. Prof. Alfredo Balena, 190 - Sala 246, 30130-100, Belo Horizonte (MG) –Brazil. E-mail: caramelli@ufmg.br.

Note: [1] Data used in preparation of this article were obtained from the Alzheimer’s Disease Neuroimaging Initiative (ADNI) data-base (https://adni.loni.usc.edu). As such, the investigators within the ADNI contributed to the design and implementation of ADNI and/or provided data but did not participate in analysis or writing of this report. A complete listing of ADNI investigators can be found at: https://adni.loni.usc.edu/wp-content/uploads/how_to_apply/ADNI_Acknowledgement_List.pdf

Abstract: Background:A cheap and minimum-invasive method for early identification of Alzheimer’s disease (AD) pathogenesis is key to disease management and the success of emerging treatments targeting the prodromal phases of the disease. Objective:To develop a machine learning-based blood panel to predict the progression from mild cognitive impairment (MCI) to dementia due to AD within a four-year time-to-conversion horizon. Methods:We created over one billion models to predict the probability of conversion from MCI to dementia due to AD and chose the best-performing one. We used Alzheimer’s Disease Neuroimaging Initiative (ADNI) data of 379 MCI individuals in the baseline visit, from which 176 converted to AD dementia. Results:We developed a machine learning-based panel composed of 12 plasma proteins (ApoB, Calcitonin, C-peptide, CRP, IGFBP-2, Interleukin-3, Interleukin-8, PARC, Serotransferrin, THP, TLSP 1-309, and TN-C), and which yielded an AUC of 0.91, accuracy of 0.91, sensitivity of 0.84, and specificity of 0.98 for predicting the risk of MCI patients converting to dementia due to AD in a horizon of up to four years. Conclusion:The proposed machine learning model was able to accurately predict the risk of MCI patients converting to dementia due to AD in a horizon of up to four years, suggesting that this model could be used as a minimum-invasive tool for clinical decision support. Further studies are needed to better clarify the possible pathophysiological links with the reported proteins.

Keywords: Alzheimer’s disease, artificial intelligence, biomarkers, machine learning, proteomics

DOI: 10.3233/JAD-220256

Journal: Journal of Alzheimer's Disease, vol. 88, no. 2, pp. 549-561, 2022