{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,6,10]],"date-time":"2026-06-10T10:57:34Z","timestamp":1781089054051,"version":"3.54.1"},"reference-count":22,"publisher":"MDPI AG","issue":"12","license":[{"start":{"date-parts":[[2026,6,6]],"date-time":"2026-06-06T00:00:00Z","timestamp":1780704000000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":["Diagnostics"],"abstract":"<jats:p>Background: Plasma biomarkers are promoted as scalable tools for the staging of Alzheimer\u2019s disease (AD), yet head-to-head comparisons against the clinical scales used to define diagnostic labels remain scarce. Reported gains from machine learning fusion of clinical and biomarker features may reflect label circularity rather than biological signals, and quantifying this circularity is a central aim of the present work. Methods: From the Alzheimer\u2019s Disease Neuroimaging Initiative (ADNI), we assembled 655 participants (CN = 296, MCI = 168, and AD = 191) with concurrent plasma biomarkers (pT217, A\u03b242\/40, NfL, and GFAP), clinical scales (MMSE, CDR-SB, and FAQ), APOE genotype, and demographics. Three pre-specified feature sets (clinical-only, biomarker plus demographic\u2013genetic, and full fusion) were compared across four classifiers (Logistic Regression, SVM, Random Forest, and XGBoost) using repeated, nested cross-validation (5-fold \u00d7 3 outer, 5-fold inner) with balanced class weighting. Because the external Center for Neurodegeneration and Translational Neuroscience (CNTN) cohort (n=130) measures pT181 rather than pT217 and lacks A\u03b242\/40, external evaluation used a separate reduced feature panel (NfL, GFAP, APOE, age, sex, and education), not the proposed pT217-inclusive panel. Results: Clinical scales alone reached a three-class AUC-OVR of 0.9539\u00b10.0041, and fusion reached 0.9559\u00b10.0046, an indistinguishable gain. Because MMSE, CDR-SB, and FAQ partly determine ADNI diagnostic labels, both estimates are circularity-inflated upper bounds and do not reflect independent classification power. Independent of this circularity, the internal plasma plus demographic\u2013genetic model still achieved AUC-OVR =0.7455\u00b10.0150, with pT217 as the dominant contributor. Pairwise discrimination was excellent for CN vs. AD (1.0000) and MCI vs. AD (0.9739) but markedly weaker for CN vs. MCI (0.9302 for fused and 0.6972 for plasma only). The separate reduced-feature model, which contains neither pT217 nor A\u03b242\/40, transferred to CNTN with AUC-OVR =0.702 (95% CI 0.635\u20130.764). Conclusions: Apparent fusion gains in ADNI are largely a consequence of label circularity. After removing the circular clinical features, the internal pT217-inclusive plasma model supports three-class CN\/MCI\/AD screening at AUC \u22480.74 and a reduced panel without pT217 transfers to an independent cohort at AUC \u22480.70. These values provide a realistic performance estimate for blood-based AD staging under the current feature set, diagnostic label structure, and cohort design, and richer feature sets or pathology-anchored labels may shift this estimate. MCI detection remains the principal bottleneck.<\/jats:p>","DOI":"10.3390\/diagnostics16121755","type":"journal-article","created":{"date-parts":[[2026,6,8]],"date-time":"2026-06-08T02:32:46Z","timestamp":1780885966000},"page":"1755","update-policy":"https:\/\/doi.org\/10.3390\/mdpi_crossmark_policy","source":"Crossref","is-referenced-by-count":0,"title":["Machine Learning-Based Multiclass Classification of Cognitive Stages Using Plasma Biomarkers, Clinical Assessments, and Genetic Features: A Repeated, Nested Cross-Validation Study in ADNI with External Evaluation in CNTN"],"prefix":"10.3390","volume":"16","author":[{"ORCID":"https:\/\/orcid.org\/0009-0007-6741-9382","authenticated-orcid":false,"given":"Jiayuan","family":"Xu","sequence":"first","affiliation":[{"name":"Department of Electrical and Electronic Engineering, University of Manchester, Manchester M13 9PL, UK"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-1582-1664","authenticated-orcid":false,"given":"Fumie","family":"Costen","sequence":"additional","affiliation":[{"name":"Department of Electrical and Electronic Engineering, University of Manchester, Manchester M13 9PL, UK"}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"1968","published-online":{"date-parts":[[2026,6,6]]},"reference":[{"key":"ref_1","unstructured":"World Health Organization (2023). Dementia: Key Facts, WHO. Available online: https:\/\/www.who.int\/news-room\/fact-sheets\/detail\/dementia."},{"key":"ref_2","doi-asserted-by":"crossref","first-page":"e105","DOI":"10.1016\/S2468-2667(21)00249-8","article-title":"Estimation of the global prevalence of dementia in 2019 and forecasted prevalence in 2050","volume":"7","author":"Nichols","year":"2022","journal-title":"Lancet Public Health"},{"key":"ref_3","doi-asserted-by":"crossref","first-page":"535","DOI":"10.1016\/j.jalz.2018.02.018","article-title":"NIA-AA Research Framework: Toward a biological definition of Alzheimer\u2019s disease","volume":"14","author":"Jack","year":"2018","journal-title":"Alzheimers Dement."},{"key":"ref_4","doi-asserted-by":"crossref","first-page":"9","DOI":"10.1056\/NEJMoa2212948","article-title":"Lecanemab in Early Alzheimer\u2019s Disease","volume":"388","author":"Swanson","year":"2023","journal-title":"N. Engl. J. Med."},{"key":"ref_5","doi-asserted-by":"crossref","first-page":"954","DOI":"10.1038\/s41591-021-01382-x","article-title":"Biomarkers for neurodegenerative diseases","volume":"27","author":"Hansson","year":"2021","journal-title":"Nat. Med."},{"key":"ref_6","doi-asserted-by":"crossref","first-page":"e20200861","DOI":"10.1084\/jem.20200861","article-title":"Blood plasma phosphorylated-tau isoforms track CNS change in Alzheimer\u2019s disease","volume":"217","author":"Horie","year":"2020","journal-title":"J. Exp. Med."},{"key":"ref_7","doi-asserted-by":"crossref","first-page":"255","DOI":"10.1001\/jamaneurol.2023.5319","article-title":"Diagnostic accuracy of a plasma phosphorylated tau 217 immunoassay for Alzheimer disease pathology","volume":"81","author":"Ashton","year":"2024","journal-title":"JAMA Neurol."},{"key":"ref_8","doi-asserted-by":"crossref","first-page":"643","DOI":"10.1111\/joim.12816","article-title":"Biomarkers for Alzheimer\u2019s disease: Current status and prospects for the future","volume":"284","author":"Blennow","year":"2018","journal-title":"J. Intern. Med."},{"key":"ref_9","doi-asserted-by":"crossref","first-page":"1471","DOI":"10.1001\/jamaneurol.2021.3671","article-title":"Differences Between Plasma and Cerebrospinal Fluid Glial Fibrillary Acidic Protein Levels Across the Alzheimer\u2019s Disease Continuum","volume":"78","author":"Benedet","year":"2021","journal-title":"JAMA Neurol."},{"key":"ref_10","doi-asserted-by":"crossref","first-page":"456","DOI":"10.1148\/radiol.2018180958","article-title":"A deep learning model to predict a diagnosis of Alzheimer disease by using 18F-FDG PET of the brain","volume":"290","author":"Ding","year":"2019","journal-title":"Radiology"},{"key":"ref_11","doi-asserted-by":"crossref","first-page":"211","DOI":"10.1016\/j.neubiorev.2020.04.026","article-title":"Artificial intelligence and neuropsychological measures: The case of Alzheimer\u2019s disease","volume":"114","author":"Battista","year":"2020","journal-title":"Neurosci. Biobehav. Rev."},{"key":"ref_12","doi-asserted-by":"crossref","first-page":"68","DOI":"10.1016\/j.neuroimage.2017.06.061","article-title":"Cross-validation failure: Small sample sizes lead to large error bars","volume":"180","author":"Varoquaux","year":"2018","journal-title":"Neuroimage"},{"key":"ref_13","first-page":"2079","article-title":"On over-fitting in model selection and subsequent selection bias in performance evaluation","volume":"11","author":"Cawley","year":"2010","journal-title":"J. Mach. Learn. Res."},{"key":"ref_14","doi-asserted-by":"crossref","first-page":"106","DOI":"10.1038\/nrneurol.2012.263","article-title":"Apolipoprotein E and Alzheimer disease: Risk, mechanisms and therapy","volume":"9","author":"Liu","year":"2013","journal-title":"Nat. Rev. Neurol."},{"key":"ref_15","unstructured":"(2026, February 01). Alzheimer\u2019s Disease Neuroimaging Initiative (ADNI). Available online: https:\/\/adni.loni.usc.edu."},{"key":"ref_16","unstructured":"(2026, January 06). Center for Neurodegeneration and Translational Neuroscience (CNTN). Available online: https:\/\/nevadacntn.org\/."},{"key":"ref_17","doi-asserted-by":"crossref","unstructured":"Van Calster, B., McLernon, D.J., van Smeden, M., Wynants, L., and Steyerberg, E.W. (2019). Calibration: The Achilles heel of predictive analytics. BMC Med., 17.","DOI":"10.1186\/s12916-019-1466-7"},{"key":"ref_18","unstructured":"Lundberg, S.M., and Lee, S.I. (2017). A unified approach to interpreting model predictions. Advances in Neural Information Processing Systems 30, Curran Associates, Inc."},{"key":"ref_19","doi-asserted-by":"crossref","first-page":"1085","DOI":"10.1038\/s41591-024-02869-z","article-title":"Highly accurate blood test for Alzheimer\u2019s disease is similar or superior to clinical cerebrospinal fluid tests","volume":"30","author":"Schindler","year":"2024","journal-title":"Nat. Med."},{"key":"ref_20","doi-asserted-by":"crossref","first-page":"4987","DOI":"10.1002\/alz.13083","article-title":"Comparing machine learning-derived MRI-based and blood-based neurodegeneration biomarkers in predicting syndromal conversion in early AD","volume":"19","author":"Cai","year":"2023","journal-title":"Alzheimers Dement."},{"key":"ref_21","doi-asserted-by":"crossref","first-page":"102712","DOI":"10.1016\/j.nicl.2021.102712","article-title":"Cross-cohort generalizability of deep and conventional machine learning for MRI-based diagnosis and prediction of Alzheimer\u2019s disease","volume":"31","author":"Bron","year":"2021","journal-title":"NeuroImage Clin."},{"key":"ref_22","doi-asserted-by":"crossref","unstructured":"Zhao, Y., Zhang, J., Chen, Y., and Jiang, J. (2022). A Novel Deep Learning Radiomics Model to Discriminate AD, MCI and NC: An Exploratory Study Based on Tau PET Scans from ADNI. Brain Sci., 12.","DOI":"10.3390\/brainsci12081067"}],"container-title":["Diagnostics"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/www.mdpi.com\/2075-4418\/16\/12\/1755\/pdf","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,6,10]],"date-time":"2026-06-10T10:00:16Z","timestamp":1781085616000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.mdpi.com\/2075-4418\/16\/12\/1755"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,6,6]]},"references-count":22,"journal-issue":{"issue":"12","published-online":{"date-parts":[[2026,6]]}},"alternative-id":["diagnostics16121755"],"URL":"https:\/\/doi.org\/10.3390\/diagnostics16121755","relation":{},"ISSN":["2075-4418"],"issn-type":[{"value":"2075-4418","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,6,6]]}}}