{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"institution":[{"name":"bioRxiv"}],"indexed":{"date-parts":[[2026,7,30]],"date-time":"2026-07-30T10:29:33Z","timestamp":1785407373899,"version":"3.56.0"},"posted":{"date-parts":[[2026,2,4]]},"group-title":"Bioinformatics","reference-count":55,"publisher":"openRxiv","license":[{"start":{"date-parts":[[2026,2,4]],"date-time":"2026-02-04T00:00:00Z","timestamp":1770163200000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/www.biorxiv.org\/about\/FAQ#license"}],"funder":[{"id":[{"id":"https:\/\/ror.org\/05k73zm37","id-type":"ROR","asserted-by":"publisher"}]},{"id":[{"id":"https:\/\/ror.org\/00ckakm23","id-type":"ROR","asserted-by":"publisher"}]},{"id":[{"id":"https:\/\/ror.org\/01qe64w73","id-type":"ROR","asserted-by":"publisher"}]},{"id":[{"id":"https:\/\/ror.org\/0574hzg66","id-type":"ROR","asserted-by":"publisher"}]}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":[],"accepted":{"date-parts":[[2026,5,18]]},"abstract":"<jats:title>Abstract<\/jats:title>\n                <jats:p>Amyloid-\u03b2 (A\u03b2) PET imaging is a core biomarker and is sufficient for the biological diagnosis of Alzheimer\u2019s disease (AD). Here, we aimed to identify biologically meaningful subgroups across the continuum of A\u03b2 accumulation using a data-driven deep learning approach, without imposing predefined thresholds for A\u03b2 negativity or positivity.<\/jats:p>\n                <jats:p>We analyzed 3,110 A\u03b2 PET scans from the Alzheimer\u2019s Disease Neuroimaging Initiative and Anti-Amyloid Treatment in Asymptomatic Alzheimer\u2019s Disease studies to develop petVAE, a two-dimensional variational autoencoder. The model accurately reconstructed scans without prior labeling, selection by scanner or region of interest. Latent representations of scans extracted from petVAE were used to visualize and cluster the AD continuum. Clustering yielded four groups: two predominantly A\u03b2 negative (A\u03b2-, A\u03b2-+) and two predominantly A\u03b2 positive (A\u03b2+, A\u03b2++).<\/jats:p>\n                <jats:p>\n                  All clusters differed in standardized uptake value ratio (\n                  <jats:italic>p<\/jats:italic>\n                  &lt; 1.64\u00d710\u207b\u2078) and cerebrospinal fluid (CSF) A\u03b2 (\n                  <jats:italic>p<\/jats:italic>\n                  &lt; 0.02), demonstrating petVAE\u2019s ability to assign scans along the A\u03b2 continuum. Extreme clusters (A\u03b2-, A\u03b2++) resembled conventional A\u03b2 negative and positive groups and differed in cognition,\n                  <jats:italic>APOE<\/jats:italic>\n                  \u03b54 prevalence, A\u03b2 and tau CSF biomarkers (\n                  <jats:italic>p<\/jats:italic>\n                  &lt; 3\u00d710\u207b\u2076). Intermediate clusters (A\u03b2-+, A\u03b2+) showed higher odds of carrying at least one\n                  <jats:italic>APOE<\/jats:italic>\n                  \u03b54 allele versus A\u03b2- (\n                  <jats:italic>p<\/jats:italic>\n                  &lt; 0.03). Participants in A\u03b2+ or A\u03b2++ clusters exhibited faster progression to AD (A\u03b2+ hazard ratio = 2.42, A\u03b2++ HR = 9.43;\n                  <jats:italic>p<\/jats:italic>\n                  &lt; 1.17\u00d710\u207b\u2077).\n                <\/jats:p>\n                <jats:p>\n                  Thus,\n                  <jats:italic>petVAE<\/jats:italic>\n                  was capable of reconstructing PET scans while extracting latent features that capture the AD continuum and define biologically meaningful subgroups, enabling data-driven characterization of preclinical disease stages.\n                <\/jats:p>","DOI":"10.64898\/2026.02.02.703218","type":"posted-content","created":{"date-parts":[[2026,2,4]],"date-time":"2026-02-04T23:15:28Z","timestamp":1770246928000},"source":"Crossref","is-referenced-by-count":0,"title":["petVAE: A Data-Driven Model for Identifying Amyloid PET Subgroups Across the Alzheimer\u2019s Disease Continuum"],"prefix":"10.64898","author":[{"ORCID":"https:\/\/orcid.org\/0000-0002-2906-6279","authenticated-orcid":false,"given":"Arina A.","family":"Tagmazian","sequence":"first","affiliation":[{"name":"Institute for Molecular Medicine Finland (FIMM), HiLIFE, University of Helsinki, Helsinki, Finland"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-3234-912X","authenticated-orcid":false,"given":"Claudia","family":"Schwarz","sequence":"additional","affiliation":[{"name":"Department of Neurology, University Medicine Greifswald, Greifswald, Germany"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-6745-8441","authenticated-orcid":false,"given":"Catharina","family":"Lange","sequence":"additional","affiliation":[{"name":"Department of Nuclear Medicine, Charit\u00e9 - Universit\u00e4tsmedizin Berlin, Corporate Member of Freie Universit\u00e4t Berlin and Humboldt-Universit\u00e4t zu Berlin, Berlin, Germany"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-9818-6370","authenticated-orcid":false,"given":"Esa","family":"Pitk\u00e4nen","sequence":"additional","affiliation":[{"name":"Institute for Molecular Medicine Finland (FIMM), HiLIFE, University of Helsinki, Helsinki, Finland"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-6534-3667","authenticated-orcid":false,"given":"Eero","family":"Vuoksimaa","sequence":"additional","affiliation":[{"name":"Institute for Molecular Medicine Finland (FIMM), HiLIFE, University of Helsinki, Helsinki, Finland"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"name":"the Alzheimer\u2019s Disease Neuroimaging Initiative","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"54368","reference":[{"key":"2026052010551698000_2026.02.02.703218v2.1","doi-asserted-by":"crossref","first-page":"2116","DOI":"10.1093\/ageing\/afab144","article-title":"Total cost of care increases significantly from early to mild Alzheimer\u2019s disease: 5-year ALSOVA follow-up","volume":"50","year":"2021","journal-title":"Age Ageing"},{"key":"2026052010551698000_2026.02.02.703218v2.2","doi-asserted-by":"crossref","unstructured":"2025 Alzheimer\u2019s disease facts and figures. Alzheimer\u2019s & Dementia. 2025;21: e70235.","DOI":"10.1002\/alz.70235"},{"key":"2026052010551698000_2026.02.02.703218v2.3","doi-asserted-by":"crossref","unstructured":"Schindler SE , Jasielec MS , Weng H , Hassenstab JJ , Grober E , McCue LM , et al. Neuropsychological measures that detect early impairment and decline in preclinical Alzheimer disease. Neurobiology of aging. 2017;56: 25.","DOI":"10.1016\/j.neurobiolaging.2017.04.004"},{"key":"2026052010551698000_2026.02.02.703218v2.4","doi-asserted-by":"publisher","DOI":"10.1002\/ana.24040"},{"key":"2026052010551698000_2026.02.02.703218v2.5","doi-asserted-by":"publisher","DOI":"10.1146\/annurev-clinpsy-072720-014213"},{"key":"2026052010551698000_2026.02.02.703218v2.6","doi-asserted-by":"publisher","DOI":"10.1038\/nrneurol.2012.241"},{"key":"2026052010551698000_2026.02.02.703218v2.7","doi-asserted-by":"publisher","DOI":"10.1016\/j.jalz.2011.03.004"},{"key":"2026052010551698000_2026.02.02.703218v2.8","doi-asserted-by":"publisher","DOI":"10.1016\/j.jalz.2018.02.018"},{"key":"2026052010551698000_2026.02.02.703218v2.9","doi-asserted-by":"crossref","first-page":"5143","DOI":"10.1002\/alz.13859","article-title":"Revised criteria for diagnosis and staging of Alzheimer\u2019s disease: Alzheimer\u2019s Association Workgroup","volume":"20","year":"2024","journal-title":"Alzheimer\u2019s & Dementia"},{"key":"2026052010551698000_2026.02.02.703218v2.10","doi-asserted-by":"crossref","unstructured":"Klunk WE , Koeppe RA , Price JC , Benzinger T , Devous MD Sr , Jagust W , et al. The Centiloid Project: Standardizing Quantitative Amyloid Plaque Estimation by PET. Alzheimer\u2019s & dementia: the journal of the Alzheimer\u2019s Association. 2014;11: 1.","DOI":"10.1016\/j.jalz.2014.07.003"},{"key":"2026052010551698000_2026.02.02.703218v2.11","doi-asserted-by":"publisher","DOI":"10.2967\/jnmt.125.271678"},{"key":"2026052010551698000_2026.02.02.703218v2.12","doi-asserted-by":"publisher","DOI":"10.21203\/rs.3.rs-7015694\/v1"},{"key":"2026052010551698000_2026.02.02.703218v2.13","doi-asserted-by":"crossref","first-page":"102","DOI":"10.1002\/cns3.5","article-title":"Machine learning and deep learning in medicine and neuroimaging","volume":"1","year":"2023","journal-title":"Annals of the Child Neurology Society"},{"key":"2026052010551698000_2026.02.02.703218v2.14","first-page":"1","article-title":"Deep Learning Based Alzheimer Disease Diagnosis: A Comprehensive Review","volume":"5","year":"2024","journal-title":"SN Computer Science"},{"key":"2026052010551698000_2026.02.02.703218v2.15","doi-asserted-by":"crossref","unstructured":"Kingma DP , Welling M. An Introduction to Variational Autoencoders. arXiv [cs.LG]. 2019. Available: http:\/\/arxiv.org\/abs\/1906.02691","DOI":"10.1561\/9781680836233"},{"key":"2026052010551698000_2026.02.02.703218v2.16","doi-asserted-by":"crossref","unstructured":"Choi H , Ha S , Kang H , Lee H , Lee DS, for the Alzheimer\u2019s Disease Neuroimaging Initiative. Deep learning only by normal brain PET identify unheralded brain anomalies. EBioMedicine. 2019;43: 447.","DOI":"10.1016\/j.ebiom.2019.04.022"},{"key":"2026052010551698000_2026.02.02.703218v2.17","doi-asserted-by":"publisher","DOI":"10.1101\/2023.08.15.553412"},{"key":"2026052010551698000_2026.02.02.703218v2.18","doi-asserted-by":"publisher","DOI":"10.1007\/s00259-023-06440-9"},{"key":"2026052010551698000_2026.02.02.703218v2.19","doi-asserted-by":"crossref","unstructured":"Li Y , Wang X , Ge Q , Graeber MB , Yan S , Li J , et al. Brain tau PET-based identification and characterization of subpopulations in patients with Alzheimer\u2019s disease using deep learning-derived saliency maps. EJNMMI Phys. 2025;12: 55.","DOI":"10.1186\/s40658-025-00761-4"},{"key":"2026052010551698000_2026.02.02.703218v2.20","first-page":"251385","article-title":"Staging Tau Pathology in Alzheimer\u2019s Disease Using Contrastive Learning on PET Imaging Data","volume":"66","year":"2025","journal-title":"Journal of Nuclear Medicine"},{"key":"2026052010551698000_2026.02.02.703218v2.21","doi-asserted-by":"crossref","first-page":"umae019","DOI":"10.1093\/radadv\/umae019","article-title":"Deep learning-based amyloid PET harmonization to predict cognitive decline in non-demented elderly","volume":"1","year":"2024","journal-title":"Radiology Advances"},{"key":"2026052010551698000_2026.02.02.703218v2.22","doi-asserted-by":"crossref","first-page":"830","DOI":"10.1001\/jamaneurol.2017.0892","article-title":"Association of Longitudinal Cognitive Decline With Amyloid Burden in Middle-aged and Older Adults: Evidence for a Dose-Response Relationship","volume":"74","year":"2017","journal-title":"JAMA Neurol"},{"key":"2026052010551698000_2026.02.02.703218v2.23","doi-asserted-by":"publisher","DOI":"10.1212\/WNL.0000000000011222"},{"key":"2026052010551698000_2026.02.02.703218v2.24","doi-asserted-by":"publisher","DOI":"10.1212\/WNL.0b013e3181cb3e25"},{"key":"2026052010551698000_2026.02.02.703218v2.25","doi-asserted-by":"publisher","DOI":"10.1002\/ana.23908"},{"key":"2026052010551698000_2026.02.02.703218v2.26","doi-asserted-by":"publisher","DOI":"10.1002\/ana.23650"},{"key":"2026052010551698000_2026.02.02.703218v2.27","doi-asserted-by":"publisher","DOI":"10.1016\/S1474-4422(12)70142-4"},{"key":"2026052010551698000_2026.02.02.703218v2.28","doi-asserted-by":"publisher","DOI":"10.2967\/jnumed.114.153494"},{"key":"2026052010551698000_2026.02.02.703218v2.29","doi-asserted-by":"crossref","first-page":"961","DOI":"10.1001\/jamaneurol.2014.803","article-title":"The Preclinical Alzheimer Cognitive Composite: Measuring Amyloid-Related Decline","volume":"71","year":"2014","journal-title":"JAMA Neurol"},{"key":"2026052010551698000_2026.02.02.703218v2.30","doi-asserted-by":"crossref","unstructured":"Avants BB , Tustison N , Johnson H. Advanced Normalization Tools (ANTS). 2009. Available: https:\/\/psychiatry.ucsd.edu\/research\/programs-centers\/snl\/_files\/ants2.pdf","DOI":"10.54294\/uvnhin"},{"key":"2026052010551698000_2026.02.02.703218v2.31","doi-asserted-by":"publisher","DOI":"10.1002\/hbm.10123"},{"key":"2026052010551698000_2026.02.02.703218v2.32","doi-asserted-by":"publisher","DOI":"10.1016\/j.neuroimage.2007.11.034"},{"key":"2026052010551698000_2026.02.02.703218v2.33","unstructured":"Nair V , Hinton GE . Rectified linear units improve restricted boltzmann machines. Proceedings of the 27th international conference on machine learning (ICML-10). 2010."},{"key":"2026052010551698000_2026.02.02.703218v2.34","unstructured":"Maas AL , Awni Y , Hannun AY . Rectifier nonlinearities improve neural network acoustic models. Proc icml. 2013;30."},{"key":"2026052010551698000_2026.02.02.703218v2.35","unstructured":"Van Rossum G , Drake FL . Python Language Reference Manual. Network Theory.; 2003."},{"key":"2026052010551698000_2026.02.02.703218v2.36","unstructured":"Paszke A , Gross S , Massa F , Lerer A , Bradbury J , Chanan G , et al. PyTorch: An Imperative Style, High-Performance Deep Learning Library. 2019. Available: http:\/\/arxiv.org\/abs\/1912.01703"},{"key":"2026052010551698000_2026.02.02.703218v2.37","doi-asserted-by":"publisher","DOI":"10.48550\/arXiv.1711.06853"},{"key":"2026052010551698000_2026.02.02.703218v2.38","doi-asserted-by":"publisher","DOI":"10.5281\/zenodo.13936989"},{"key":"2026052010551698000_2026.02.02.703218v2.39","doi-asserted-by":"crossref","unstructured":"Abraham A , Pedregosa F , Eickenberg M , Gervais P , Mueller A , Kossaifi J , et al. Machine learning for neuroimaging with scikit-learn. Front Neuroinform. 2014;8: 14.","DOI":"10.3389\/fninf.2014.00014"},{"key":"2026052010551698000_2026.02.02.703218v2.40","doi-asserted-by":"publisher","DOI":"10.1016\/j.jneumeth.2016.03.001"},{"key":"2026052010551698000_2026.02.02.703218v2.41","doi-asserted-by":"crossref","first-page":"3061","DOI":"10.1007\/s00259-021-05662-z","article-title":"Image-level trajectory inference of tau pathology using variational autoencoder for Flortaucipir PET","volume":"49","year":"2022","journal-title":"Eur J Nucl Med Mol Imaging"},{"key":"2026052010551698000_2026.02.02.703218v2.42","doi-asserted-by":"publisher","DOI":"10.1126\/science.1566067"},{"key":"2026052010551698000_2026.02.02.703218v2.43","doi-asserted-by":"crossref","unstructured":"Sun Y , Zhao Y , Hu K , Wang M , Liu Y , Liu B , et al. Distinct spatiotemporal subtypes of amyloid deposition are associated with diverging disease profiles in cognitively normal and mild cognitive impairment individuals. Transl Psychiatry. 2023;13: 35.","DOI":"10.1038\/s41398-023-02328-2"},{"key":"2026052010551698000_2026.02.02.703218v2.44","doi-asserted-by":"crossref","first-page":"5434","DOI":"10.1002\/alz.14036","article-title":"Spatial extent as a sensitive amyloid-PET metric in preclinical Alzheimer\u2019s disease","volume":"20","year":"2024","journal-title":"Alzheimers Dement"},{"key":"2026052010551698000_2026.02.02.703218v2.45","doi-asserted-by":"crossref","first-page":"391","DOI":"10.1038\/s43587-023-00380-7","article-title":"CSF tau phosphorylation occupancies at T217 and T205 represent improved biomarkers of amyloid and tau pathology in Alzheimer\u2019s disease","volume":"3","year":"2023","journal-title":"Nat Aging"},{"key":"2026052010551698000_2026.02.02.703218v2.46","doi-asserted-by":"crossref","first-page":"6365","DOI":"10.1002\/alz.14132","article-title":"Predicting continuous amyloid PET values with CSF tau phosphorylation occupancies","volume":"20","year":"2024","journal-title":"Alzheimers Dement"},{"key":"2026052010551698000_2026.02.02.703218v2.47","doi-asserted-by":"publisher","DOI":"10.23736\/S1824-4785.19.03124-8"},{"key":"2026052010551698000_2026.02.02.703218v2.48","doi-asserted-by":"publisher","DOI":"10.1007\/s00259-012-2088-x"},{"key":"2026052010551698000_2026.02.02.703218v2.49","doi-asserted-by":"publisher","DOI":"10.1002\/ana.20009"},{"key":"2026052010551698000_2026.02.02.703218v2.50","doi-asserted-by":"crossref","unstructured":"Iaccarino L , Burnham SC , Tunali I , Wang J , Navitsky M , Arora AK , et al. A practical overview of the use of amyloid-PET Centiloid values in clinical trials and research. NeuroImage Clin. 2025;46: 103765.","DOI":"10.1016\/j.nicl.2025.103765"},{"key":"2026052010551698000_2026.02.02.703218v2.51","doi-asserted-by":"crossref","first-page":"945","DOI":"10.3233\/JAD-150814","article-title":"Optimization of statistical single subject analysis of brain FDG PET for the prognosis of mild cognitive impairment-to-Alzheimer\u2019s disease conversion","volume":"49","year":"2016","journal-title":"J Alzheimers Dis"},{"key":"2026052010551698000_2026.02.02.703218v2.52","doi-asserted-by":"publisher","DOI":"10.1007\/s00259-022-05784-y"},{"key":"2026052010551698000_2026.02.02.703218v2.53","doi-asserted-by":"crossref","unstructured":"Kinahan PE , Fletcher JW . PET\/CT Standardized Uptake Values (SUVs) in Clinical Practice and Assessing Response to Therapy. Seminars in ultrasound, CT, and MR. 2010;31: 496.","DOI":"10.1053\/j.sult.2010.10.001"},{"key":"2026052010551698000_2026.02.02.703218v2.54","doi-asserted-by":"publisher","DOI":"10.2967\/jnumed.118.210518"},{"key":"2026052010551698000_2026.02.02.703218v2.55","doi-asserted-by":"publisher","DOI":"10.1212\/WNL.0000000000004643"}],"container-title":[],"original-title":[],"link":[{"URL":"https:\/\/syndication.highwire.org\/content\/doi\/10.64898\/2026.02.02.703218","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,5,20]],"date-time":"2026-05-20T17:56:15Z","timestamp":1779299775000},"score":1,"resource":{"primary":{"URL":"http:\/\/biorxiv.org\/lookup\/doi\/10.64898\/2026.02.02.703218"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,2,4]]},"references-count":55,"URL":"https:\/\/doi.org\/10.64898\/2026.02.02.703218","relation":{},"subject":[],"published":{"date-parts":[[2026,2,4]]},"subtype":"preprint"}}