{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,9,10]],"date-time":"2026-09-10T11:08:06Z","timestamp":1789038486757,"version":"build-2803163510"},"reference-count":57,"publisher":"Frontiers Media SA","license":[{"start":{"date-parts":[[2022,11,2]],"date-time":"2022-11-02T00:00:00Z","timestamp":1667347200000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"funder":[{"DOI":"10.13039\/501100004329","name":"Slovenian Research Agency","doi-asserted-by":"publisher","award":["P1-0389"],"award-info":[{"award-number":["P1-0389"]}],"id":[{"id":"10.13039\/501100004329","id-type":"DOI","asserted-by":"publisher"}]}],"content-domain":{"domain":["frontiersin.org"],"crossmark-restriction":true},"short-container-title":["Front. Aging Neurosci."],"abstract":"<jats:sec><jats:title>Background<\/jats:title><jats:p>Metabolic brain imaging with 2-[<jats:sup>18<\/jats:sup>F]fluoro-2-deoxy-D-glucose positron emission tomography (FDG PET) is a supportive diagnostic and differential diagnostic tool for neurodegenerative dementias. In the clinic, scans are usually visually interpreted. However, computer-aided approaches can improve diagnostic accuracy. We aimed to build two machine learning classifiers, based on two sets of FDG PET-derived features, for differential diagnosis of common dementia syndromes.<\/jats:p><\/jats:sec><jats:sec><jats:title>Methods<\/jats:title><jats:p>We analyzed FDG PET scans from three dementia cohorts [63 dementia due to Alzheimer\u2019s disease (AD), 79 dementia with Lewy bodies (DLB) and 23 frontotemporal dementia (FTD)], and 41 normal controls (NCs). Patients\u2019 clinical diagnosis at follow-up (25\u2009\u00b1\u200920\u2009months after scanning) or cerebrospinal fluid biomarkers for Alzheimer\u2019s disease was considered a gold standard. FDG PET scans were first visually evaluated. Scans were pre-processed, and two sets of features extracted: (1) the expressions of previously identified metabolic brain patterns, and (2) the mean uptake value in 95 regions of interest (ROIs). Two multi-class support vector machine (SVM) classifiers were tested and their diagnostic performance assessed and compared to visual reading. Class-specific regional feature importance was assessed with Shapley Additive Explanations.<\/jats:p><\/jats:sec><jats:sec><jats:title>Results<\/jats:title><jats:p>Pattern- and ROI-based classifier achieved higher overall accuracy than expert readers (78% and 80% respectively, vs. 71%). Both SVM classifiers performed similarly to one another and to expert readers in AD (F1\u2009=\u20090.74, 0.78, and 0.78) and DLB (F1\u2009=\u20090.81, 0.81, and 0.78). SVM classifiers outperformed expert readers in FTD (F1\u2009=\u20090.87, 0.83, and 0.63), but not in NC (F1\u2009=\u20090.71, 0.75, and 0.92). Visualization of the SVM model showed bilateral temporal cortices and cerebellum to be the most important features for AD; occipital cortices, hippocampi and parahippocampi, amygdala, and middle temporal lobes for DLB; bilateral frontal cortices, middle and anterior cingulum for FTD; and bilateral angular gyri, pons, and vermis for NC.<\/jats:p><\/jats:sec><jats:sec><jats:title>Conclusion<\/jats:title><jats:p>Multi-class SVM classifiers based on the expression of characteristic metabolic brain patterns or ROI glucose uptake, performed better than experts in the differential diagnosis of common dementias using FDG PET scans. Experts performed better in the recognition of normal scans and a combined approach may yield optimal results in the clinical setting.<\/jats:p><\/jats:sec>","DOI":"10.3389\/fnagi.2022.1005731","type":"journal-article","created":{"date-parts":[[2022,11,2]],"date-time":"2022-11-02T10:00:01Z","timestamp":1667383201000},"update-policy":"https:\/\/doi.org\/10.3389\/crossmark-policy","source":"Crossref","is-referenced-by-count":34,"title":["Automated differential diagnosis of dementia syndromes using FDG PET and machine learning"],"prefix":"10.3389","volume":"14","author":[{"given":"Matej","family":"Perovnik","sequence":"first","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"An","family":"Vo","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Nha","family":"Nguyen","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Jan","family":"Jam\u0161ek","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Toma\u017e","family":"Rus","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Chris C.","family":"Tang","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Maja","family":"Tro\u0161t","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"David","family":"Eidelberg","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"1965","published-online":{"date-parts":[[2022,11,2]]},"reference":[{"key":"ref1","doi-asserted-by":"publisher","first-page":"103502","DOI":"10.1016\/j.artint.2021.103502","article-title":"Explaining individual predictions when features are dependent: more accurate approximations to Shapley values","volume":"298","author":"Aas","year":"2021","journal-title":"Artif. Intell."},{"key":"ref2","doi-asserted-by":"publisher","first-page":"103","DOI":"10.1038\/s41582-018-0107-7","article-title":"Dementia with Lewy bodies \u2014 from scientific knowledge to clinical insights","volume":"15","author":"Arnaoutoglou","year":"2019","journal-title":"Nat. Rev. Neurol."},{"key":"ref3","doi-asserted-by":"publisher","first-page":"367","DOI":"10.1016\/j.jalz.2018.02.001","article-title":"2018 Alzheimer\u2019s disease facts and figures","volume":"14","author":"Association","year":"2018","journal-title":"Alzheimers Dement."},{"key":"ref4","doi-asserted-by":"publisher","first-page":"1672","DOI":"10.1016\/S0140-6736(15)00461-4","article-title":"Frontotemporal dementia","volume":"386","author":"Bang","year":"2015","journal-title":"Lancet"},{"key":"ref5","doi-asserted-by":"publisher","first-page":"837","DOI":"10.2967\/jnumed.118.219097","article-title":"Principal components analysis of brain metabolism predicts development of Alzheimer dementia","volume":"60","author":"Blazhenets","year":"2019","journal-title":"J. Nucl. Med."},{"key":"ref6","author":"Borchert","year":"2021"},{"key":"ref7","doi-asserted-by":"publisher","first-page":"684","DOI":"10.1148\/rg.343135065","article-title":"Brain PET in suspected dementia: patterns of altered FDG metabolism","volume":"34","author":"Brown","year":"2014","journal-title":"Radiographics"},{"key":"ref8","doi-asserted-by":"publisher","first-page":"439","DOI":"10.1097\/WCO.0000000000000838","article-title":"Machine learning for classification and prediction of brain diseases: recent advances and upcoming challenges","volume":"33","author":"Burgos","year":"2020","journal-title":"Curr. Opin. Neurol."},{"key":"ref9","doi-asserted-by":"publisher","first-page":"182","DOI":"10.1109\/4235.996017","article-title":"A fast and elitist multiobjective genetic algorithm: NSGA-II","volume":"6","author":"Deb","year":"2002","journal-title":"IEEE Trans. Evol. Comput."},{"key":"ref10","doi-asserted-by":"publisher","first-page":"1","DOI":"10.3389\/fnagi.2021.708932","article-title":"Genetic algorithms for optimized diagnosis of Alzheimer\u2019s disease and Frontotemporal dementia using Fluorodeoxyglucose positron emission tomography imaging","volume":"13","author":"D\u00edaz-\u00c1lvarez","year":"2022","journal-title":"Front. Aging Neurosci."},{"key":"ref11","doi-asserted-by":"publisher","first-page":"1490","DOI":"10.1016\/j.neuroimage.2009.09.017","article-title":"Differential effects of global and cerebellar normalization on detection and differentiation of dementia in FDG-PET studies","volume":"49","author":"Dukart","year":"2010","journal-title":"Neuroimage"},{"key":"ref12","doi-asserted-by":"publisher","first-page":"548","DOI":"10.1016\/j.tins.2009.06.003","article-title":"Metabolic brain networks in neurodegenerative disorders: a functional imaging approach","volume":"32","author":"Eidelberg","year":"2009","journal-title":"Trends Neurosci."},{"key":"ref13","doi-asserted-by":"publisher","first-page":"563","DOI":"10.1007\/s00259-021-05483-0","article-title":"A 3D deep learning model to predict the diagnosis of dementia with Lewy bodies, Alzheimer\u2019s disease, and mild cognitive impairment using brain 18F-FDG PET","volume":"49","author":"Etminani","year":"2021","journal-title":"Eur. J. Nucl. Med. Mol. Imaging"},{"key":"ref14","doi-asserted-by":"publisher","first-page":"102267","DOI":"10.1016\/j.nicl.2020.102267","article-title":"Evaluating 2-[18F]FDG-PET in differential diagnosis of dementia using a data-driven decision model","volume":"27","author":"Gjerum","year":"2020","journal-title":"NeuroImage Clin."},{"key":"ref15","doi-asserted-by":"publisher","first-page":"1847","DOI":"10.1172\/JCI66029","article-title":"Thinking laterally about neurodegenerative proteinopathies","volume":"123","author":"Golde","year":"2013","journal-title":"J. Clin. Invest."},{"key":"ref16","doi-asserted-by":"publisher","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":"ref17","doi-asserted-by":"publisher","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":"ref18","doi-asserted-by":"publisher","first-page":"13236","DOI":"10.1038\/s41598-018-31653-6","article-title":"Machine learning identified an Alzheimer\u2019s disease-related FDG-PET pattern which is also expressed in Lewy body dementia and Parkinson\u2019s disease dementia","volume":"8","author":"Katako","year":"2018","journal-title":"Sci. Rep."},{"key":"ref19","doi-asserted-by":"publisher","first-page":"1","DOI":"10.1093\/cercor\/bhx267","article-title":"Network structure and function in Parkinson\u2019s disease","volume":"28","author":"Ko","year":"2018","journal-title":"Cereb. Cortex"},{"key":"ref20","doi-asserted-by":"publisher","first-page":"59","DOI":"10.1111\/ene.13439","article-title":"Alzheimer\u2019s disease","volume":"25","author":"Lane","year":"2018","journal-title":"Eur. J. Neurol."},{"key":"ref21","doi-asserted-by":"publisher","first-page":"503","DOI":"10.1007\/BF01589116","article-title":"On the limited memory BFGS method for large scale optimization","volume":"45","author":"Liu","year":"1989","journal-title":"Math. Program."},{"key":"ref22","doi-asserted-by":"publisher","first-page":"56","DOI":"10.1038\/s42256-019-0138-9","article-title":"From local explanations to global understanding with explainable AI for trees","volume":"2","author":"Lundberg","year":"2020","journal-title":"Nat. Mach. Intell."},{"key":"ref23","author":"Lundberg","year":"2017"},{"key":"ref24","doi-asserted-by":"publisher","first-page":"234","DOI":"10.1109\/TBME.2011.2170986","article-title":"Design and application of a generic clinical decision support system for multiscale data","volume":"59","author":"Mattila","year":"2012","journal-title":"I.E.E.E. Trans. Biomed. Eng."},{"key":"ref25","doi-asserted-by":"publisher","first-page":"88","DOI":"10.1212\/WNL.0000000000004058","article-title":"Diagnosis and management of dementia with Lewy bodies fourth consensus report of the DLB consortium","volume":"89","author":"McKeith","year":"2017","journal-title":"Neurology"},{"key":"ref26","doi-asserted-by":"publisher","first-page":"263","DOI":"10.1016\/j.jalz.2011.03.005","article-title":"The diagnosis of dementia due to Alzheimer\u2019s disease: recommendations from the National Institute on Aging-Alzheimer\u2019s Association workgroups on diagnostic guidelines for Alzheimer\u2019s disease","volume":"7","author":"McKhann","year":"2011","journal-title":"Alzheimers Dement."},{"key":"ref27","doi-asserted-by":"publisher","first-page":"440","DOI":"10.1038\/s41582-020-0377-8","article-title":"Applications of machine learning to diagnosis and treatment of neurodegenerative diseases","volume":"16","author":"Myszczynska","year":"2020","journal-title":"Nat. Rev. Neurol."},{"key":"ref28","doi-asserted-by":"publisher","first-page":"583","DOI":"10.1016\/j.dadm.2018.07.009","article-title":"A multivariate metabolic imaging marker for behavioral variant frontotemporal dementia. Alzheimer\u2019s dement. Diagnosis","volume":"10","author":"Nazem","year":"2018","journal-title":"Assess. Dis. Monit."},{"key":"ref29","doi-asserted-by":"publisher","first-page":"1509","DOI":"10.1007\/s00259-018-4035-y","article-title":"Clinical utility of FDG-PET for the differential diagnosis among the main forms of dementia","volume":"45","author":"Nestor","year":"2018","journal-title":"Eur. J. Nucl. Med. Mol. Imaging"},{"key":"ref30","doi-asserted-by":"publisher","first-page":"547","DOI":"10.1212\/WNL.0b013e31823b9c5e","article-title":"Visual assessment vs. quantitative assessment of 11C-PIB PET and 18F-FDG PET for detection of Alzheimer\u2019s disease","volume":"48","author":"Ng","year":"2007","journal-title":"J. Nucl. Med."},{"key":"ref31","doi-asserted-by":"publisher","first-page":"88","DOI":"10.1016\/S1474-4422(18)30403-4","article-title":"Global, regional, and national burden of Alzheimer\u2019s disease and other dementias, 1990\u20132016: a systematic analysis for the global burden of disease study 2016","volume":"18","author":"Nichols","year":"2019","journal-title":"Lancet Neurol."},{"key":"ref32","doi-asserted-by":"publisher","first-page":"1557","DOI":"10.1007\/s00259-018-4030-3","article-title":"Automated assessment of FDG-PET for differential diagnosis in patients with neurodegenerative disorders","volume":"45","author":"Nobili","year":"2018","journal-title":"Eur. J. Nucl. Med. Mol. Imaging"},{"key":"ref33","doi-asserted-by":"publisher","first-page":"2763","DOI":"10.1038\/S41598-022-06663-0","article-title":"A replication study, systematic review and meta-analysis of automated image-based diagnosis in parkinsonism","volume":"12","author":"Papathoma","year":"2022","journal-title":"Sci. Rep."},{"key":"ref34","doi-asserted-by":"publisher","first-page":"11752","DOI":"10.1038\/s41598-022-15667-9","article-title":"Identification and validation of Alzheimer\u2019s disease-related metabolic brain pattern in biomarker confirmed Alzheimer\u2019s dementia patients","volume":"12","author":"Perovnik","year":"","journal-title":"Sci. Rep."},{"key":"ref35","doi-asserted-by":"publisher","first-page":"103080","DOI":"10.1016\/j.nicl.2022.103080","article-title":"Metabolic brain pattern in dementia with Lewy bodies: relationship to Alzheimer\u2019s disease topography","volume":"35","author":"Perovnik","year":"","journal-title":"NeuroImage Clin."},{"key":"ref36","year":"2019"},{"key":"ref37","doi-asserted-by":"publisher","first-page":"2034","DOI":"10.1212\/WNL.0b013e31823b9c5e","article-title":"Amyloid vs FDG-PET in the differential diagnosis of AD and FTLD","volume":"77","author":"Rabinovici","year":"2011","journal-title":"Neurology"},{"key":"ref38","doi-asserted-by":"publisher","first-page":"2456","DOI":"10.1093\/brain\/awr179","article-title":"Sensitivity of revised diagnostic criteria for the behavioural variant of frontotemporal dementia","volume":"134","author":"Rascovsky","year":"2011","journal-title":"Brain"},{"key":"ref39","doi-asserted-by":"publisher","first-page":"116189","DOI":"10.1016\/j.neuroimage.2019.116189","article-title":"Automated anatomical labelling atlas 3","volume":"206","author":"Rolls","year":"2020","journal-title":"Neuroimage"},{"key":"ref40","doi-asserted-by":"publisher","first-page":"e056357","DOI":"10.1002\/ALZ.056357","article-title":"Abnormal metabolic brain network in behavioral variant of frontotemporal dementia","volume":"17","author":"Rus","year":"2021","journal-title":"Alzheimers Dement."},{"key":"ref41","doi-asserted-by":"publisher","first-page":"2901","DOI":"10.1007\/s00259-020-04785-z","article-title":"Differential diagnosis of parkinsonian syndromes: a comparison of clinical and automated - metabolic brain patterns\u2019 based approach","volume":"47","author":"Rus","year":"2020","journal-title":"Eur. J. Nucl. Med. Mol. Imaging"},{"key":"ref42","doi-asserted-by":"publisher","first-page":"2899","DOI":"10.1016\/j.neuroimage.2010.10.025","article-title":"Scaled subprofile modeling of resting state imaging data in Parkinson\u2019s disease: methodological issues","volume":"54","author":"Spetsieris","year":"2011","journal-title":"Neuroimage"},{"key":"ref43","doi-asserted-by":"publisher","first-page":"117568","DOI":"10.1016\/j.neuroimage.2020.117568","article-title":"Spectral guided sparse inverse covariance estimation of metabolic networks in Parkinson\u2019s disease","volume":"226","author":"Spetsieris","year":"2021","journal-title":"Neuroimage"},{"key":"ref44","doi-asserted-by":"publisher","first-page":"2563","DOI":"10.1073\/pnas.1411011112","article-title":"Metabolic resting-state brain networks in health and disease","volume":"112","author":"Spetsieris","year":"2015","journal-title":"Proc. Natl. Acad. Sci."},{"key":"ref45","doi-asserted-by":"publisher","first-page":"382","DOI":"10.1038\/nn.4513","article-title":"Glucose utilization: still in the synapse","volume":"20","author":"Stoessl","year":"2017","journal-title":"Nat. Neurosci."},{"key":"ref46","doi-asserted-by":"publisher","first-page":"1049","DOI":"10.1523\/JNEUROSCI.4188-09.2010","article-title":"Abnormalities in metabolic network activity precede the onset of motor symptoms in Parkinson\u2019s disease","volume":"30","author":"Tang","year":"2010","journal-title":"J. Neurosci."},{"key":"ref47","doi-asserted-by":"publisher","first-page":"882","DOI":"10.1136\/jnnp.2009.194779","article-title":"Molecular imaging in the diagnosis of Alzheimer\u2019s disease: visual assessment of [11C]PIB and [18F]FDDNP PET images","volume":"81","author":"Tolboom","year":"2010","journal-title":"J. Neurol. Neurosurg. Psychiatry"},{"key":"ref48","doi-asserted-by":"publisher","first-page":"23","DOI":"10.1007\/s00401-012-0983-7","article-title":"CSF biomarkers cutoffs: the importance of coincident neuropathological diseases","volume":"124","author":"Toledo","year":"2012","journal-title":"Acta Neuropathol."},{"key":"ref49","doi-asserted-by":"publisher","first-page":"507","DOI":"10.1007\/s00234-017-1821-3","article-title":"Abnormal metabolic brain network associated with Parkinson\u2019s disease: replication on a new European sample","volume":"59","author":"Tom\u0161e","year":"2017","journal-title":"Neuroradiology"},{"key":"ref50","doi-asserted-by":"publisher","first-page":"613","DOI":"10.1016\/j.nicl.2017.06.012","article-title":"Five-class differential diagnostics of neurodegenerative diseases using random undersampling boosting","volume":"15","author":"Tong","year":"2017","journal-title":"NeuroImage Clin."},{"key":"ref51","doi-asserted-by":"publisher","first-page":"60","DOI":"10.2967\/jnumed.115.161992","article-title":"Automated differential diagnosis of early parkinsonism using metabolic brain networks: a validation study","volume":"57","author":"Tripathi","year":"2016","journal-title":"J. Nucl. Med."},{"key":"ref52","doi-asserted-by":"publisher","first-page":"13","DOI":"10.15274\/NRJ-2014-10002","article-title":"Differential diagnosis of neurodegenerative dementias using metabolic phenotypes on F-18 FDG PET\/CT","volume":"27","author":"Tripathi","year":"2014","journal-title":"Neuroradiol. J."},{"key":"ref53","doi-asserted-by":"publisher","first-page":"273","DOI":"10.1006\/nimg.2001.0978","article-title":"Automated anatomical labeling of activations in SPM using a macroscopic anatomical parcellation of the MNI MRI single-subject brain","volume":"15","author":"Tzourio-Mazoyer","year":"2002","journal-title":"Neuroimage"},{"key":"ref54","doi-asserted-by":"publisher","first-page":"2103","DOI":"10.1007\/s00259-009-1264-0","article-title":"EANM procedure guidelines for PET brain imaging using [18F]FDG, version 2","volume":"36","author":"Varrone","year":"2009","journal-title":"Eur. J. Nucl. Med. Mol. Imaging"},{"key":"ref55","doi-asserted-by":"publisher","first-page":"254","DOI":"10.1177\/0891988716654984","article-title":"Imaging in dementia with Lewy bodies: an overview","volume":"29","author":"Watson","year":"2016","journal-title":"J. Geriatr. Psychiatry Neurol."},{"key":"ref56","doi-asserted-by":"publisher","first-page":"1","DOI":"10.1155\/2014\/421743","article-title":"Automated identification of dementia using FDG-PET imaging","volume":"2014","author":"Xia","year":"2014","journal-title":"Biomed. Res. Int."},{"key":"ref57","doi-asserted-by":"publisher","first-page":"244","DOI":"10.3174\/ajnr.A3665","article-title":"Visual-statistical interpretation of 18F-FDG-PET images for characteristic Alzheimer patterns in a multicenter study: inter-rater concordance and relationship to automated quantitative evaluation","volume":"35","author":"Yamane","year":"2014","journal-title":"Am. J. Neuroradiol."}],"container-title":["Frontiers in Aging Neuroscience"],"original-title":[],"link":[{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/fnagi.2022.1005731\/full","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2022,11,2]],"date-time":"2022-11-02T10:00:04Z","timestamp":1667383204000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/fnagi.2022.1005731\/full"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2022,11,2]]},"references-count":57,"alternative-id":["10.3389\/fnagi.2022.1005731"],"URL":"https:\/\/doi.org\/10.3389\/fnagi.2022.1005731","relation":{},"ISSN":["1663-4365"],"issn-type":[{"value":"1663-4365","type":"electronic"}],"subject":[],"published":{"date-parts":[[2022,11,2]]},"article-number":"1005731"}}