{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,10,13]],"date-time":"2026-10-13T18:03:33Z","timestamp":1791914613768,"version":"4.3.4"},"reference-count":42,"publisher":"Public Library of Science (PLoS)","issue":"8","license":[{"start":{"date-parts":[[2022,8,10]],"date-time":"2022-08-10T00:00:00Z","timestamp":1660089600000},"content-version":"vor","delay-in-days":0,"URL":"http:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"funder":[{"DOI":"10.13039\/100000057","name":"National Institute of General Medical Sciences of the National Institutes of Health","doi-asserted-by":"publisher","award":["K23 GM140268"],"award-info":[{"award-number":["K23 GM140268"]}],"id":[{"id":"10.13039\/100000057","id-type":"DOI","asserted-by":"publisher"}]},{"DOI":"10.13039\/100000062","name":"National Institute of Diabetes and Digestive and Kidney Diseases","doi-asserted-by":"publisher","award":["K01 DK120784"],"award-info":[{"award-number":["K01 DK120784"]}],"id":[{"id":"10.13039\/100000062","id-type":"DOI","asserted-by":"publisher"}]},{"DOI":"10.13039\/100000057","name":"National Institute of General Medical Sciences","doi-asserted-by":"publisher","award":["R01GM110240"],"award-info":[{"award-number":["R01GM110240"]}],"id":[{"id":"10.13039\/100000057","id-type":"DOI","asserted-by":"publisher"}]},{"name":"University of Florida Research, and Gator Trust","award":["AWD09459"],"award-info":[{"award-number":["AWD09459"]}]},{"DOI":"10.13039\/100000057","name":"National Institute of General Medical Sciences","doi-asserted-by":"publisher","award":["R01GM114290"],"award-info":[{"award-number":["R01GM114290"]}],"id":[{"id":"10.13039\/100000057","id-type":"DOI","asserted-by":"publisher"}]},{"DOI":"10.13039\/100000049","name":"National Institute on Aging","doi-asserted-by":"publisher","award":["R01AG121647"],"award-info":[{"award-number":["R01AG121647"]}],"id":[{"id":"10.13039\/100000049","id-type":"DOI","asserted-by":"publisher"}]},{"name":"National Science CAREER Award","award":["1750192"],"award-info":[{"award-number":["1750192"]}]},{"DOI":"10.13039\/100000049","name":"National Institute on Aging","doi-asserted-by":"publisher","award":["P30AG028740"],"award-info":[{"award-number":["P30AG028740"]}],"id":[{"id":"10.13039\/100000049","id-type":"DOI","asserted-by":"publisher"}]},{"DOI":"10.13039\/100000049","name":"National Institute on Aging","doi-asserted-by":"publisher","award":["R01AG05533"],"award-info":[{"award-number":["R01AG05533"]}],"id":[{"id":"10.13039\/100000049","id-type":"DOI","asserted-by":"publisher"}]},{"DOI":"10.13039\/100000070","name":"National Institute of Biomedical Imaging and Bioengineering","doi-asserted-by":"publisher","award":["1R21EB027344"],"award-info":[{"award-number":["1R21EB027344"]}],"id":[{"id":"10.13039\/100000070","id-type":"DOI","asserted-by":"publisher"}]},{"DOI":"10.13039\/100000057","name":"National Institute of General Medical Sciences","doi-asserted-by":"publisher","award":["R01GM-110240"],"award-info":[{"award-number":["R01GM-110240"]}],"id":[{"id":"10.13039\/100000057","id-type":"DOI","asserted-by":"publisher"}]},{"name":"W. Martin Smith Interdisciplinary Patient Quality and Safety Award (IPQSA),"},{"name":"Sepsis and Critical Illness Research Center Award"},{"DOI":"10.13039\/100000057","name":"National Institute of General Medical Sciences","doi-asserted-by":"publisher","award":["R01 GM110240"],"award-info":[{"award-number":["R01 GM110240"]}],"id":[{"id":"10.13039\/100000057","id-type":"DOI","asserted-by":"publisher"}]},{"name":"University of Florida Research","award":["AWD09458"],"award-info":[{"award-number":["AWD09458"]}]}],"content-domain":{"domain":["www.plosdigitalhealth.org"],"crossmark-restriction":false},"short-container-title":["PLOS Digit Health"],"abstract":"<jats:p>Mistrust is a major barrier to implementing deep learning in healthcare settings. Entrustment could be earned by conveying model certainty, or the probability that a given model output is accurate, but the use of uncertainty estimation for deep learning entrustment is largely unexplored, and there is no consensus regarding optimal methods for quantifying uncertainty. Our purpose is to critically evaluate methods for quantifying uncertainty in deep learning for healthcare applications and propose a conceptual framework for specifying certainty of deep learning predictions. We searched Embase, MEDLINE, and PubMed databases for articles relevant to study objectives, complying with PRISMA guidelines, rated study quality using validated tools, and extracted data according to modified CHARMS criteria. Among 30 included studies, 24 described medical imaging applications. All imaging model architectures used convolutional neural networks or a variation thereof. The predominant method for quantifying uncertainty was Monte Carlo dropout, producing predictions from multiple networks for which different neurons have dropped out and measuring variance across the distribution of resulting predictions. Conformal prediction offered similar strong performance in estimating uncertainty, along with ease of interpretation and application not only to deep learning but also to other machine learning approaches. Among the six articles describing non-imaging applications, model architectures and uncertainty estimation methods were heterogeneous, but predictive performance was generally strong, and uncertainty estimation was effective in comparing modeling methods. Overall, the use of model learning curves to quantify epistemic uncertainty (attributable to model parameters) was sparse. Heterogeneity in reporting methods precluded the performance of a meta-analysis. Uncertainty estimation methods have the potential to identify rare but important misclassifications made by deep learning models and compare modeling methods, which could build patient and clinician trust in deep learning applications in healthcare. Efficient maturation of this field will require standardized guidelines for reporting performance and uncertainty metrics.<\/jats:p>","DOI":"10.1371\/journal.pdig.0000085","type":"journal-article","created":{"date-parts":[[2022,8,10]],"date-time":"2022-08-10T17:28:25Z","timestamp":1660152505000},"page":"e0000085","update-policy":"https:\/\/doi.org\/10.1371\/journal.pdig.corrections_policy","source":"Crossref","is-referenced-by-count":73,"title":["Uncertainty-aware deep learning in healthcare: A scoping review"],"prefix":"10.1371","volume":"1","author":[{"ORCID":"https:\/\/orcid.org\/0000-0001-5354-443X","authenticated-orcid":true,"given":"Tyler J.","family":"Loftus","sequence":"first","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-5304-7027","authenticated-orcid":true,"given":"Benjamin","family":"Shickel","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Matthew M.","family":"Ruppert","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Jeremy A.","family":"Balch","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Tezcan","family":"Ozrazgat-Baslanti","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-1881-9941","authenticated-orcid":true,"given":"Patrick J.","family":"Tighe","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Philip A.","family":"Efron","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"William R.","family":"Hogan","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-4530-2048","authenticated-orcid":true,"given":"Parisa","family":"Rashidi","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Gilbert R.","family":"Upchurch","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-5745-2863","authenticated-orcid":true,"given":"Azra","family":"Bihorac","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"340","published-online":{"date-parts":[[2022,8,10]]},"reference":[{"issue":"1","key":"pdig.0000085.ref001","doi-asserted-by":"crossref","first-page":"1879","DOI":"10.1038\/s41598-019-38491-0","article-title":"DeepSOFA: A Continuous Acuity Score for Critically Ill Patients using Clinically Interpretable Deep Learning.","volume":"9","author":"B Shickel","year":"2019","journal-title":"Sci Rep."},{"issue":"1","key":"pdig.0000085.ref002","doi-asserted-by":"crossref","first-page":"e1919396","DOI":"10.1001\/jamanetworkopen.2019.19396","article-title":"Assessment of a Machine Learning Model Applied to Harmonized Electronic Health Record Data for the Prediction of Incident Atrial Fibrillation","volume":"3","author":"P Tiwari","year":"2020","journal-title":"JAMA Netw Open"},{"issue":"7639","key":"pdig.0000085.ref003","doi-asserted-by":"crossref","first-page":"115","DOI":"10.1038\/nature21056","article-title":"Dermatologist-level classification of skin cancer with deep neural networks","volume":"542","author":"A Esteva","year":"2017","journal-title":"Nature"},{"issue":"7788","key":"pdig.0000085.ref004","doi-asserted-by":"crossref","first-page":"89","DOI":"10.1038\/s41586-019-1799-6","article-title":"International evaluation of an AI system for breast cancer screening","volume":"577","author":"SM McKinney","year":"2020","journal-title":"Nature"},{"issue":"3","key":"pdig.0000085.ref005","first-page":"3","article-title":"Autonomy and common ground in human-robot interaction: A field study (vol 22, pg 42, 2007).","volume":"22","author":"K Stubbs","year":"2007","journal-title":"Ieee Intell Syst"},{"issue":"23","key":"pdig.0000085.ref006","doi-asserted-by":"crossref","first-page":"2482","DOI":"10.1177\/154193120605002304","article-title":"Human-Automation Collaboration in Dynamic Mission Planning: A Challenge Requiring an Ecological Approach","volume":"50","author":"MP Linegang","year":"2006","journal-title":"Proceedings of the Human Factors and Ergonomics Society Annual Meeting"},{"key":"pdig.0000085.ref007","doi-asserted-by":"crossref","first-page":"1","DOI":"10.1016\/j.artint.2018.07.007","article-title":"Explanation in artificial intelligence: Insights from the social sciences","volume":"267","author":"T. Miller","year":"2019","journal-title":"Artif Intell"},{"key":"pdig.0000085.ref008","unstructured":"Tonekaboni S, Joshi S, McCradden MD, Goldenberg A. What Clinicians Want: Contextualizing Explainable Machine Learning for Clinical End Use. In: Finale D-V, Jim F, Ken J, David K, Rajesh R, Byron W, et al., editors. Proceedings of the 4th Machine Learning for Healthcare Conference; Proceedings of Machine Learning Research: PMLR; 2019. p. 359\u201380."},{"key":"pdig.0000085.ref009","article-title":"The Elephant in the Room.","author":"A Rosenfeld","year":"2018","journal-title":"arXiv:1808.03305 [cs.CV]."},{"issue":"10","key":"pdig.0000085.ref010","doi-asserted-by":"crossref","first-page":"e1001744","DOI":"10.1371\/journal.pmed.1001744","article-title":"Critical appraisal and data extraction for systematic reviews of prediction modelling studies: the CHARMS checklist.","volume":"11","author":"KG Moons","year":"2014","journal-title":"PLoS Med"},{"issue":"3","key":"pdig.0000085.ref011","doi-asserted-by":"crossref","first-page":"272","DOI":"10.1177\/1525822X08317166","article-title":"Using pooled kappa to summarize interrater agreement across many items.","volume":"20","author":"H De Vries","year":"2008","journal-title":"Field Method"},{"issue":"1","key":"pdig.0000085.ref012","doi-asserted-by":"crossref","first-page":"159","DOI":"10.2307\/2529310","article-title":"The measurement of observer agreement for categorical data","volume":"33","author":"JR Landis","year":"1977","journal-title":"Biometrics"},{"issue":"3","key":"pdig.0000085.ref013","doi-asserted-by":"crossref","first-page":"457","DOI":"10.1007\/s10994-021-05946-3","article-title":"Aleatoric and epistemic uncertainty in machine learning: an introduction to concepts and methods.","volume":"110","author":"E Hullermeier","year":"2021","journal-title":"Mach Learn."},{"key":"pdig.0000085.ref014","unstructured":"Gal Y, Ghahramani Z. Dropout as a Bayesian Approximation: Representing Model Uncertainty in Deep Learning. In: Maria Florina B, Kilian QW, editors. Proceedings of The 33rd International Conference on Machine Learning; Proceedings of Machine Learning Research: PMLR; 2016. p. 1050\u20139."},{"key":"pdig.0000085.ref015","doi-asserted-by":"crossref","first-page":"101790","DOI":"10.1016\/j.media.2020.101790","article-title":"Integrating uncertainty in deep neural networks for MRI based stroke analysis","volume":"65","author":"L Herzog","year":"2020","journal-title":"Med Image Anal"},{"key":"pdig.0000085.ref016","doi-asserted-by":"crossref","first-page":"101885","DOI":"10.1016\/j.media.2020.101885","article-title":"Super-Resolved q-Space deep learning with uncertainty quantification","volume":"67","author":"Y Qin","year":"2021","journal-title":"Med Image Anal"},{"key":"pdig.0000085.ref017","doi-asserted-by":"crossref","first-page":"117366","DOI":"10.1016\/j.neuroimage.2020.117366","article-title":"Uncertainty modelling in deep learning for safer neuroimage enhancement: Demonstration in diffusion MRI.","volume":"225","author":"R Tanno","year":"2021","journal-title":"Neuroimage"},{"key":"pdig.0000085.ref018","first-page":"75","volume-title":"Improving Aleatoric Uncertainty Quantification in Multi-annotated Medical Image Segmentation with Normalizing Flows. Uncertainty for Safe Utilization of Machine Learning in Medical Imaging, and Perinatal Imaging, Placental and Preterm Image Analysis","author":"M Valiuddin","year":"2021"},{"issue":"2","key":"pdig.0000085.ref019","doi-asserted-by":"crossref","first-page":"371","DOI":"10.1109\/JBHI.2020.2996300","article-title":"Deep Learning With Conformal Prediction for Hierarchical Analysis of Large-Scale Whole-Slide Tissue Images","volume":"25","author":"H Wieslander","year":"2021","journal-title":"IEEE J Biomed Health Inform"},{"key":"pdig.0000085.ref020","doi-asserted-by":"crossref","first-page":"331","DOI":"10.1016\/j.neucom.2019.09.106","article-title":"Audio-visual domain adaptation using conditional semi-supervised Generative Adversarial Networks.","volume":"397","author":"C Athanasiadis","year":"2020","journal-title":"Neurocomputing."},{"key":"pdig.0000085.ref021","first-page":"23","volume-title":"Hierarchical brain parcellation with uncertainty. Uncertainty for Safe Utilization of Machine Learning in Medical Imaging, and Graphs in Biomedical Image Analysis","author":"MS Graham","year":"2020"},{"key":"pdig.0000085.ref022","volume-title":"Distance metric learning using graph convolutional networks: Application to functional brain networks. International Conference on Medical Image Computing and Computer-Assisted Intervention","author":"SI Ktena","year":"2017"},{"key":"pdig.0000085.ref023","first-page":"12","volume-title":"Probabilistic image registration via deep multi-class classification: characterizing uncertainty. Uncertainty for Safe Utilization of Machine Learning in Medical Imaging and Clinical Image-Based Procedures","author":"A Sedghi","year":"2019"},{"issue":"3","key":"pdig.0000085.ref024","doi-asserted-by":"crossref","first-page":"1269","DOI":"10.1021\/acs.jcim.8b00542","article-title":"Deep Confidence: A Computationally Efficient Framework for Calculating Reliable Prediction Errors for Deep Neural Networks.","volume":"59","author":"I Cortes-Ciriano","year":"2019","journal-title":"J Chem Inf Model."},{"issue":"6","key":"pdig.0000085.ref025","doi-asserted-by":"crossref","first-page":"2697","DOI":"10.1021\/acs.jcim.9b00975","article-title":"Evaluating Scalable Uncertainty Estimation Methods for Deep Learning-Based Molecular Property Prediction.","volume":"60","author":"G Scalia","year":"2020","journal-title":"J Chem Inf Model."},{"issue":"6","key":"pdig.0000085.ref026","doi-asserted-by":"crossref","first-page":"2648","DOI":"10.1021\/acs.jcim.1c00208","article-title":"Deep Learning-Based Conformal Prediction of Toxicity.","volume":"61","author":"J Zhang","year":"2021","journal-title":"J Chem Inf Model."},{"issue":"7","key":"pdig.0000085.ref027","doi-asserted-by":"crossref","first-page":"3330","DOI":"10.1021\/acs.jcim.9b00297","article-title":"Reliable Prediction Errors for Deep Neural Networks Using Test-Time Dropout","volume":"59","author":"I Cortes-Ciriano","year":"2019","journal-title":"Journal of Chemical Information and Modeling"},{"key":"pdig.0000085.ref028","article-title":"StoCast: Stochastic Disease Forecasting with Progression Uncertainty","author":"X Teng","year":"2020","journal-title":"IEEE J Biomed Health Inform"},{"key":"pdig.0000085.ref029","doi-asserted-by":"crossref","DOI":"10.1016\/j.media.2020.101653","article-title":"Deep learning uncertainty and confidence calibration for the five-class polyp classification from colonoscopy","volume":"62","author":"G Carneiro","year":"2020","journal-title":"Medical Image Analysis"},{"issue":"9","key":"pdig.0000085.ref030","doi-asserted-by":"crossref","first-page":"2599","DOI":"10.1109\/JBHI.2020.2972694","article-title":"Coarse-to-Fine Adversarial Networks and Zone-Based Uncertainty Analysis for NK\/T-Cell Lymphoma Segmentation in CT\/PET Images","volume":"24","author":"X Hu","year":"2020","journal-title":"IEEE J Biomed Health Inform"},{"key":"pdig.0000085.ref031","doi-asserted-by":"crossref","first-page":"101724","DOI":"10.1016\/j.media.2020.101724","article-title":"Expert-validated estimation of diagnostic uncertainty for deep neural networks in diabetic retinopathy detection","volume":"64","author":"MS Ayhan","year":"2020","journal-title":"Med Image Anal"},{"issue":"1","key":"pdig.0000085.ref032","doi-asserted-by":"crossref","first-page":"431","DOI":"10.1109\/TMI.2020.3029161","article-title":"Uncertainty Aware Temporal-Ensembling Model for Semi-Supervised ABUS Mass Segmentation","volume":"40","author":"X Cao","year":"2021","journal-title":"IEEE Trans Med Imaging"},{"issue":"12","key":"pdig.0000085.ref033","doi-asserted-by":"crossref","first-page":"3431","DOI":"10.1109\/JBHI.2020.2983730","article-title":"UD-MIL: Uncertainty-Driven Deep Multiple Instance Learning for OCT Image Classification","volume":"24","author":"X Wang","year":"2020","journal-title":"IEEE J Biomed Health Inform"},{"key":"pdig.0000085.ref034","doi-asserted-by":"crossref","first-page":"101715","DOI":"10.1016\/j.media.2020.101715","article-title":"DR|GRADUATE: Uncertainty-aware deep learning-based diabetic retinopathy grading in eye fundus images","volume":"63","author":"T Araujo","year":"2020","journal-title":"Med Image Anal"},{"issue":"1","key":"pdig.0000085.ref035","doi-asserted-by":"crossref","first-page":"239","DOI":"10.1109\/TMI.2020.3025065","article-title":"Uncertainty Quantification in Deep MRI Reconstruction","volume":"40","author":"V Edupuganti","year":"2021","journal-title":"IEEE Trans Med Imaging"},{"key":"pdig.0000085.ref036","doi-asserted-by":"crossref","first-page":"101557","DOI":"10.1016\/j.media.2019.101557","article-title":"Exploring uncertainty measures in deep networks for Multiple sclerosis lesion detection and segmentation","volume":"59","author":"T Nair","year":"2020","journal-title":"Med Image Anal"},{"key":"pdig.0000085.ref037","doi-asserted-by":"crossref","first-page":"6","DOI":"10.3389\/fncom.2020.00006","article-title":"Demystifying Brain Tumor Segmentation Networks: Interpretability and Uncertainty Analysis.","volume":"14","author":"P Natekar","year":"2020","journal-title":"Front Comput Neurosci"},{"issue":"1","key":"pdig.0000085.ref038","doi-asserted-by":"crossref","first-page":"87","DOI":"10.1109\/TMI.2019.2919951","article-title":"Exploiting Epistemic Uncertainty of Anatomy Segmentation for Anomaly Detection in Retinal OCT","volume":"39","author":"P Seebock","year":"2020","journal-title":"IEEE Trans Med Imaging"},{"issue":"10","key":"pdig.0000085.ref039","doi-asserted-by":"crossref","first-page":"E549","DOI":"10.1016\/S2589-7500(20)30219-3","article-title":"Guidelines for clinical trial protocols for interventions involving artificial intelligence: the SPIRIT-AI extension","volume":"2","author":"SC Rivera","year":"2020","journal-title":"Lancet Digit Health"},{"issue":"9","key":"pdig.0000085.ref040","doi-asserted-by":"crossref","first-page":"1364","DOI":"10.1038\/s41591-020-1034-x","article-title":"Reporting guidelines for clinical trial reports for interventions involving artificial intelligence: the CONSORT-AI extension","volume":"26","author":"X Liu","year":"2020","journal-title":"Nat Med"},{"issue":"5","key":"pdig.0000085.ref041","doi-asserted-by":"crossref","first-page":"623","DOI":"10.1097\/CCM.0000000000004246","article-title":"Development and Reporting of Prediction Models: Guidance for Authors From Editors of Respiratory, Sleep, and Critical Care Journals.","volume":"48","author":"DE Leisman","year":"2020","journal-title":"Crit Care Med."},{"issue":"1","key":"pdig.0000085.ref042","doi-asserted-by":"crossref","first-page":"22","DOI":"10.1513\/AnnalsATS.201808-564PS","article-title":"Control of Confounding and Reporting of Results in Causal Inference Studies. Guidance for Authors from Editors of Respiratory, Sleep, and Critical Care Journals.","volume":"16","author":"DJ Lederer","year":"2019","journal-title":"Ann Am Thorac Soc."}],"container-title":["PLOS Digital Health"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/dx.plos.org\/10.1371\/journal.pdig.0000085","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2022,8,10]],"date-time":"2022-08-10T17:29:28Z","timestamp":1660152568000},"score":1,"resource":{"primary":{"URL":"https:\/\/dx.plos.org\/10.1371\/journal.pdig.0000085"}},"subtitle":[],"editor":[{"given":"Yuan","family":"Lai","sequence":"first","affiliation":[],"role":[{"vocabulary":"crossref","role":"editor"}]}],"short-title":[],"issued":{"date-parts":[[2022,8,10]]},"references-count":42,"journal-issue":{"issue":"8","published-online":{"date-parts":[[2022,8,10]]}},"URL":"https:\/\/doi.org\/10.1371\/journal.pdig.0000085","relation":{},"ISSN":["2767-3170"],"issn-type":[{"value":"2767-3170","type":"electronic"}],"subject":[],"published":{"date-parts":[[2022,8,10]]}}}