{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,5,26]],"date-time":"2026-05-26T16:05:53Z","timestamp":1779811553821,"version":"3.53.1"},"reference-count":34,"publisher":"Springer Science and Business Media LLC","license":[{"start":{"date-parts":[[2026,5,26]],"date-time":"2026-05-26T00:00:00Z","timestamp":1779753600000},"content-version":"tdm","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0"},{"start":{"date-parts":[[2026,5,26]],"date-time":"2026-05-26T00:00:00Z","timestamp":1779753600000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0"}],"funder":[{"DOI":"10.13039\/501100011699","name":"Siemens Healthineers","doi-asserted-by":"publisher","award":["LSHM20103"],"award-info":[{"award-number":["LSHM20103"]}],"id":[{"id":"10.13039\/501100011699","id-type":"DOI","asserted-by":"publisher"}]},{"DOI":"10.13039\/100016036","name":"Health~Holland","doi-asserted-by":"publisher","award":["LSHM20103"],"award-info":[{"award-number":["LSHM20103"]}],"id":[{"id":"10.13039\/100016036","id-type":"DOI","asserted-by":"publisher"}]}],"content-domain":{"domain":["link.springer.com"],"crossmark-restriction":false},"short-container-title":["Magn Reson Mater Phy"],"abstract":"<jats:title>Abstract<\/jats:title>\n                  <jats:sec>\n                    <jats:title>Purpose<\/jats:title>\n                    <jats:p>Although undersampling combined with deep learning (DL)-based reconstruction shortens MRI acquisition, it increases the chance of inaccuracies, highlighting the need for quantifiable uncertainty measures. Two inference-time perturbation strategies, echo-train dropout (ET-Drop) and Gaussian noise Monte Carlo sampling (GN-MC), were compared in terms of the correlation between their variance-based uncertainty maps and absolute reconstruction error in DL-accelerated T2w prostate MRI.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Methods<\/jats:title>\n                    <jats:p>\n                      This retrospective multi-center study used a publicly available dataset with 312\u00a0k-spaces from NYU for training and a dataset with 120\u00a0k-spaces from University Medical Center Groningen for external validation. Fully sampled 3\u00a0T data were retrospectively undersampled to acceleration factors R\u2009=\u20093 and R\u2009=\u20096 and reconstructed by a vSHARP model. Per slice, five GN-MC perturbations were reconstructed by adding complex noise at 2.5\u03c3, and five ET-Drop perturbations, created by omitting non-central echo trains. Voxel-wise aleatoric uncertainty was defined as the variance (\u03c3\n                      <jats:sup>2<\/jats:sup>\n                      ) across these reconstructions and correlated with absolute reconstruction error over whole slices and within the prostate.\n                    <\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Results<\/jats:title>\n                    <jats:p>\n                      Both uncertainties yielded moderate slice-level correlations with absolute error. At R\u2009=\u20093, ET-Drop slightly outperformed GN-MC (median\n                      <jats:italic>\u03c1<\/jats:italic>\n                      \u2009=\u20090.39 vs 0.35;\n                      <jats:italic>p<\/jats:italic>\n                      \u2009&lt;\u20090.001). At R\u2009=\u20096, the ranking reversed (0.44 vs 0.40;\n                      <jats:italic>p<\/jats:italic>\n                      \u2009&lt;\u20090.001). Correlations within the prostate fell to 0.10\u20130.15. ET-Drop variance maps were dominated by coil sensitivities.\n                    <\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Conclusion<\/jats:title>\n                    <jats:p>Both perturbation strategies yield variance-based uncertainty maps that correlate moderately with voxel-wise error. More importantly, they consistently highlighted acquisition-related fragility, supporting the role of uncertainty mapping as a useful quality-control tool in prostate MRI.<\/jats:p>\n                  <\/jats:sec>","DOI":"10.1007\/s10334-026-01369-0","type":"journal-article","created":{"date-parts":[[2026,5,26]],"date-time":"2026-05-26T15:35:33Z","timestamp":1779809733000},"update-policy":"https:\/\/doi.org\/10.1007\/springer_crossmark_policy","source":"Crossref","is-referenced-by-count":0,"title":["Aleatoric uncertainty in accelerated prostate MRI reconstruction: echo-train dropout versus Gaussian noise Monte Carlo sampling"],"prefix":"10.1007","author":[{"ORCID":"https:\/\/orcid.org\/0000-0002-7128-0818","authenticated-orcid":false,"given":"Quintin","family":"van Lohuizen","sequence":"first","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Stefan J.","family":"Fransen","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Henkjan","family":"Huisman","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Jelmer M.","family":"Wolterink","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Thomas C.","family":"Kwee","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Derya","family":"Yakar","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Frank F. J.","family":"Simonis","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"297","published-online":{"date-parts":[[2026,5,26]]},"reference":[{"issue":"4","key":"1369_CR1","doi-asserted-by":"publisher","first-page":"887","DOI":"10.1002\/jmri.24850","volume":"42","author":"M Zaitsev","year":"2015","unstructured":"Zaitsev M, Maclaren J, Herbst M (2015) Motion artifacts in MRI: a complex problem with many partial solutions. J Magn Reson Imaging 42(4):887\u2013901. https:\/\/doi.org\/10.1002\/jmri.24850","journal-title":"J Magn Reson Imaging"},{"issue":"6","key":"1369_CR2","doi-asserted-by":"publisher","first-page":"1202","DOI":"10.1002\/mrm.10171","volume":"47","author":"MA Griswold","year":"2002","unstructured":"Griswold MA, Jakob PM, Heidemann RM, Nittka M, Jellus V, Wang J, Kiefer B, Haase A (2002) Generalized autocalibrating partially parallel acquisitions (GRAPPA). Magn Reson Med 47(6):1202\u20131210. https:\/\/doi.org\/10.1002\/mrm.10171","journal-title":"Magn Reson Med"},{"issue":"6","key":"1369_CR3","doi-asserted-by":"publisher","first-page":"1182","DOI":"10.1002\/mrm.21391","volume":"58","author":"M Lustig","year":"2007","unstructured":"Lustig M, Donoho D, Pauly JM (2007) Sparse MRI: the application of compressed sensing for rapid MR imaging. Magn Reson Med 58(6):1182\u20131195. https:\/\/doi.org\/10.1002\/mrm.21391","journal-title":"Magn Reson Med"},{"issue":"48","key":"1369_CR4","doi-asserted-by":"publisher","first-page":"30088","DOI":"10.1073\/pnas.1907377117","volume":"117","author":"V Antun","year":"2020","unstructured":"Antun V, Renna F, Poon C, Adcock B, Hansen AC (2020) On instabilities of deep learning in image reconstruction and the potential costs of AI. Proc Natl Acad Sci U S A 117(48):30088\u201330095. https:\/\/doi.org\/10.1073\/pnas.1907377117","journal-title":"Proc Natl Acad Sci U S A"},{"issue":"2","key":"1369_CR5","doi-asserted-by":"publisher","first-page":"394","DOI":"10.1109\/TMI.2018.2865356","volume":"38","author":"HK Aggarwal","year":"2019","unstructured":"Aggarwal HK, Mani MP, Jacob M (2019) MoDL: model-based deep learning architecture for inverse problems. IEEE Trans Med Imaging 38(2):394\u2013405. https:\/\/doi.org\/10.1109\/TMI.2018.2865356","journal-title":"IEEE Trans Med Imaging"},{"issue":"6","key":"1369_CR6","doi-asserted-by":"publisher","first-page":"3055","DOI":"10.1002\/mrm.26977","volume":"79","author":"K Hammernik","year":"2018","unstructured":"Hammernik K, Klatzer T, Kobler E et al (2018) Learning a variational network for reconstruction of accelerated MRI data. Magn Reson Med 79(6):3055\u20133071. https:\/\/doi.org\/10.1002\/mrm.26977","journal-title":"Magn Reson Med"},{"issue":"1","key":"1369_CR7","doi-asserted-by":"publisher","first-page":"13710","DOI":"10.1038\/s41598-020-70551-8","volume":"10","author":"T K\u00fcstner","year":"2020","unstructured":"K\u00fcstner T, Fuin N, Hammernik K et al (2020) CINENet: deep learning-based 3D cardiac CINE MRI reconstruction with multi-coil complex-valued 4D spatio-temporal convolutions. Sci Rep 10(1):13710. https:\/\/doi.org\/10.1038\/s41598-020-70551-8","journal-title":"Sci Rep"},{"issue":"2","key":"1369_CR8","doi-asserted-by":"publisher","first-page":"491","DOI":"10.1109\/TMI.2017.2760978","volume":"37","author":"J Schlemper","year":"2018","unstructured":"Schlemper J, Caballero J, Hajnal JV, Price AN, Rueckert D (2018) A deep cascade of convolutional neural networks for dynamic MR image reconstruction. IEEE Trans Med Imaging 37(2):491\u2013503. https:\/\/doi.org\/10.1109\/TMI.2017.2760978","journal-title":"IEEE Trans Med Imaging"},{"key":"1369_CR9","doi-asserted-by":"publisher","DOI":"10.1016\/j.mri.2024.110266","volume":"115","author":"G Yiasemis","year":"2025","unstructured":"Yiasemis G, Moriakov N, Sonke JJ, Teuwen J (2025) vSHARP: variable splitting half-quadratic ADMM algorithm for reconstruction of inverse-problems. Magn Reson Imaging 115:110266. https:\/\/doi.org\/10.1016\/j.mri.2024.110266","journal-title":"Magn Reson Imaging"},{"issue":"1","key":"1369_CR10","doi-asserted-by":"publisher","first-page":"439","DOI":"10.1002\/mrm.27420","volume":"81","author":"M Ak\u00e7akaya","year":"2019","unstructured":"Ak\u00e7akaya M, Moeller S, Weing\u00e4rtner S, U\u011furbil K (2019) Scan-specific robust artificial-neural-networks for k-space interpolation (RAKI) reconstruction: database-free deep learning for fast imaging. Magn Reson Med 81(1):439\u2013453. https:\/\/doi.org\/10.1002\/mrm.27420","journal-title":"Magn Reson Med"},{"issue":"6","key":"1369_CR11","doi-asserted-by":"publisher","first-page":"1848","DOI":"10.1002\/mrm.26081","volume":"76","author":"D Lee","year":"2016","unstructured":"Lee D, Jin KH, Kim EY, Park SH, Ye JC (2016) Acceleration of MR parameter mapping using annihilating filter-based low rank Hankel matrix (ALOHA). Magn Reson Med 76(6):1848\u20131864. https:\/\/doi.org\/10.1002\/mrm.26081","journal-title":"Magn Reson Med"},{"issue":"6","key":"1369_CR12","doi-asserted-by":"publisher","DOI":"10.1148\/ryai.210313","volume":"4","author":"A Radmanesh","year":"2022","unstructured":"Radmanesh A, Muckley MJ, Murrell T et al (2022) Exploring the acceleration limits of deep learning variational network-based two-dimensional brain MRI. Radiol Artif Intell 4(6):e210313. https:\/\/doi.org\/10.1148\/ryai.210313","journal-title":"Radiol Artif Intell"},{"issue":"1","key":"1369_CR13","doi-asserted-by":"publisher","first-page":"128","DOI":"10.1109\/MSP.2019.2950640","volume":"37","author":"F Knoll","year":"2020","unstructured":"Knoll F, Hammernik K, Zhang C et al (2020) Deep-learning methods for parallel magnetic resonance imaging reconstruction: a survey of the current approaches, trends, and issues. IEEE Signal Process Mag 37(1):128\u2013140. https:\/\/doi.org\/10.1109\/MSP.2019.2950640","journal-title":"IEEE Signal Process Mag"},{"key":"1369_CR14","unstructured":"Kendall A, Gal Y. What uncertainties do we need in Bayesian deep learning for computer vision? arXiv [Preprint]. 2017. Available from: https:\/\/arxiv.org\/abs\/1703.04977"},{"issue":"2","key":"1369_CR15","doi-asserted-by":"publisher","first-page":"279","DOI":"10.1109\/TMI.2021.3112040","volume":"41","author":"D Narnhofer","year":"2022","unstructured":"Narnhofer D, Effland A, Kobler E, Hammernik K, Knoll F, Pock T (2022) Bayesian uncertainty estimation of learned variational MRI reconstruction. IEEE Trans Med Imaging 41(2):279\u2013291. https:\/\/doi.org\/10.1109\/TMI.2021.3112040","journal-title":"IEEE Trans Med Imaging"},{"key":"1369_CR16","unstructured":"Riquelme C, Tucker G, Snoek J. Deep Bayesian bandits showdown: an empirical comparison of Bayesian deep networks for Thompson sampling. arXiv [Preprint]. 2018. Available from: https:\/\/arxiv.org\/abs\/1802.09127"},{"key":"1369_CR17","doi-asserted-by":"crossref","unstructured":"Huttinga NR, Bruijnen T, van den Berg CA, Sbrizzi A. Gaussian processes for real-time 3D motion and uncertainty estimation during MR-guided radiotherapy. arXiv [Preprint]. 2022. Available from: https:\/\/arxiv.org\/abs\/2204.09873","DOI":"10.1016\/j.media.2023.102843"},{"key":"1369_CR18","doi-asserted-by":"publisher","DOI":"10.1109\/ACCESS.2022.3195211","volume":"10","author":"M Frueh","year":"2022","unstructured":"Frueh M, Schilling A, Gatidis S, Kuestner T (2022) Real time landmark detection for within- and cross subject tracking with minimal human supervision. IEEE Access 10:81192-202. https:\/\/doi.org\/10.1109\/ACCESS.2022.3195211","journal-title":"IEEE Access"},{"key":"1369_CR19","doi-asserted-by":"publisher","DOI":"10.1016\/j.neuroimage.2020.117366","volume":"225","author":"R Tanno","year":"2021","unstructured":"Tanno R, Worrall DE, Kaden E et al (2021) Uncertainty modelling in deep learning for safer neuroimage enhancement: demonstration in diffusion MRI. Neuroimage 225:117366. https:\/\/doi.org\/10.1016\/j.neuroimage.2020.117366","journal-title":"Neuroimage"},{"issue":"1","key":"1369_CR20","doi-asserted-by":"publisher","first-page":"239","DOI":"10.1109\/TMI.2020.3025065","volume":"40","author":"V Edupuganti","year":"2021","unstructured":"Edupuganti V, Mardani M, Vasanawala S, Pauly J (2021) Uncertainty quantification in deep MRI reconstruction. IEEE Trans Med Imaging 40(1):239\u2013250. https:\/\/doi.org\/10.1109\/TMI.2020.3025065","journal-title":"IEEE Trans Med Imaging"},{"key":"1369_CR21","unstructured":"Gal Y, Ghahramani Z. Dropout as a Bayesian approximation: representing model uncertainty in deep learning. Proceedings of the 33rd International Conference on Machine Learning. PMLR; 2016:1050\u20131059."},{"key":"1369_CR22","doi-asserted-by":"crossref","unstructured":"Schlemper J, Castro DC, Bai W, et al. Bayesian deep learning for accelerated MR image reconstruction. Proceedings of the International Workshop on Machine Learning for Medical Image Reconstruction. Springer; 2018:64\u201371.","DOI":"10.1007\/978-3-030-00129-2_8"},{"key":"1369_CR23","unstructured":"Avci MY, Li Z, Fan Q, Huang SY, Bilgic B, Tian Q. Improving accuracy and uncertainty quantification of deep learning based quantitative MRI using Monte Carlo dropout. arXiv [Preprint]. 2021. Available from: https:\/\/arxiv.org\/abs\/2112.01587"},{"key":"1369_CR24","unstructured":"Lakshminarayanan B, Pritzel A, Blundell C. Simple and scalable predictive uncertainty estimation using deep ensembles. arXiv [Preprint]. 2016. Available from: https:\/\/arxiv.org\/abs\/1612.01474"},{"issue":"1","key":"1369_CR25","doi-asserted-by":"publisher","first-page":"289","DOI":"10.1002\/mrm.30030","volume":"92","author":"T K\u00fcstner","year":"2024","unstructured":"K\u00fcstner T, Hammernik K, Rueckert D, Hepp T, Gatidis S (2024) Predictive uncertainty in deep learning-based MR image reconstruction using deep ensembles: evaluation on the fastMRI data set. Magn Reson Med 92(1):289\u2013302. https:\/\/doi.org\/10.1002\/mrm.30030","journal-title":"Magn Reson Med"},{"key":"1369_CR26","doi-asserted-by":"publisher","unstructured":"Yiasemis G, Moriakov N, Sonke J-J, Teuwen J. Deep Cardiac MRI Reconstruction with ADMM. In: Camara O, et al., eds. Statistical Atlases and Computational Models of the Heart. Regular and CMRxRecon Challenge Papers. Lecture Notes in Computer Science, 14507. Springer; 2024:386\u2013394. https:\/\/doi.org\/10.1007\/978-3-031-52448-6_45.","DOI":"10.1007\/978-3-031-52448-6_45"},{"key":"1369_CR27","doi-asserted-by":"publisher","DOI":"10.1016\/j.media.2025.103485","volume":"101","author":"J Lyu","year":"2025","unstructured":"Lyu J, Qin C, Wang S et al (2025) The state-of-the-art in cardiac MRI reconstruction: results of the CMRxRecon challenge in MICCAI 2023. Med Image Anal 101:103485. https:\/\/doi.org\/10.1016\/j.media.2025.103485","journal-title":"Med Image Anal"},{"key":"1369_CR28","doi-asserted-by":"publisher","first-page":"404","DOI":"10.1038\/s41597-024-03252-w","volume":"11","author":"R Tibrewala","year":"2024","unstructured":"Tibrewala R, Dutt T, Tong A et al (2024) FastMRI Prostate: A public, biparametric MRI dataset to advance machine learning for prostate cancer imaging. Sci Data 11:404. https:\/\/doi.org\/10.1038\/s41597-024-03252-w","journal-title":"Sci Data"},{"key":"1369_CR29","doi-asserted-by":"publisher","first-page":"33","DOI":"10.1016\/j.mri.2023.12.012","volume":"107","author":"G Yiasemis","year":"2024","unstructured":"Yiasemis G, S\u00e1nchez CI, Sonke JJ, Teuwen J (2024) On retrospective k-space subsampling schemes for deep MRI reconstruction. Magn Reson Imaging 107:33\u201346. https:\/\/doi.org\/10.1016\/j.mri.2023.12.012","journal-title":"Magn Reson Imaging"},{"key":"1369_CR30","doi-asserted-by":"publisher","DOI":"10.1016\/j.ejrai.2025.100017","volume":"2","author":"Q van Lohuizen","year":"2025","unstructured":"van Lohuizen Q, Fransen SJ, Yiasemis G et al (2025) Deep learning reconstruction for prostate MRI: impact of field-of-view selection on global and regional image quality. Eur J Radiol Artif Intell 2:100017. https:\/\/doi.org\/10.1016\/j.ejrai.2025.100017","journal-title":"Eur J Radiol Artif Intell"},{"issue":"2","key":"1369_CR31","doi-asserted-by":"publisher","DOI":"10.1148\/radiol.241613","volume":"314","author":"T Akinci D\u2019Antonoli","year":"2025","unstructured":"Akinci D\u2019Antonoli T, Berger LK, Indrakanti AK et al (2025) TotalSegmentator MRI: robust sequence-independent segmentation of multiple anatomic structures in MRI. Radiology 314(2):e241613. https:\/\/doi.org\/10.1148\/radiol.241613","journal-title":"Radiology"},{"issue":"2","key":"1369_CR32","doi-asserted-by":"publisher","first-page":"203","DOI":"10.1038\/s41592-020-01008-z","volume":"18","author":"F Isensee","year":"2021","unstructured":"Isensee F, Jaeger PF, Kohl SAA, Petersen J, Maier-Hein KH (2021) nnU-Net: a self-configuring method for deep learning-based biomedical image segmentation. Nat Methods 18(2):203\u2013211. https:\/\/doi.org\/10.1038\/s41592-020-01008-z","journal-title":"Nat Methods"},{"issue":"5","key":"1369_CR33","doi-asserted-by":"publisher","first-page":"952","DOI":"10.1002\/(SICI)1522-2594(199911)42:5<952::AID-MRM16>3.0.CO;2-S","volume":"42","author":"KP Pruessmann","year":"1999","unstructured":"Pruessmann KP, Weiger M, Scheidegger MB, Boesiger P (1999) SENSE: sensitivity encoding for fast MRI. Magn Reson Med 42(5):952\u2013962","journal-title":"Magn Reson Med"},{"issue":"11","key":"1369_CR34","doi-asserted-by":"publisher","first-page":"6955","DOI":"10.1002\/mp.16543","volume":"50","author":"S Brahma","year":"2023","unstructured":"Brahma S, Kolbitsch C, Martin J, Schaeffter T, Kofler A (2023) Data-efficient Bayesian learning for radial dynamic MR reconstruction. Med Phys 50(11):6955\u20136977. https:\/\/doi.org\/10.1002\/mp.16543","journal-title":"Med Phys"}],"container-title":["Magnetic Resonance Materials in Physics, Biology and Medicine"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/link.springer.com\/content\/pdf\/10.1007\/s10334-026-01369-0.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/link.springer.com\/article\/10.1007\/s10334-026-01369-0","content-type":"text\/html","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/link.springer.com\/content\/pdf\/10.1007\/s10334-026-01369-0.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,5,26]],"date-time":"2026-05-26T15:35:37Z","timestamp":1779809737000},"score":1,"resource":{"primary":{"URL":"https:\/\/link.springer.com\/10.1007\/s10334-026-01369-0"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,5,26]]},"references-count":34,"alternative-id":["1369"],"URL":"https:\/\/doi.org\/10.1007\/s10334-026-01369-0","relation":{},"ISSN":["1352-8661"],"issn-type":[{"value":"1352-8661","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,5,26]]},"assertion":[{"value":"27 November 2025","order":1,"name":"received","label":"Received","group":{"name":"ArticleHistory","label":"Article History"}},{"value":"30 April 2026","order":2,"name":"revised","label":"Revised","group":{"name":"ArticleHistory","label":"Article History"}},{"value":"1 May 2026","order":3,"name":"accepted","label":"Accepted","group":{"name":"ArticleHistory","label":"Article History"}},{"value":"26 May 2026","order":4,"name":"first_online","label":"First Online","group":{"name":"ArticleHistory","label":"Article History"}},{"order":1,"name":"Ethics","group":{"name":"EthicsHeading","label":"Declarations"}},{"value":"The authors have no relevant financial or non-financial interests to disclose.","order":2,"name":"Ethics","group":{"name":"EthicsHeading","label":"Conflict of interest"}},{"value":"This retrospective study was approved by the Medical Ethical Review Board of the University Medical Center Groningen (METc 2022\/082).","order":3,"name":"Ethics","group":{"name":"EthicsHeading","label":"Ethical approval"}},{"value":"The requirement for informed consent was waived by the ethics committee due to the retrospective nature of the study.","order":4,"name":"Ethics","group":{"name":"EthicsHeading","label":"Informed consent"}}]}}