{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,7,29]],"date-time":"2026-07-29T13:39:25Z","timestamp":1785332365917,"version":"3.55.0"},"reference-count":33,"publisher":"Wiley","issue":"2","license":[{"start":{"date-parts":[[2024,12,23]],"date-time":"2024-12-23T00:00:00Z","timestamp":1734912000000},"content-version":"vor","delay-in-days":0,"URL":"http:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"content-domain":{"domain":["aapm.onlinelibrary.wiley.com"],"crossmark-restriction":true},"short-container-title":["J Applied Clin Med Phys"],"published-print":{"date-parts":[[2025,2]]},"abstract":"<jats:title>Abstract<\/jats:title><jats:sec><jats:title>Purpose<\/jats:title><jats:p>To quantitatively evaluate the performance of two types of recurrent neural networks (RNNs), long short\u2010term memory (LSTM) and gated recurrent units (GRU), using Monte Carlo dropout (MCD) to predict pharmacokinetic (PK) parameters from dynamic contrast\u2010enhanced magnetic resonance imaging (DCE\u2010MRI) data.<\/jats:p><\/jats:sec><jats:sec><jats:title>Methods<\/jats:title><jats:p>DCE\u2010MRI data for simulation studies were synthesized using the extended Tofts model and a population\u2010averaged arterial input function (AIF). The ranges of PK parameters for training the RNNs were determined from data of patients with brain tumors. The effects of the number of training samples, number of hidden units, dropout rate (DR), and bolus arrival time delay and dispersion in AIF on the accuracy of the PK parameters were investigated, and the uncertainties for different DRs and peak signal\u2010to\u2010noise ratios (PSNRs) were quantified. For comparison, PK parameters were estimated using the nonlinear least\u2010squares method. In the clinical studies, the PK parameter and uncertainty images were generated by applying the trained RNNs to DCE\u2010MRI data.<\/jats:p><\/jats:sec><jats:sec><jats:title>Results<\/jats:title><jats:p>Compared with GRU, the computational cost for training the LSTM was significantly higher. The prediction accuracy of GRU decreased with decreasing numbers of training samples and hidden units, whereas the performance of LSTM remained stable. Despite an increased computational cost, MCD reduced the prediction error at low PSNR and improved the quality of PK parameter images. The simulation results recommended using a DR of 0.25\u20130.5 at low PSNR and \u2264 0.25 for other PSNRs. The clinical studies recommended using a DR of 0.25 and 0.5 for LSTM and GRU, respectively.<\/jats:p><\/jats:sec><jats:sec><jats:title>Conclusions<\/jats:title><jats:p>MCD is effective in quantifying uncertainty in PK parameter prediction from DCE\u2010MRI data and improves their performance, particularly at low PSNR; however, at the expense of increased computational cost. This study helps deepen our understanding of RNNs with MCD and select suitable hyperparameters for creating an RNN architecture for DCE\u2010MRI studies.<\/jats:p><\/jats:sec>","DOI":"10.1002\/acm2.14586","type":"journal-article","created":{"date-parts":[[2024,12,24]],"date-time":"2024-12-24T07:41:41Z","timestamp":1735026101000},"update-policy":"https:\/\/doi.org\/10.1002\/crossmark_policy","source":"Crossref","is-referenced-by-count":1,"title":["Performance of recurrent neural networks with Monte Carlo dropout for predicting pharmacokinetic parameters from dynamic contrast\u2010enhanced magnetic resonance imaging data"],"prefix":"10.1002","volume":"26","author":[{"given":"Kenya","family":"Murase","sequence":"first","affiliation":[{"name":"Department of Future Diagnostic Radiology Graduate School of Medicine Osaka University  Suita Osaka Japan"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Atsushi","family":"Nakamoto","sequence":"additional","affiliation":[{"name":"Department of Future Diagnostic Radiology Graduate School of Medicine Osaka University  Suita Osaka Japan"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Noriyuki","family":"Tomiyama","sequence":"additional","affiliation":[{"name":"Department of Diagnostic and Interventional Radiology Graduate School of Medicine Osaka University  Suita Osaka Japan"}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"311","published-online":{"date-parts":[[2024,12,23]]},"reference":[{"key":"e_1_2_10_2_1","doi-asserted-by":"publisher","DOI":"10.1002\/jmri.10176"},{"key":"e_1_2_10_3_1","doi-asserted-by":"publisher","DOI":"10.1002\/nbm.1273"},{"key":"e_1_2_10_4_1","doi-asserted-by":"publisher","DOI":"10.1038\/nrclinonc.2012.2"},{"key":"e_1_2_10_5_1","doi-asserted-by":"publisher","DOI":"10.1002\/(SICI)1522-2586(199909)10:3<223::AID-JMRI2>3.0.CO;2-S"},{"key":"e_1_2_10_6_1","doi-asserted-by":"publisher","DOI":"10.1088\/0031-9155\/50\/9\/N02"},{"key":"e_1_2_10_7_1","doi-asserted-by":"publisher","DOI":"10.1002\/mrm.20022"},{"key":"e_1_2_10_8_1","doi-asserted-by":"publisher","DOI":"10.3389\/fneur.2018.01147"},{"key":"e_1_2_10_9_1","doi-asserted-by":"publisher","DOI":"10.1109\/TMI.2019.2953901"},{"key":"e_1_2_10_10_1","doi-asserted-by":"publisher","DOI":"10.1002\/mp.14222"},{"key":"e_1_2_10_11_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.media.2022.102512"},{"key":"e_1_2_10_12_1","doi-asserted-by":"publisher","DOI":"10.1162\/neco.1997.9.8.1735"},{"key":"e_1_2_10_13_1","doi-asserted-by":"crossref","unstructured":"ChoK vanMerrienboerB BahdanauD BougaresF SchwenkH BengioY.Learning phrase representations using RNN encoder\u2010decoder for statistical machine translation. arXiv preprint arXiv:1406.1078 2014.10.48550\/arXiv.1406.1078","DOI":"10.3115\/v1\/D14-1179"},{"key":"e_1_2_10_14_1","unstructured":"ChungJ GulcehreC ChoK BengioY.Empirical evaluation of gated recurrent neural networks on sequence modeling. arXiv preprint arXiv:1412.3555 2014.10.48550\/arXiv.1412.3555"},{"key":"e_1_2_10_15_1","doi-asserted-by":"publisher","DOI":"10.1038\/s41467\u2010023\u201043317\u20109"},{"key":"e_1_2_10_16_1","doi-asserted-by":"publisher","DOI":"10.1088\/1361\u20106560\/ad0284"},{"issue":"56","key":"e_1_2_10_17_1","first-page":"1929","article-title":"Dropout: a simple way to prevent neural networks from overfitting","volume":"15","author":"Srivastava N","year":"2014","journal-title":"J Mach Learn Res"},{"key":"e_1_2_10_18_1","unstructured":"GalY GhahramaniZ.Dropout as a Bayesian approximation: representing model uncertainty in deep learning. arXiv preprint arXiv:1506.02142 2015.10.48550\/arXiv.1506.02142"},{"key":"e_1_2_10_19_1","doi-asserted-by":"publisher","DOI":"10.1002\/mrm.29909"},{"key":"e_1_2_10_20_1","doi-asserted-by":"publisher","DOI":"10.1007\/s00330-015-4012-9"},{"key":"e_1_2_10_21_1","doi-asserted-by":"publisher","DOI":"10.1002\/jmri.27495"},{"key":"e_1_2_10_22_1","doi-asserted-by":"publisher","DOI":"10.1002\/mrm.20873"},{"key":"e_1_2_10_23_1","unstructured":"KingmaDP BaJ.Adam: a method for stochastic optimization. arXiv preprint arXiv:1412.6980 2014.10.48550\/arXiv.1412.6980"},{"key":"e_1_2_10_24_1","doi-asserted-by":"publisher","DOI":"10.2307\/2532051"},{"key":"e_1_2_10_25_1","doi-asserted-by":"publisher","DOI":"10.1002\/mrm.25135"},{"key":"e_1_2_10_26_1","doi-asserted-by":"publisher","DOI":"10.1016\/0730-725X(87)90021-X"},{"key":"e_1_2_10_27_1","doi-asserted-by":"publisher","DOI":"10.1097\/01.rli.0000184756.66360.d3"},{"issue":"1","key":"e_1_2_10_28_1","first-page":"46","article-title":"Influence of partial volume effect on venous output and arterial input function","volume":"27","author":"Schaaf I","year":"2006","journal-title":"AJNR Am J Neuroradiol"},{"key":"e_1_2_10_29_1","unstructured":"CahuantziR ChenX G\u00fcttelS.A comparison of LSTM and GRU networks for learning symbolic sequences. arXiv preprint arXiv:2107.02248 2021.10.48550\/arXiv.2107.02248"},{"key":"e_1_2_10_30_1","unstructured":"VerdojaF KyrkiV.Notes on the Behavior of MC Dropout. arXiv preprint arXiv:2008.02627.2020.10.48550\/arXiv.2008.02627"},{"key":"e_1_2_10_31_1","doi-asserted-by":"publisher","DOI":"10.1088\/0031\u20109155\/46\/12\/306"},{"key":"e_1_2_10_32_1","doi-asserted-by":"publisher","DOI":"10.1002\/mrm.21066"},{"key":"e_1_2_10_33_1","doi-asserted-by":"publisher","DOI":"10.5152\/dir.2022.19512"},{"key":"e_1_2_10_34_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.artmed.2019.101769"}],"container-title":["Journal of Applied Clinical Medical Physics"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/aapm.onlinelibrary.wiley.com\/doi\/pdf\/10.1002\/acm2.14586","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2025,4,12]],"date-time":"2025-04-12T04:50:20Z","timestamp":1744433420000},"score":1,"resource":{"primary":{"URL":"https:\/\/aapm.onlinelibrary.wiley.com\/doi\/10.1002\/acm2.14586"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2024,12,23]]},"references-count":33,"journal-issue":{"issue":"2","published-print":{"date-parts":[[2025,2]]}},"alternative-id":["10.1002\/acm2.14586"],"URL":"https:\/\/doi.org\/10.1002\/acm2.14586","archive":["Portico"],"relation":{},"ISSN":["1526-9914","1526-9914"],"issn-type":[{"value":"1526-9914","type":"print"},{"value":"1526-9914","type":"electronic"}],"subject":[],"published":{"date-parts":[[2024,12,23]]},"assertion":[{"value":"2024-07-26","order":0,"name":"received","label":"Received","group":{"name":"publication_history","label":"Publication History"}},{"value":"2024-11-14","order":2,"name":"accepted","label":"Accepted","group":{"name":"publication_history","label":"Publication History"}},{"value":"2024-12-23","order":3,"name":"published","label":"Published","group":{"name":"publication_history","label":"Publication History"}}],"article-number":"e14586"}}