{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,10,17]],"date-time":"2026-10-17T05:10:11Z","timestamp":1792213811335,"version":"4.4.0"},"reference-count":29,"publisher":"Georg Thieme Verlag KG","issue":"05","content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":["Rofo"],"published-print":{"date-parts":[[2021,5]]},"abstract":"<jats:p>\n            Purpose\u2002A recently developed deep learning model (U-Net) approximated the clinical performance of radiologists in the prediction of clinically significant prostate cancer (sPC) from prostate MRI. Here, we compare the agreement between lesion segmentations by U-Net with manual lesion segmentations performed by different radiologists.<\/jats:p><jats:p>\n            Materials and Methods\u2002165 patients with suspicion for sPC underwent targeted and systematic fusion biopsy following 3\u00a0Tesla multiparametric MRI (mpMRI). Five sets of segmentations were generated retrospectively: segmentations of clinical lesions, independent segmentations by three radiologists, and fully automated bi-parametric U-Net segmentations. Per-lesion agreement was calculated for each rater by averaging Dice coefficients with all overlapping lesions from other raters. Agreement was compared using descriptive statistics and linear mixed models.<\/jats:p><jats:p>\n            Results\u2002The mean Dice coefficient for manual segmentations showed only moderate agreement at 0.48\u20130.52, reflecting the difficult visual task of determining the outline of otherwise jointly detected lesions. U-net segmentations were significantly smaller than manual segmentations (p\u200a&lt;\u200a0.0001) and exhibited a lower mean Dice coefficient of 0.22, which was significantly lower compared to manual segmentations (all p\u200a&lt;\u200a0.0001). These differences remained after correction for lesion size and were unaffected between sPC and non-sPC lesions and between peripheral and transition zone lesions.<\/jats:p><jats:p>\n            Conclusion\u2002Knowledge of the order of agreement of manual segmentations of different radiologists is important to set the expectation value for artificial intelligence (AI) systems in the task of prostate MRI lesion segmentation. Perfect agreement (Dice coefficient of one) should not be expected for AI. Lower Dice coefficients of U-Net compared to manual segmentations are only partially explained by smaller segmentation sizes and may result from a focus on the lesion core and a small relative lesion center shift. Although it is primarily important that AI detects sPC correctly, the Dice coefficient for overlapping lesions from multiple raters can be used as a secondary measure for segmentation quality in future studies.<\/jats:p><jats:p>\n            Key Points:\u2002<\/jats:p><jats:p>\n            Citation Format\n          <\/jats:p>","DOI":"10.1055\/a-1290-8070","type":"journal-article","created":{"date-parts":[[2020,11,20]],"date-time":"2020-11-20T01:59:21Z","timestamp":1605837561000},"page":"559-573","source":"Crossref","is-referenced-by-count":24,"title":["Comparison of Prostate MRI Lesion Segmentation Agreement Between Multiple Radiologists and a Fully Automatic Deep Learning System"],"prefix":"10.1055","volume":"193","author":[{"given":"Patrick","family":"Schelb","sequence":"additional","affiliation":[{"name":"Division of Radiology, German Cancer Research Center (DKFZ), Heidelberg, Germany"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Anoshirwan Andrej","family":"Tavakoli","sequence":"additional","affiliation":[{"name":"Division of Radiology, German Cancer Research Center (DKFZ), Heidelberg, Germany"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Teeravut","family":"Tubtawee","sequence":"additional","affiliation":[{"name":"Division of Radiology, German Cancer Research Center (DKFZ), Heidelberg, Germany"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Thomas","family":"Hielscher","sequence":"additional","affiliation":[{"name":"Division of Biostatistics, German Cancer Research Center (DKFZ), Heidelberg, Germany"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Jan-Philipp","family":"Radtke","sequence":"additional","affiliation":[{"name":"Department of Urology, University of Heidelberg Medical Center, Heidelberg, Germany"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Magdalena","family":"G\u00f6rtz","sequence":"additional","affiliation":[{"name":"Department of Urology, University of Heidelberg Medical Center, Heidelberg, Germany"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Viktoria","family":"Sch\u00fctz","sequence":"additional","affiliation":[{"name":"Department of Urology, University of Heidelberg Medical Center, Heidelberg, Germany"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Tristan Anselm","family":"Kuder","sequence":"additional","affiliation":[{"name":"Division of Medical Physics, German Cancer Research Center (DKFZ), Heidelberg, Germany"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-7476-292X","authenticated-orcid":false,"given":"Lars","family":"Schimm\u00f6ller","sequence":"additional","affiliation":[{"name":"University Dusseldorf, Medical Faculty, Department of\u00a0Diagnostic and Interventional Radiology, Dusseldorf, Germany"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Albrecht","family":"Stenzinger","sequence":"additional","affiliation":[{"name":"Institute of Pathology, University of Heidelberg Medical Center, Heidelberg, Germany"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Markus","family":"Hohenfellner","sequence":"additional","affiliation":[{"name":"Department of Urology, University of Heidelberg Medical Center, Heidelberg, Germany"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Heinz-Peter","family":"Schlemmer","sequence":"additional","affiliation":[{"name":"Division of Radiology, German Cancer Research Center (DKFZ), Heidelberg, Germany"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-4811-0087","authenticated-orcid":false,"given":"David","family":"Bonekamp","sequence":"additional","affiliation":[{"name":"Division of Radiology, German Cancer Research Center (DKFZ), Heidelberg, Germany"}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"194","published-online":{"date-parts":[[2020,11,19]]},"reference":[{"key":"ref1","doi-asserted-by":"publisher","first-page":"87","DOI":"10.1016\/j.juro.2014.07.098","article-title":"Comparative analysis of transperineal template saturation prostate biopsy versus magnetic resonance imaging targeted biopsy with magnetic resonance imaging-ultrasound fusion guidance","volume":"193","author":"J P Radtke","year":"2015","journal-title":"J Urol"},{"key":"ref2","doi-asserted-by":"publisher","first-page":"713","DOI":"10.1016\/j.eururo.2013.05.059","article-title":"Magnetic resonance imaging\/ultrasound-fusion biopsy significantly upgrades prostate cancer versus systematic 12-core transrectal ultrasound biopsy","volume":"64","author":"M M Siddiqui","year":"2013","journal-title":"Eur Urol"},{"key":"ref3","doi-asserted-by":"publisher","first-page":"815","DOI":"10.1016\/S0140-6736(16)32401-1","article-title":"Diagnostic accuracy of multi-parametric MRI and TRUS biopsy in prostate cancer (PROMIS): a paired validating confirmatory study","volume":"389","author":"H U Ahmed","year":"2017","journal-title":"Lancet (London, England)"},{"key":"ref4","doi-asserted-by":"publisher","first-page":"1767","DOI":"10.1056\/NEJMoa1801993","article-title":"MRI-Targeted or Standard Biopsy for Prostate-Cancer Diagnosis","volume":"378","author":"V Kasivisvanathan","year":"2018","journal-title":"The New England journal of medicine"},{"key":"ref5","doi-asserted-by":"publisher","DOI":"10.1007\/s00330-018-5751-1","article-title":"Histopathological to multiparametric MRI spatial mapping of extended systematic sextant and MR\/TRUS-fusion-targeted biopsy of the prostate","author":"D Bonekamp","year":"2018","journal-title":"European radiology"},{"key":"ref6","doi-asserted-by":"publisher","first-page":"29","DOI":"10.1016\/j.euo.2018.01.002","article-title":"Association Between Prostate Imaging Reporting and Data System (PI-RADS) Score for the Index Lesion and Multifocal, Clinically Significant Prostate Cancer","volume":"1","author":"A Stabile","year":"2018","journal-title":"Eur Urol Oncol"},{"key":"ref7","doi-asserted-by":"crossref","first-page":"385","DOI":"10.1016\/j.eururo.2018.05.035","article-title":"Prostate Imaging-Reporting and Data System Steering Committee: PI-RADS v2 Status Update and Future Directions","volume":"75","author":"A R Padhani","year":"2018","journal-title":"Eur Urol"},{"key":"ref8","doi-asserted-by":"publisher","first-page":"16","DOI":"10.1016\/j.eururo.2015.08.052","article-title":"PI-RADS Prostate Imaging \u2013 Reporting and Data System: 2015, Version 2","volume":"69","author":"J C Weinreb","year":"2016","journal-title":"Eur Urol"},{"key":"ref11","doi-asserted-by":"publisher","first-page":"607","DOI":"10.1148\/radiol.2019190938","article-title":"Classification of Cancer at Prostate MRI: Deep Learning versus Clinical PI-RADS Assessment","volume":"293","author":"P Schelb","year":"2019","journal-title":"Radiology"},{"key":"ref12","doi-asserted-by":"publisher","first-page":"19518","DOI":"10.1038\/s41598-019-55972-4","article-title":"Prostate Cancer Detection using Deep Convolutional Neural Networks","volume":"9","author":"S Yoo","year":"2019","journal-title":"Scientific reports"},{"key":"ref13","doi-asserted-by":"publisher","DOI":"10.1002\/jmri.27204","article-title":"Deep-Learning-Based Artificial Intelligence for PI-RADS Classification to Assist Multiparametric Prostate MRI Interpretation: A Development Study","author":"T Sanford","year":"2020","journal-title":"J Magn Reson Imaging"},{"key":"ref14","doi-asserted-by":"publisher","first-page":"677","DOI":"10.1148\/rg.313105139","article-title":"Advancements in MR imaging of the prostate: from diagnosis to interventions","volume":"31","author":"D Bonekamp","year":"2011","journal-title":"Radiographics"},{"key":"ref15","doi-asserted-by":"publisher","first-page":"1301","DOI":"10.1038\/s41591-019-0508-1","article-title":"Clinical-grade computational pathology using weakly supervised deep learning on whole slide images","volume":"25","author":"G Campanella","year":"2019","journal-title":"Nat Med"},{"key":"ref16","doi-asserted-by":"publisher","first-page":"1127","DOI":"10.1109\/TMI.2017.2789181","article-title":"Automated Detection of Clinically Significant Prostate Cancer in mp-MRI Images Based on an End-to-End Deep Neural Network","volume":"37","author":"Z Wang","year":"2018","journal-title":"IEEE Trans Med Imaging"},{"key":"ref17","doi-asserted-by":"crossref","first-page":"297","DOI":"10.2307\/1932409","article-title":"Measures of the amount of ecologic association between species","volume":"26","author":"L R Dice","year":"1945","journal-title":"Ecology"},{"key":"ref18","doi-asserted-by":"crossref","first-page":"579","DOI":"10.1002\/jmri.25372","article-title":"Accuracy and agreement of PIRADSv2 for prostate cancer mpMRI: A multireader study","volume":"45","author":"M D Greer","year":"2017","journal-title":"J Magn Reson Imaging"},{"key":"ref19","doi-asserted-by":"publisher","first-page":"746","DOI":"10.1007\/s00330-011-2377-y","article-title":"ESUR prostate MR guidelines 2012","volume":"22","author":"J O Barentsz","year":"2012","journal-title":"European radiology"},{"key":"ref20","doi-asserted-by":"publisher","first-page":"253","DOI":"10.1055\/s-0032-1330270","article-title":"[PI-RADS classification: structured reporting for MRI of the prostate]","volume":"185","author":"M Rothke","year":"2013","journal-title":"Rofo"},{"key":"ref21","doi-asserted-by":"publisher","first-page":"846","DOI":"10.1016\/j.eururo.2015.12.052","article-title":"Multiparametric Magnetic Resonance Imaging (MRI) and MRI-Transrectal Ultrasound Fusion Biopsy for Index Tumor Detection: Correlation with Radical Prostatectomy Specimen","volume":"70","author":"J P Radtke","year":"2016","journal-title":"Eur Urol"},{"key":"ref22","doi-asserted-by":"publisher","first-page":"568","DOI":"10.1111\/bju.12132","article-title":"Definitions of terms, processes and a minimum dataset for transperineal prostate biopsies: a standardization approach of the Ginsburg Study Group for Enhanced Prostate Diagnostics","volume":"112","author":"T H Kuru","year":"2013","journal-title":"BJU Int"},{"key":"ref23","doi-asserted-by":"publisher","first-page":"441","DOI":"10.3414\/ME11-02-0031","article-title":"MITK diffusion imaging","volume":"51","author":"K H Fritzsche","year":"2012","journal-title":"Methods Inf Med"},{"key":"ref24","doi-asserted-by":"publisher","first-page":"607","DOI":"10.1007\/s11548-013-0840-8","article-title":"The Medical Imaging Interaction Toolkit: challenges and advances: 10 years of open-source development","volume":"8","author":"M Nolden","year":"2013","journal-title":"Int J Comput Assist Radiol Surg"},{"key":"ref25","doi-asserted-by":"crossref","first-page":"568","DOI":"10.1111\/bju.12132","article-title":"Definitions of terms, processes and a minimum dataset for transperineal prostate biopsies: a standardization approach of the G insburg S tudy G roup for E nhanced P rostate D iagnostics","volume":"112","author":"T H Kuru","year":"2013","journal-title":"BJU international"},{"key":"ref26","doi-asserted-by":"publisher","first-page":"433","DOI":"10.1111\/apm.12533","article-title":"International Society of Urological Pathology (ISUP) grading of prostate cancer \u2013 An ISUP consensus on contemporary grading","volume":"124","author":"L Egevad","year":"2016","journal-title":"APMIS"},{"key":"ref28","doi-asserted-by":"crossref","first-page":"24","DOI":"10.1148\/radiol.2261021292","article-title":"Towards complete and accurate reporting of studies of diagnostic accuracy: the STARD initiative","volume":"226","author":"P M Bossuyt","year":"2003","journal-title":"Radiology"},{"key":"ref29","doi-asserted-by":"crossref","first-page":"359","DOI":"10.1016\/j.media.2013.12.002","article-title":"Evaluation of prostate segmentation algorithms for MRI: the PROMISE12 challenge","volume":"18","author":"G Litjens","year":"2014","journal-title":"Medical image analysis"},{"key":"ref30","doi-asserted-by":"publisher","first-page":"128","DOI":"10.1148\/radiol.2018173064","article-title":"Radiomic Machine Learning for Characterization of Prostate Lesions with MRI: Comparison to ADC Values","volume":"289","author":"D Bonekamp","year":"2018","journal-title":"Radiology"},{"key":"ref31","doi-asserted-by":"publisher","first-page":"340","DOI":"10.1016\/j.eururo.2019.02.033","article-title":"Prostate Imaging Reporting and Data System Version 2.1: 2019 Update of Prostate Imaging Reporting and Data System Version 2","volume":"76","author":"B Turkbey","year":"2019","journal-title":"Eur Urol"},{"key":"ref32","first-page":"2","article-title":"Explainable artificial intelligence (xai)","volume":"2","author":"D Gunning","year":"2017","journal-title":"Defense Advanced Research Projects Agency (DARPA), nd Web"}],"container-title":["R\u00f6Fo - Fortschritte auf dem Gebiet der R\u00f6ntgenstrahlen und der bildgebenden Verfahren"],"original-title":[],"language":"de","link":[{"URL":"http:\/\/www.thieme-connect.de\/products\/ejournals\/pdf\/10.1055\/a-1290-8070.pdf","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2021,4,27]],"date-time":"2021-04-27T22:41:55Z","timestamp":1619563315000},"score":1,"resource":{"primary":{"URL":"http:\/\/www.thieme-connect.de\/DOI\/DOI?10.1055\/a-1290-8070"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2020,11,19]]},"references-count":29,"journal-issue":{"issue":"05","published-online":{"date-parts":[[2021,4,27]]},"published-print":{"date-parts":[[2021,5]]}},"URL":"https:\/\/doi.org\/10.1055\/a-1290-8070","relation":{},"ISSN":["1438-9029","1438-9010"],"issn-type":[{"value":"1438-9029","type":"print"},{"value":"1438-9010","type":"electronic"}],"subject":[],"published":{"date-parts":[[2020,11,19]]}}}