{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,5,29]],"date-time":"2026-05-29T22:59:50Z","timestamp":1780095590834,"version":"3.54.0"},"reference-count":16,"publisher":"Springer Science and Business Media LLC","issue":"3","license":[{"start":{"date-parts":[[2023,2,3]],"date-time":"2023-02-03T00:00:00Z","timestamp":1675382400000},"content-version":"tdm","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0"},{"start":{"date-parts":[[2023,2,3]],"date-time":"2023-02-03T00:00:00Z","timestamp":1675382400000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0"}],"funder":[{"name":"Universit\u00e4tsklinikum Freiburg"}],"content-domain":{"domain":["link.springer.com"],"crossmark-restriction":false},"short-container-title":["HNO"],"published-print":{"date-parts":[[2023,3]]},"abstract":"<jats:title>Zusammenfassung<\/jats:title><jats:sec>\n                <jats:title>Hintergrund<\/jats:title>\n                <jats:p>Maligne Speicheldr\u00fcsentumoren sind aufgrund ihrer gro\u00dfen Anzahl an histopathologischen Entit\u00e4ten, ihres seltenen Auftretens und der Vielfalt der klinischen und histologischen Pr\u00e4sentation eine besondere Herausforderung in der Diagnostik. Ziel der vorliegenden Arbeit ist es, die Anwendung von konvolutionalen neuronalen Netzen (CNN) als Hilfsmittel bei der histologischen Diagnose von malignen Speicheldr\u00fcsentumoren zu untersuchen und zu vergleichen.<\/jats:p>\n              <\/jats:sec><jats:sec>\n                <jats:title>Methoden<\/jats:title>\n                <jats:p>Dazu wurden 118 histologische Schnitte von Speicheldr\u00fcsenkarzinompr\u00e4paraten von 68\u00a0Patienten hochaufl\u00f6send digitalisiert. Diese virtuellen Schnitte wurden in kleine Bildausschnitte unterteilt. Die 83.819 Bilder wurden in 4\u00a0Kategorien eingeteilt: Hintergrund, Binde- und St\u00fctzgewebe, nichtneoplastisches Speicheldr\u00fcsengewebe und Speicheldr\u00fcsenkarzinomgewebe, wobei unter der letzten Kategorie die Entit\u00e4ten adenoidzystisches Karzinom, Adenokarzinom (\u201enot otherwise specified\u201c), Azinuszellkarzinom, Basalzellkarzinom, Mukoepidermoidkarzinom und das myoepitheliale Karzinom zusammengefasst wurden. Die kategorisierten Bilder wurden dann in einem Trainings\u2011, Validierungs- und Testlauf von mit dem Bilddatensatz ImageNet vortrainierten CNN (Inception ResNet v2, Inception v3, ResNet152, Xception) in verschiedenen Pixelgr\u00f6\u00dfen verarbeitet.<\/jats:p>\n              <\/jats:sec><jats:sec>\n                <jats:title>Ergebnisse<\/jats:title>\n                <jats:p>Die Accuracy-Werte reichten von 18,8\u201384,7\u202f% \u00fcber alle Netzarchitekturen und Pixelgr\u00f6\u00dfen, dabei erreichte das Inception-v3-Netz den h\u00f6chsten Wert bei 500\u202f\u00d7\u2009500 Pixel. Die erreichten Recall-Werte bzw. die Sensitivit\u00e4t f\u00fcr verschiedene Pixelgr\u00f6\u00dfen lagen bei bis zu 85\u202f% (Inception-v3-Netz bei 1000\u202f\u00d7\u20091000 Pixel). Der minimal erreichte F1-Score misst 0,07 f\u00fcr das Inception ResNet v2 sowie das Inception v3 bei jeweils 100\u202f\u00d7\u2009100 Pixeln, der maximal erreichte F1-Score lag bei 0,72 f\u00fcr das Xception bei 1000\u202f\u00d7\u20091000 Pixeln. Das Netz mit den k\u00fcrzesten Trainingszeiten war das Inception v3, das allen anderen getesteten Netzen bei jeder Pixelgr\u00f6\u00dfe \u00fcberlegen war.<\/jats:p>\n              <\/jats:sec><jats:sec>\n                <jats:title>Schlussfolgerung<\/jats:title>\n                <jats:p>In der vorliegenden Arbeit konnte erstmals f\u00fcr den Bereich der histopathologischen Analyse von Speicheldr\u00fcsenhistologien die Anwendbarkeit von CNN dargestellt und ein Vergleich der Performance verschiedener Netzarchitekturen angestellt werden. Die Ergebnisse lassen einen deutlichen potenziellen Nutzen zuk\u00fcnftiger Anwendungen erkennen.<\/jats:p>\n              <\/jats:sec>","DOI":"10.1007\/s00106-023-01276-z","type":"journal-article","created":{"date-parts":[[2023,2,3]],"date-time":"2023-02-03T15:08:37Z","timestamp":1675436917000},"page":"170-176","update-policy":"https:\/\/doi.org\/10.1007\/springer_crossmark_policy","source":"Crossref","is-referenced-by-count":7,"title":["Ein Vergleich von 4\u00a0konvolutionalen neuronalen Netzen in der histopathologischen Diagnostik von Speicheldr\u00fcsenkarzinomen","Comparison of four convolutional neural networks for histopathological diagnosis of salivary gland carcinomas"],"prefix":"10.1007","volume":"71","author":[{"given":"Tobias","family":"Schulz","sequence":"first","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Christoph","family":"Becker","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Gian","family":"Kayser","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"297","published-online":{"date-parts":[[2023,2,3]]},"reference":[{"issue":"1","key":"1276_CR1","doi-asserted-by":"publisher","first-page":"35","DOI":"10.1038\/nmeth.3707","volume":"13","author":"N Rusk","year":"2015","unstructured":"Rusk\u00a0N (2015) Deep learning. Nat Methods 13(1):35","journal-title":"Nat Methods"},{"key":"1276_CR2","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-319-42999-1","volume-title":"Deep learning and convolutional neural networks for medical image computing","author":"L Lu","year":"2017","unstructured":"Lu\u00a0L, Zheng\u00a0Y, Carneiro\u00a0G (2017) Deep learning and convolutional neural networks for medical image computing. Springer, Cham"},{"issue":"3","key":"1276_CR3","doi-asserted-by":"publisher","first-page":"257","DOI":"10.1007\/s12194-017-0406-5","volume":"10","author":"K Suzuki","year":"2017","unstructured":"Suzuki\u00a0K (2017) Overview of deep learning in medical imaging. Radiol Phys Technol 10(3):257\u2013273","journal-title":"Radiol Phys Technol"},{"issue":"6","key":"1276_CR4","doi-asserted-by":"publisher","first-page":"84","DOI":"10.1145\/3065386","volume":"60","author":"BA Krizhevsky","year":"2012","unstructured":"Krizhevsky\u00a0BA, Sutskever\u00a0I, Hinton\u00a0GE (2012) Cnn\u5b9e\u9645\u8bad\u7ec3\u7684. Commun ACM 60(6):84\u201390","journal-title":"Commun ACM"},{"key":"1276_CR5","doi-asserted-by":"publisher","first-page":"317","DOI":"10.1016\/j.ebiom.2017.12.026","volume":"27","author":"P Khosravi","year":"2018","unstructured":"Khosravi\u00a0P, Kazemi\u00a0E, Imielinski\u00a0M, Elemento\u00a0O, Hajirasouliha\u00a0I (2018) Deep convolutional neural networks enable discrimination of heterogeneous digital pathology images. eBioMedicine 27:317\u2013328. https:\/\/doi.org\/10.1016\/j.ebiom.2017.12.026","journal-title":"eBioMedicine"},{"key":"1276_CR6","doi-asserted-by":"publisher","first-page":"1","DOI":"10.1155\/2018\/2362108","volume":"2018","author":"A-A Nahid","year":"2018","unstructured":"Nahid\u00a0A\u2011A, Mehrabi\u00a0MA, Kong\u00a0Y (2018) Histopathological breast cancer image classification by deep neural network techniques guided by local clustering. Biomed Res Int 2018:1\u201320","journal-title":"Biomed Res Int"},{"key":"1276_CR7","doi-asserted-by":"publisher","first-page":"737","DOI":"10.1007\/978-3-319-93000-8_83","volume":"10882","author":"A Rakhlin","year":"2018","unstructured":"Rakhlin\u00a0A, Shvets\u00a0A, Iglovikov\u00a0V, Kalinin\u00a0AA (2018) Deep convolutional neural networks for breast cancer histology image analysis. Lect Notes Comput Sci 10882:737\u2013744","journal-title":"Lect Notes Comput Sci"},{"issue":"1","key":"1276_CR8","doi-asserted-by":"publisher","first-page":"131","DOI":"10.1007\/s10278-018-0112-9","volume":"32","author":"H Yoon","year":"2019","unstructured":"Yoon\u00a0H, Lee\u00a0J, Oh\u00a0JE, Kim\u00a0HR, Lee\u00a0S, Chang\u00a0HJ et\u00a0al (2019) Tumor identification in colorectal histology images using a\u00a0convolutional neural network. J\u00a0Digit Imaging 32(1):131\u2013140","journal-title":"J Digit Imaging"},{"issue":"40","key":"1276_CR9","doi-asserted-by":"publisher","first-page":"6207","DOI":"10.3748\/wjg.v26.i40.6207","volume":"26","author":"HJ Jang","year":"2020","unstructured":"Jang\u00a0HJ, Lee\u00a0A, Kang\u00a0J, Song\u00a0IH, Lee\u00a0SH (2020) Prediction of clinically actionable genetic alterations from colorectal cancer histopathology images using deep learning. World J Gastroenterol 26(40):6207\u20136223","journal-title":"World J Gastroenterol"},{"issue":"2","key":"1276_CR10","doi-asserted-by":"publisher","first-page":"1139","DOI":"10.1007\/s10103-021-03365-3","volume":"37","author":"Z Yang","year":"2022","unstructured":"Yang\u00a0Z, Shang\u00a0J, Liu\u00a0C, Zhang\u00a0J, Liang\u00a0Y (2022) Classification of oral salivary gland tumors based on texture features in optical coherence tomography images. Lasers Med Sci 37(2):1139\u20131146. https:\/\/doi.org\/10.1007\/s10103-021-03365-3","journal-title":"Lasers Med Sci"},{"issue":"1","key":"1276_CR11","doi-asserted-by":"publisher","first-page":"1","DOI":"10.1038\/s41598-019-50313-x","volume":"9","author":"M Halicek","year":"2019","unstructured":"Halicek\u00a0M, Shahedi\u00a0M, Little\u00a0JV, Chen\u00a0AY, Myers\u00a0LL, Sumer\u00a0BD et al (2019) Head and neck cancer detection in digitized whole-slide histology using convolutional neural networks. Sci Rep 9(1):1\u201311. https:\/\/doi.org\/10.1038\/s41598-019-50313-x","journal-title":"Sci Rep"},{"key":"1276_CR12","doi-asserted-by":"publisher","first-page":"102921","DOI":"10.1016\/j.bspc.2021.102921","volume":"69","author":"DFD dos Santos","year":"2021","unstructured":"dos Santos\u00a0DFD, de Faria\u00a0PR, Traven\u00e7olo\u00a0BAN, do Nascimento\u00a0MZ (2021) Automated detection of tumor regions from oral histological whole slide images using fully convolutional neural networks. Biomed Signal Process Control 69:102921. https:\/\/doi.org\/10.1016\/j.bspc.2021.102921","journal-title":"Biomed Signal Process Control"},{"issue":"1","key":"1276_CR13","doi-asserted-by":"publisher","first-page":"55","DOI":"10.1007\/s12105-017-0795-0","volume":"11","author":"RR Seethala","year":"2017","unstructured":"Seethala\u00a0RR, Stenman\u00a0G (2017) Update from the 4th Edition of the World Health Organization classification of head and neck tumours: tumors of the salivary gland. Head Neck Pathol 11(1):55\u201367","journal-title":"Head Neck Pathol"},{"issue":"1","key":"1276_CR14","doi-asserted-by":"publisher","first-page":"1","DOI":"10.1038\/s41598-017-17204-5","volume":"7","author":"P Bankhead","year":"2017","unstructured":"Bankhead\u00a0P, Loughrey\u00a0MB, Fern\u00e1ndez\u00a0JA, Dombrowski\u00a0Y, McArt\u00a0DG, Dunne\u00a0PD et\u00a0al (2017) QuPath: Open source software for digital pathology image analysis. Sci Rep 7(1):1\u20137","journal-title":"Sci Rep"},{"issue":"7639","key":"1276_CR15","doi-asserted-by":"publisher","first-page":"115","DOI":"10.1038\/nature21056","volume":"542","author":"A Esteva","year":"2017","unstructured":"Esteva\u00a0A, Kuprel\u00a0B, Novoa\u00a0RA, Ko\u00a0J, Swetter\u00a0SM, Blau\u00a0HM et al (2017) Dermatologist-level classification of skin cancer with deep neural networks. Nature 542(7639):115\u2013118","journal-title":"Nature"},{"issue":"4","key":"1276_CR16","doi-asserted-by":"publisher","first-page":"829","DOI":"10.1007\/s00432-018-02834-7","volume":"145","author":"PR Jeyaraj","year":"2019","unstructured":"Jeyaraj\u00a0PR, Samuel Nadar\u00a0ER (2019) Computer-assisted medical image classification for early diagnosis of oral cancer employing deep learning algorithm. J\u00a0Cancer Res Clin Oncol 145(4):829\u2013837. https:\/\/doi.org\/10.1007\/s00432-018-02834-7","journal-title":"J Cancer Res Clin Oncol"}],"container-title":["HNO"],"original-title":[],"language":"de","link":[{"URL":"https:\/\/link.springer.com\/content\/pdf\/10.1007\/s00106-023-01276-z.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/link.springer.com\/article\/10.1007\/s00106-023-01276-z\/fulltext.html","content-type":"text\/html","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/link.springer.com\/content\/pdf\/10.1007\/s00106-023-01276-z.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2023,2,23]],"date-time":"2023-02-23T09:04:26Z","timestamp":1677143066000},"score":1,"resource":{"primary":{"URL":"https:\/\/link.springer.com\/10.1007\/s00106-023-01276-z"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2023,2,3]]},"references-count":16,"journal-issue":{"issue":"3","published-print":{"date-parts":[[2023,3]]}},"alternative-id":["1276"],"URL":"https:\/\/doi.org\/10.1007\/s00106-023-01276-z","relation":{},"ISSN":["0017-6192","1433-0458"],"issn-type":[{"value":"0017-6192","type":"print"},{"value":"1433-0458","type":"electronic"}],"subject":[],"published":{"date-parts":[[2023,2,3]]},"assertion":[{"value":"20 January 2023","order":1,"name":"accepted","label":"Accepted","group":{"name":"ArticleHistory","label":"Article History"}},{"value":"3 February 2023","order":2,"name":"first_online","label":"First Online","group":{"name":"ArticleHistory","label":"Article History"}},{"order":1,"name":"Ethics","group":{"name":"EthicsHeading","label":"Einhaltung ethischer Richtlinien"}},{"value":"T.\u00a0Schulz, C.\u00a0Becker und G.\u00a0Kayser geben an, dass kein Interessenkonflikt besteht.","order":2,"name":"Ethics","group":{"name":"EthicsHeading","label":"Interessenkonflikt"}},{"value":"Alle beschriebenen Untersuchungen am Menschen oder an menschlichem Gewebe wurden mit Zustimmung der zust\u00e4ndigen Ethikkommission, im Einklang mit nationalem Recht sowie gem\u00e4\u00df der Deklaration von Helsinki von 1975 (in der aktuellen, \u00fcberarbeiteten Fassung) durchgef\u00fchrt. F\u00fcr die aufgef\u00fchrten Studien gelten die jeweils dort angegebenen ethischen Richtlinien. Die Studie wurde von der Ethikkommission gepr\u00fcft und genehmigt (Ethikkommissions-Vorlage Nummer 167\/18).","order":3,"name":"Ethics","group":{"name":"EthicsHeading","label":"Interessenkonflikt"}}]}}