<?xml version="1.0" encoding="utf-8"?>
<export-example>
  <doc>
    <id>4045</id>
    <completedYear/>
    <publishedYear>2023</publishedYear>
    <thesisYearAccepted/>
    <language>eng</language>
    <pageFirst/>
    <pageLast/>
    <pageNumber>16</pageNumber>
    <edition/>
    <issue>14</issue>
    <volume>2023</volume>
    <articleNumber>100301</articleNumber>
    <type>article</type>
    <publisherName>Elsevier</publisherName>
    <publisherPlace>Amsterdam</publisherPlace>
    <creatingCorporation/>
    <contributingCorporation/>
    <belongsToBibliography>1</belongsToBibliography>
    <completedDate>--</completedDate>
    <publishedDate>--</publishedDate>
    <thesisDateAccepted>--</thesisDateAccepted>
    <title language="eng">Pan-tumor T-lymphocyte detection using deep neural networks: Recommendations for transfer learning in immunohistochemistry</title>
    <abstract language="eng">The success of immuno-oncology treatments promises long-term cancer remission for an increasing number of patients. The response to checkpoint inhibitor drugs has shown a correlation with the presence of immune cells in the tumor and tumor microenvironment. An in-depth understanding of the spatial localization of immune cells is therefore critical for understanding the tumor’s immune landscape and predicting drug response. Computer-aided systems are well suited for efficiently quantifying immune cells in their spatial context. Conventional image analysis approaches are often based on color features and therefore require a high level of manual interaction. More robust image analysis methods based on deep learning are expected to decrease this reliance on human interaction and improve the reproducibility of immune cell scoring. However, these methods require sufficient training data and previous work has reported low robustness of these algorithms when they are tested on out-of-distribution data from different pathology labs or samples from different organs. In this work, we used a new image analysis pipeline to explicitly evaluate the robustness of marker-labeled lymphocyte quantification algorithms depending on the number of training samples before and after being transferred to a new tumor indication. For these experiments, we adapted the RetinaNet architecture for the task of T-lymphocyte detection and employed transfer learning to bridge the domain gap between tumor indications and reduce the annotation costs for unseen domains. On our test set, we achieved human-level performance for almost all tumor indications with an average precision of 0.74 in-domain and 0.72–0.74 cross-domain. From our results, we derive recommendations for model development regarding annotation extent, training sample selection, and label extraction for the development of robust algorithms for immune cell scoring. By extending the task of marker-labeled lymphocyte quantification to a multi-class detection task, the pre-requisite for subsequent analyses, e.g., distinguishing lymphocytes in the tumor stroma from tumor-infiltrating lymphocytes, is met.</abstract>
    <parentTitle language="eng">Journal of Pathology Informatics</parentTitle>
    <identifier type="issn">2153-3539</identifier>
    <identifier type="urn">urn:nbn:de:bvb:573-40458</identifier>
    <enrichment key="opus_doi_flag">true</enrichment>
    <enrichment key="opus_import_data">{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2023,9,13]],"date-time":"2023-09-13T21:24:30Z","timestamp":1694640270410},"reference-count":32,"publisher":"Elsevier BV","license":[{"start":{"date-parts":[[2023,1,1]],"date-time":"2023-01-01T00:00:00Z","timestamp":1672531200000},"content-version":"tdm","delay-in-days":0,"URL":"https:\/\/www.elsevier.com\/tdm\/userlicense\/1.0\/"},{"start":{"date-parts":[[2023,2,11]],"date-time":"2023-02-11T00:00:00Z","timestamp":1676073600000},"content-version":"vor","delay-in-days":41,"URL":"http:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"content-domain":{"domain":["elsevier.com","sciencedirect.com"],"crossmark-restriction":true},"short-container-title":["Journal of Pathology Informatics"],"published-print":{"date-parts":[[2023]]},"DOI":"10.1016\/j.jpi.2023.100301","type":"journal-article","created":{"date-parts":[[2023,2,27]],"date-time":"2023-02-27T07:08:06Z","timestamp":1677481686000},"page":"100301","update-policy":"http:\/\/dx.doi.org\/10.1016\/elsevier_cm_policy","source":"Crossref","is-referenced-by-count":0,"title":["Pan-tumor T-lymphocyte detection using deep neural networks: Recommendations for transfer learning in immunohistochemistry"],"prefix":"10.1016","volume":"14","author":[{"given":"Frauke","family":"Wilm","sequence":"first","affiliation":[]},{"given":"Christian","family":"Ihling","sequence":"additional","affiliation":[]},{"given":"G\u00e1bor","family":"M\u00e9hes","sequence":"additional","affiliation":[]},{"given":"Luigi","family":"Terracciano","sequence":"additional","affiliation":[]},{"given":"Chlo\u00e9","family":"Puget","sequence":"additional","affiliation":[]},{"given":"Robert","family":"Klopfleisch","sequence":"additional","affiliation":[]},{"given":"Peter","family":"Sch\u00fcffler","sequence":"additional","affiliation":[]},{"given":"Marc","family":"Aubreville","sequence":"additional","affiliation":[]},{"given":"Andreas","family":"Maier","sequence":"additional","affiliation":[]},{"given":"Thomas","family":"Mrowiec","sequence":"additional","affiliation":[]},{"given":"Katharina","family":"Breininger","sequence":"additional","affiliation":[]}],"member":"78","reference":[{"key":"10.1016\/j.jpi.2023.100301_bb0005","series-title":"Semin Cancer Biol","first-page":"151","article-title":"Scoring of tumor-infiltrating lymphocytes: from visual estimation to machine learning","volume":"52","author":"Klauschen","year":"2018"},{"issue":"4","key":"10.1016\/j.jpi.2023.100301_bb0010","doi-asserted-by":"crossref","first-page":"405","DOI":"10.1354\/vp.42-4-405","article-title":"Technical aspects of immunohistochemistry","volume":"42","author":"Ramos-Vara","year":"2005","journal-title":"Vet Pathol"},{"key":"10.1016\/j.jpi.2023.100301_bb0015","doi-asserted-by":"crossref","first-page":"116471","DOI":"10.1016\/j.eswa.2021.116471","article-title":"Deep learning-based instance segmentation for the precise automated quantification of digital breast cancer immunohistochemistry images","volume":"193","author":"Priego-Torres","year":"2022","journal-title":"Expert Syst Appl"},{"key":"10.1016\/j.jpi.2023.100301_bb0020","series-title":"Proc IEEE Int Symp Comput Based Med Syst","first-page":"200","article-title":"Automatic lymphocyte detection on gastric cancer IHC images using deep learning","author":"Garcia","year":"2017"},{"key":"10.1016\/j.jpi.2023.100301_bb0025","series-title":"Mach Learn Med Imaging","first-page":"17","article-title":"Deep learning based automatic immune cell detection for immunohistochemistry images","author":"Chen","year":"2014"},{"issue":"1","key":"10.1016\/j.jpi.2023.100301_bb0030","first-page":"1","article-title":"Deep learning-based image analysis methods for brightfield-acquired multiplex immunohistochemistry images","volume":"15","author":"Fassler","year":"2020","journal-title":"Diagn Pathol"},{"issue":"5","key":"10.1016\/j.jpi.2023.100301_bb0035","doi-asserted-by":"crossref","first-page":"735","DOI":"10.1007\/s40846-020-00545-4","article-title":"Automatic detection and counting of lymphocytes from immunohistochemistry cancer images using deep learning","volume":"40","author":"Evangeline","year":"2020","journal-title":"J Med Biol Eng"},{"key":"10.1016\/j.jpi.2023.100301_bb0040","series-title":"Medical Imaging with Deep Learning","article-title":"You only look on lymphocytes once","author":"van Rijthoven","year":"2018"},{"key":"10.1016\/j.jpi.2023.100301_bb0045","doi-asserted-by":"crossref","DOI":"10.1016\/j.media.2019.101547","article-title":"Learning to detect lymphocytes in immunohistochemistry with deep learning","volume":"58","author":"Swiderska-Chadaj","year":"2019","journal-title":"Med Image Anal"},{"issue":"2","key":"10.1016\/j.jpi.2023.100301_bb0050","doi-asserted-by":"crossref","first-page":"325","DOI":"10.1109\/JBHI.2020.3032060","article-title":"Measuring domain shift for deep learning in histopathology","volume":"25","author":"Stacke","year":"2020","journal-title":"IEEE J Biomed Health Inform"},{"key":"10.1016\/j.jpi.2023.100301_bb0055","series-title":"Biomedical Image Registration, Domain Generalisation and Out-of-Distribution Analysis: MICCAI 2021 Challenges","first-page":"5","article-title":"Domain adversarial retinanet as a reference algorithm for the mitosis domain generalization challenge","author":"Wilm","year":"2022"},{"issue":"417","key":"10.1016\/j.jpi.2023.100301_bb0060","first-page":"1","article-title":"A completely annotated whole slide image dataset of canine breast cancer to aid human breast cancer research","volume":"7","author":"Aubreville","year":"2020","journal-title":"Sci Data"},{"key":"10.1016\/j.jpi.2023.100301_bb0065","series-title":"TNM Classification of Malignant Tumours","author":"Sobin","year":"2011"},{"issue":"1","key":"10.1016\/j.jpi.2023.100301_bb0070","doi-asserted-by":"crossref","first-page":"1","DOI":"10.1038\/s41598-021-83827-4","article-title":"EXACT: a collaboration toolset for algorithm-aided annotation of images with annotation version control","volume":"11","author":"Marzahl","year":"2021","journal-title":"Sci Rep"},{"key":"10.1016\/j.jpi.2023.100301_bb0075","series-title":"Bildverarb Med","first-page":"241","article-title":"Influence of inter-annotator variability on automatic mitotic figure assessment","author":"Wilm","year":"2021"},{"issue":"1","key":"10.1016\/j.jpi.2023.100301_bb0080","doi-asserted-by":"crossref","first-page":"1","DOI":"10.1038\/s41598-020-65958-2","article-title":"Deep learning-based quantification of pulmonary hemosiderophages in cytology slides","volume":"10","author":"Marzahl","year":"2020","journal-title":"Sci Rep"},{"issue":"1","key":"10.1016\/j.jpi.2023.100301_bb0085","doi-asserted-by":"crossref","first-page":"37","DOI":"10.1177\/001316446002000104","article-title":"A coefficient of agreement for nominal scales","volume":"20","author":"Cohen","year":"1960","journal-title":"Educ Psychol Meas"},{"key":"10.1016\/j.jpi.2023.100301_bb0090","doi-asserted-by":"crossref","first-page":"159","DOI":"10.2307\/2529310","article-title":"The measurement of observer agreement for categorical data","author":"Landis","year":"1977","journal-title":"Biometrics"},{"key":"10.1016\/j.jpi.2023.100301_bb0095","series-title":"Inf Process Med Imaging","first-page":"842","article-title":"Signet ring cell detection with a semi-supervised learning framework","author":"Li","year":"2019"},{"issue":"3","key":"10.1016\/j.jpi.2023.100301_bb0100","doi-asserted-by":"crossref","first-page":"558","DOI":"10.1177\/0013164418823249","article-title":"Kappa coefficients for missing data","volume":"79","author":"De Raadt","year":"2019","journal-title":"Educ Psychol Meas"},{"key":"10.1016\/j.jpi.2023.100301_bb0105","series-title":"Proc IEEE Int Conf Comput Vis","first-page":"2980","article-title":"Focal loss for dense object detection","author":"Lin","year":"2017"},{"key":"10.1016\/j.jpi.2023.100301_bb0110","series-title":"Proc IEEE Comput Soc Conf Comput Vis Pattern Recognit","first-page":"770","article-title":"Deep residual learning for image recognition","author":"He","year":"2016"},{"issue":"3","key":"10.1016\/j.jpi.2023.100301_bb0115","doi-asserted-by":"crossref","first-page":"211","DOI":"10.1007\/s11263-015-0816-y","article-title":"ImageNet large scale visual recognition challenge","volume":"115","author":"Russakovsky","year":"2015","journal-title":"Int J Comput Vis"},{"issue":"4","key":"10.1016\/j.jpi.2023.100301_bb0120","doi-asserted-by":"crossref","first-page":"466","DOI":"10.1177\/2472555218818756","article-title":"Transfer learning with deep convolutional neural networks for classifying cellular morphological changes","volume":"24","author":"Kensert","year":"2019","journal-title":"SLAS Discov"},{"key":"10.1016\/j.jpi.2023.100301_bb0125","article-title":"Self supervised contrastive learning for digital histopathology","volume":"7","author":"Ciga","year":"2022","journal-title":"Mach Learn Appl"},{"issue":"5","key":"10.1016\/j.jpi.2023.100301_bb0130","doi-asserted-by":"crossref","first-page":"1299","DOI":"10.1109\/TMI.2016.2535302","article-title":"Convolutional neural networks for medical image analysis: full training or fine tuning?","volume":"35","author":"Tajbakhsh","year":"2016","journal-title":"IEEE Trans Med Imaging"},{"key":"10.1016\/j.jpi.2023.100301_bb0135","author":"Howard"},{"issue":"2","key":"10.1016\/j.jpi.2023.100301_bb0140","doi-asserted-by":"crossref","first-page":"108","DOI":"10.3390\/info11020108","article-title":"Fastai: A layered API for deep learning","volume":"11","author":"Howard","year":"2020","journal-title":"Information"},{"issue":"1\u20132","key":"10.1016\/j.jpi.2023.100301_bb0145","doi-asserted-by":"crossref","first-page":"83","DOI":"10.1002\/nav.3800020109","article-title":"The Hungarian method for the assignment problem","volume":"2","author":"Kuhn","year":"1955","journal-title":"Naval Res Logist Quart"},{"issue":"5","key":"10.1016\/j.jpi.2023.100301_bb0150","doi-asserted-by":"crossref","first-page":"646","DOI":"10.1016\/j.cell.2011.02.013","article-title":"Hallmarks of cancer: the next generation","volume":"144","author":"Hanahan","year":"2011","journal-title":"Cell"},{"key":"10.1016\/j.jpi.2023.100301_bb0155","doi-asserted-by":"crossref","first-page":"162","DOI":"10.3389\/fmed.2019.00162","article-title":"Learning domain-invariant representations of histological images","volume":"6","author":"Lafarge","year":"2019","journal-title":"Front Med"},{"key":"10.1016\/j.jpi.2023.100301_bb0160","series-title":"Medical Imaging with Deep Learning","article-title":"Quantifying the scanner-induced domain gap in mitosis detection","author":"Aubreville","year":"2021"}],"container-title":["Journal of Pathology Informatics"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/api.elsevier.com\/content\/article\/PII:S2153353923001153?httpAccept=text\/xml","content-type":"text\/xml","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/api.elsevier.com\/content\/article\/PII:S2153353923001153?httpAccept=text\/plain","content-type":"text\/plain","content-version":"vor","intended-application":"text-mining"}],"deposited":{"date-parts":[[2023,9,10]],"date-time":"2023-09-10T04:07:43Z","timestamp":1694318863000},"score":1,"resource":{"primary":{"URL":"https:\/\/linkinghub.elsevier.com\/retrieve\/pii\/S2153353923001153"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2023]]},"references-count":32,"alternative-id":["S2153353923001153"],"URL":"http:\/\/dx.doi.org\/10.1016\/j.jpi.2023.100301","relation":{},"ISSN":["2153-3539"],"issn-type":[{"value":"2153-3539","type":"print"}],"subject":["Computer Science Applications","Health Informatics","Pathology and Forensic Medicine"],"published":{"date-parts":[[2023]]},"assertion":[{"value":"Elsevier","name":"publisher","label":"This article is maintained by"},{"value":"Pan-tumor T-lymphocyte detection using deep neural networks: Recommendations for transfer learning in immunohistochemistry","name":"articletitle","label":"Article Title"},{"value":"Journal of Pathology Informatics","name":"journaltitle","label":"Journal Title"},{"value":"https:\/\/doi.org\/10.1016\/j.jpi.2023.100301","name":"articlelink","label":"CrossRef DOI link to publisher maintained version"},{"value":"article","name":"content_type","label":"Content Type"},{"value":"\u00a9 2023 The Author(s). Published by Elsevier Inc. on behalf of Association for Pathology Informatics.","name":"copyright","label":"Copyright"}],"article-number":"100301"}}</enrichment>
    <enrichment key="local_crossrefDocumentType">journal-article</enrichment>
    <enrichment key="local_crossrefLicence">https://www.elsevier.com/tdm/userlicense/1.0/</enrichment>
    <enrichment key="local_import_origin">crossref</enrichment>
    <enrichment key="local_doiImportPopulated">SubjectUncontrolled_1,SubjectUncontrolled_2,SubjectUncontrolled_3,PersonAuthorFirstName_1,PersonAuthorLastName_1,PersonAuthorFirstName_2,PersonAuthorLastName_2,PersonAuthorFirstName_3,PersonAuthorLastName_3,PersonAuthorFirstName_4,PersonAuthorLastName_4,PersonAuthorFirstName_5,PersonAuthorLastName_5,PersonAuthorFirstName_6,PersonAuthorLastName_6,PersonAuthorFirstName_7,PersonAuthorLastName_7,PersonAuthorFirstName_8,PersonAuthorLastName_8,PersonAuthorFirstName_9,PersonAuthorLastName_9,PersonAuthorFirstName_10,PersonAuthorLastName_10,PersonAuthorFirstName_11,PersonAuthorLastName_11,PublisherName,TitleMain_1,Language,TitleParent_1,ArticleNumber,Volume,CompletedYear,IdentifierIssn,Enrichmentlocal_crossrefLicence</enrichment>
    <enrichment key="opus.source">doi-import</enrichment>
    <enrichment key="THI_relatedIdentifier">https://doi.org/10.1016/j.jpi.2023.100301</enrichment>
    <enrichment key="THI_openaccess">ja</enrichment>
    <enrichment key="THI_articleversion">published</enrichment>
    <enrichment key="THI_review">peer-review</enrichment>
    <enrichment key="opus.doi.autoCreate">false</enrichment>
    <enrichment key="opus.urn.autoCreate">true</enrichment>
    <licence>Creative Commons BY 4.0</licence>
    <author>
      <first_name>Frauke</first_name>
      <last_name>Wilm</last_name>
    </author>
    <author>
      <first_name>Christian</first_name>
      <last_name>Ihling</last_name>
    </author>
    <author>
      <first_name>Gábor</first_name>
      <last_name>Méhes</last_name>
    </author>
    <author>
      <first_name>Luigi</first_name>
      <last_name>Terracciano</last_name>
    </author>
    <author>
      <first_name>Chloé</first_name>
      <last_name>Puget</last_name>
    </author>
    <author>
      <first_name>Robert</first_name>
      <last_name>Klopfleisch</last_name>
    </author>
    <author>
      <first_name>Peter</first_name>
      <last_name>Schüffler</last_name>
    </author>
    <author>
      <first_name>Marc</first_name>
      <last_name>Aubreville</last_name>
    </author>
    <author>
      <first_name>Andreas</first_name>
      <last_name>Maier</last_name>
    </author>
    <author>
      <first_name>Thomas</first_name>
      <last_name>Mrowiec</last_name>
    </author>
    <author>
      <first_name>Katharina</first_name>
      <last_name>Breininger</last_name>
    </author>
    <subject>
      <language>eng</language>
      <type>uncontrolled</type>
      <value>Tumor-infiltrating lymphocytes</value>
    </subject>
    <subject>
      <language>eng</language>
      <type>uncontrolled</type>
      <value>Immuno-oncology</value>
    </subject>
    <subject>
      <language>eng</language>
      <type>uncontrolled</type>
      <value>Immunohistochemistry</value>
    </subject>
    <subject>
      <language>eng</language>
      <type>uncontrolled</type>
      <value>Deep learning</value>
    </subject>
    <subject>
      <language>eng</language>
      <type>uncontrolled</type>
      <value>Transfer learning</value>
    </subject>
    <subject>
      <language>eng</language>
      <type>uncontrolled</type>
      <value>Domain adaptation</value>
    </subject>
    <collection role="open_access" number="">open_access</collection>
    <collection role="institutes" number="19309">Fakultät Informatik</collection>
    <collection role="institutes" number="19379">AImotion Bavaria</collection>
    <thesisPublisher>Technische Hochschule Ingolstadt</thesisPublisher>
    <file>https://opus4.kobv.de/opus4-haw/files/4045/1-s2.0-S2153353923001153-main.pdf</file>
  </doc>
</export-example>
