Swarm learning with weak supervision enables automatic breast cancer detection in magnetic resonance imaging
Abstract Background Over the next 5 years, new breast cancer screening guidelines recommending magnetic resonance imaging (MRI) for certain patients will significantly increase the volume of imaging data to be analyzed. While this increase poses challenges for radiologists, artificial intelligence (...
Saved in:
| Main Authors: | , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Nature Portfolio
2025-02-01
|
| Series: | Communications Medicine |
| Online Access: | https://doi.org/10.1038/s43856-024-00722-5 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1823861664589545472 |
|---|---|
| author | Oliver Lester Saldanha Jiefu Zhu Gustav Müller-Franzes Zunamys I. Carrero Nicholas R. Payne Lorena Escudero Sánchez Paul Christophe Varoutas Sreenath Kyathanahally Narmin Ghaffari Laleh Kevin Pfeiffer Marta Ligero Jakob Behner Kamarul A. Abdullah Georgios Apostolakos Chrysafoula Kolofousi Antri Kleanthous Michail Kalogeropoulos Cristina Rossi Sylwia Nowakowska Alexandra Athanasiou Raquel Perez-Lopez Ritse Mann Wouter Veldhuis Julia Camps Volkmar Schulz Markus Wenzel Sergey Morozov Alexander Ciritsis Christiane Kuhl Fiona J. Gilbert Daniel Truhn Jakob Nikolas Kather |
| author_facet | Oliver Lester Saldanha Jiefu Zhu Gustav Müller-Franzes Zunamys I. Carrero Nicholas R. Payne Lorena Escudero Sánchez Paul Christophe Varoutas Sreenath Kyathanahally Narmin Ghaffari Laleh Kevin Pfeiffer Marta Ligero Jakob Behner Kamarul A. Abdullah Georgios Apostolakos Chrysafoula Kolofousi Antri Kleanthous Michail Kalogeropoulos Cristina Rossi Sylwia Nowakowska Alexandra Athanasiou Raquel Perez-Lopez Ritse Mann Wouter Veldhuis Julia Camps Volkmar Schulz Markus Wenzel Sergey Morozov Alexander Ciritsis Christiane Kuhl Fiona J. Gilbert Daniel Truhn Jakob Nikolas Kather |
| author_sort | Oliver Lester Saldanha |
| collection | DOAJ |
| description | Abstract Background Over the next 5 years, new breast cancer screening guidelines recommending magnetic resonance imaging (MRI) for certain patients will significantly increase the volume of imaging data to be analyzed. While this increase poses challenges for radiologists, artificial intelligence (AI) offers potential solutions to manage this workload. However, the development of AI models is often hindered by manual annotation requirements and strict data-sharing regulations between institutions. Methods In this study, we present an integrated pipeline combining weakly supervised learning—reducing the need for detailed annotations—with local AI model training via swarm learning (SL), which circumvents centralized data sharing. We utilized three datasets comprising 1372 female bilateral breast MRI exams from institutions in three countries: the United States (US), Switzerland, and the United Kingdom (UK) to train models. These models were then validated on two external datasets consisting of 649 bilateral breast MRI exams from Germany and Greece. Results Upon systematically benchmarking various weakly supervised two-dimensional (2D) and three-dimensional (3D) deep learning (DL) methods, we find that the 3D-ResNet-101 demonstrates superior performance. By implementing a real-world SL setup across three international centers, we observe that these collaboratively trained models outperform those trained locally. Even with a smaller dataset, we demonstrate the practical feasibility of deploying SL internationally with on-site data processing, addressing challenges such as data privacy and annotation variability. Conclusions Combining weakly supervised learning with SL enhances inter-institutional collaboration, improving the utility of distributed datasets for medical AI training without requiring detailed annotations or centralized data sharing. |
| format | Article |
| id | doaj-art-a1bcf5ae3c6b4dc7a9c39931cc3a78b4 |
| institution | Kabale University |
| issn | 2730-664X |
| language | English |
| publishDate | 2025-02-01 |
| publisher | Nature Portfolio |
| record_format | Article |
| series | Communications Medicine |
| spelling | doaj-art-a1bcf5ae3c6b4dc7a9c39931cc3a78b42025-02-09T12:52:10ZengNature PortfolioCommunications Medicine2730-664X2025-02-015111210.1038/s43856-024-00722-5Swarm learning with weak supervision enables automatic breast cancer detection in magnetic resonance imagingOliver Lester Saldanha0Jiefu Zhu1Gustav Müller-Franzes2Zunamys I. Carrero3Nicholas R. Payne4Lorena Escudero Sánchez5Paul Christophe Varoutas6Sreenath Kyathanahally7Narmin Ghaffari Laleh8Kevin Pfeiffer9Marta Ligero10Jakob Behner11Kamarul A. Abdullah12Georgios Apostolakos13Chrysafoula Kolofousi14Antri Kleanthous15Michail Kalogeropoulos16Cristina Rossi17Sylwia Nowakowska18Alexandra Athanasiou19Raquel Perez-Lopez20Ritse Mann21Wouter Veldhuis22Julia Camps23Volkmar Schulz24Markus Wenzel25Sergey Morozov26Alexander Ciritsis27Christiane Kuhl28Fiona J. Gilbert29Daniel Truhn30Jakob Nikolas Kather31Else Kroener Fresenius Center for Digital Health, Medical Faculty Carl Gustav Carus, Technical University DresdenElse Kroener Fresenius Center for Digital Health, Medical Faculty Carl Gustav Carus, Technical University DresdenDepartment of Diagnostic and Interventional Radiology, University Hospital RWTH AachenDepartment of Diagnostic and Interventional Radiology, University Hospital RWTH AachenDepartment of Radiology, Clinical School, Cambridge Biomedical Research Centre, University of CambridgeDepartment of Radiology, Clinical School, Cambridge Biomedical Research Centre, University of CambridgeBreast Imaging Department, Mitera Hospital AthensInstitute of Diagnostic and Interventional Radiology, University Hospital ZurichElse Kroener Fresenius Center for Digital Health, Medical Faculty Carl Gustav Carus, Technical University DresdenElse Kroener Fresenius Center for Digital Health, Medical Faculty Carl Gustav Carus, Technical University DresdenElse Kroener Fresenius Center for Digital Health, Medical Faculty Carl Gustav Carus, Technical University DresdenElse Kroener Fresenius Center for Digital Health, Medical Faculty Carl Gustav Carus, Technical University DresdenDepartment of Radiology, Clinical School, Cambridge Biomedical Research Centre, University of CambridgeBreast Imaging Department, Mitera Hospital AthensBreast Imaging Department, Mitera Hospital AthensBreast Imaging Department, Mitera Hospital AthensBreast Imaging Department, Mitera Hospital AthensInstitute of Diagnostic and Interventional Radiology, University Hospital ZurichInstitute of Diagnostic and Interventional Radiology, University Hospital ZurichBreast Imaging Department, Mitera Hospital AthensRadiomics Group, Vall d’Hebron Institute of Oncology (VHIO)Department of Diagnostic Imaging, Radboud University Medical CenterImaging Division, University Medical Center UtrechtBreast Cancer Unit, Ribera Salud HospitalsFraunhofer Institute for Digital Medicine MEVISFraunhofer Institute for Digital Medicine MEVISThe European Society of Medical Imaging Informatics (EuSoMII)Institute of Diagnostic and Interventional Radiology, University Hospital ZurichDepartment of Diagnostic and Interventional Radiology, University Hospital RWTH AachenDepartment of Radiology, Clinical School, Cambridge Biomedical Research Centre, University of CambridgeDepartment of Diagnostic and Interventional Radiology, University Hospital RWTH AachenElse Kroener Fresenius Center for Digital Health, Medical Faculty Carl Gustav Carus, Technical University DresdenAbstract Background Over the next 5 years, new breast cancer screening guidelines recommending magnetic resonance imaging (MRI) for certain patients will significantly increase the volume of imaging data to be analyzed. While this increase poses challenges for radiologists, artificial intelligence (AI) offers potential solutions to manage this workload. However, the development of AI models is often hindered by manual annotation requirements and strict data-sharing regulations between institutions. Methods In this study, we present an integrated pipeline combining weakly supervised learning—reducing the need for detailed annotations—with local AI model training via swarm learning (SL), which circumvents centralized data sharing. We utilized three datasets comprising 1372 female bilateral breast MRI exams from institutions in three countries: the United States (US), Switzerland, and the United Kingdom (UK) to train models. These models were then validated on two external datasets consisting of 649 bilateral breast MRI exams from Germany and Greece. Results Upon systematically benchmarking various weakly supervised two-dimensional (2D) and three-dimensional (3D) deep learning (DL) methods, we find that the 3D-ResNet-101 demonstrates superior performance. By implementing a real-world SL setup across three international centers, we observe that these collaboratively trained models outperform those trained locally. Even with a smaller dataset, we demonstrate the practical feasibility of deploying SL internationally with on-site data processing, addressing challenges such as data privacy and annotation variability. Conclusions Combining weakly supervised learning with SL enhances inter-institutional collaboration, improving the utility of distributed datasets for medical AI training without requiring detailed annotations or centralized data sharing.https://doi.org/10.1038/s43856-024-00722-5 |
| spellingShingle | Oliver Lester Saldanha Jiefu Zhu Gustav Müller-Franzes Zunamys I. Carrero Nicholas R. Payne Lorena Escudero Sánchez Paul Christophe Varoutas Sreenath Kyathanahally Narmin Ghaffari Laleh Kevin Pfeiffer Marta Ligero Jakob Behner Kamarul A. Abdullah Georgios Apostolakos Chrysafoula Kolofousi Antri Kleanthous Michail Kalogeropoulos Cristina Rossi Sylwia Nowakowska Alexandra Athanasiou Raquel Perez-Lopez Ritse Mann Wouter Veldhuis Julia Camps Volkmar Schulz Markus Wenzel Sergey Morozov Alexander Ciritsis Christiane Kuhl Fiona J. Gilbert Daniel Truhn Jakob Nikolas Kather Swarm learning with weak supervision enables automatic breast cancer detection in magnetic resonance imaging Communications Medicine |
| title | Swarm learning with weak supervision enables automatic breast cancer detection in magnetic resonance imaging |
| title_full | Swarm learning with weak supervision enables automatic breast cancer detection in magnetic resonance imaging |
| title_fullStr | Swarm learning with weak supervision enables automatic breast cancer detection in magnetic resonance imaging |
| title_full_unstemmed | Swarm learning with weak supervision enables automatic breast cancer detection in magnetic resonance imaging |
| title_short | Swarm learning with weak supervision enables automatic breast cancer detection in magnetic resonance imaging |
| title_sort | swarm learning with weak supervision enables automatic breast cancer detection in magnetic resonance imaging |
| url | https://doi.org/10.1038/s43856-024-00722-5 |
| work_keys_str_mv | AT oliverlestersaldanha swarmlearningwithweaksupervisionenablesautomaticbreastcancerdetectioninmagneticresonanceimaging AT jiefuzhu swarmlearningwithweaksupervisionenablesautomaticbreastcancerdetectioninmagneticresonanceimaging AT gustavmullerfranzes swarmlearningwithweaksupervisionenablesautomaticbreastcancerdetectioninmagneticresonanceimaging AT zunamysicarrero swarmlearningwithweaksupervisionenablesautomaticbreastcancerdetectioninmagneticresonanceimaging AT nicholasrpayne swarmlearningwithweaksupervisionenablesautomaticbreastcancerdetectioninmagneticresonanceimaging AT lorenaescuderosanchez swarmlearningwithweaksupervisionenablesautomaticbreastcancerdetectioninmagneticresonanceimaging AT paulchristophevaroutas swarmlearningwithweaksupervisionenablesautomaticbreastcancerdetectioninmagneticresonanceimaging AT sreenathkyathanahally swarmlearningwithweaksupervisionenablesautomaticbreastcancerdetectioninmagneticresonanceimaging AT narminghaffarilaleh swarmlearningwithweaksupervisionenablesautomaticbreastcancerdetectioninmagneticresonanceimaging AT kevinpfeiffer swarmlearningwithweaksupervisionenablesautomaticbreastcancerdetectioninmagneticresonanceimaging AT martaligero swarmlearningwithweaksupervisionenablesautomaticbreastcancerdetectioninmagneticresonanceimaging AT jakobbehner swarmlearningwithweaksupervisionenablesautomaticbreastcancerdetectioninmagneticresonanceimaging AT kamarulaabdullah swarmlearningwithweaksupervisionenablesautomaticbreastcancerdetectioninmagneticresonanceimaging AT georgiosapostolakos swarmlearningwithweaksupervisionenablesautomaticbreastcancerdetectioninmagneticresonanceimaging AT chrysafoulakolofousi swarmlearningwithweaksupervisionenablesautomaticbreastcancerdetectioninmagneticresonanceimaging AT antrikleanthous swarmlearningwithweaksupervisionenablesautomaticbreastcancerdetectioninmagneticresonanceimaging AT michailkalogeropoulos swarmlearningwithweaksupervisionenablesautomaticbreastcancerdetectioninmagneticresonanceimaging AT cristinarossi swarmlearningwithweaksupervisionenablesautomaticbreastcancerdetectioninmagneticresonanceimaging AT sylwianowakowska swarmlearningwithweaksupervisionenablesautomaticbreastcancerdetectioninmagneticresonanceimaging AT alexandraathanasiou swarmlearningwithweaksupervisionenablesautomaticbreastcancerdetectioninmagneticresonanceimaging AT raquelperezlopez swarmlearningwithweaksupervisionenablesautomaticbreastcancerdetectioninmagneticresonanceimaging AT ritsemann swarmlearningwithweaksupervisionenablesautomaticbreastcancerdetectioninmagneticresonanceimaging AT wouterveldhuis swarmlearningwithweaksupervisionenablesautomaticbreastcancerdetectioninmagneticresonanceimaging AT juliacamps swarmlearningwithweaksupervisionenablesautomaticbreastcancerdetectioninmagneticresonanceimaging AT volkmarschulz swarmlearningwithweaksupervisionenablesautomaticbreastcancerdetectioninmagneticresonanceimaging AT markuswenzel swarmlearningwithweaksupervisionenablesautomaticbreastcancerdetectioninmagneticresonanceimaging AT sergeymorozov swarmlearningwithweaksupervisionenablesautomaticbreastcancerdetectioninmagneticresonanceimaging AT alexanderciritsis swarmlearningwithweaksupervisionenablesautomaticbreastcancerdetectioninmagneticresonanceimaging AT christianekuhl swarmlearningwithweaksupervisionenablesautomaticbreastcancerdetectioninmagneticresonanceimaging AT fionajgilbert swarmlearningwithweaksupervisionenablesautomaticbreastcancerdetectioninmagneticresonanceimaging AT danieltruhn swarmlearningwithweaksupervisionenablesautomaticbreastcancerdetectioninmagneticresonanceimaging AT jakobnikolaskather swarmlearningwithweaksupervisionenablesautomaticbreastcancerdetectioninmagneticresonanceimaging |