A Multicentric Dataset for Training and Benchmarking Breast Cancer Segmentation in H&E Slides
Fuente:
arXiv
Guardado en:
| Autores principales: | , , , , , , , , , , , , , , , , , , , , |
|---|---|
| Formato: | Preprint |
| Publicado: |
2025
|
| Materias: | |
| Acceso en línea: | |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| _version_ | 1866914071718133760 |
|---|---|
| author | Lems, Carlijn Tessier, Leslie Bokhorst, John-Melle van Rijthoven, Mart Aswolinskiy, Witali Pozzi, Matteo Klubickova, Natalie Dintzis, Suzanne Campora, Michela Balkenhol, Maschenka Bult, Peter Spronck, Joey Detone, Thomas Barbareschi, Mattia Munari, Enrico Bogina, Giuseppe Wesseling, Jelle Lips, Esther H. Ciompi, Francesco Meeuwsen, Frédérique van der Laak, Jeroen |
| author_facet | Lems, Carlijn Tessier, Leslie Bokhorst, John-Melle van Rijthoven, Mart Aswolinskiy, Witali Pozzi, Matteo Klubickova, Natalie Dintzis, Suzanne Campora, Michela Balkenhol, Maschenka Bult, Peter Spronck, Joey Detone, Thomas Barbareschi, Mattia Munari, Enrico Bogina, Giuseppe Wesseling, Jelle Lips, Esther H. Ciompi, Francesco Meeuwsen, Frédérique van der Laak, Jeroen |
| contents | Automated semantic segmentation of whole-slide images (WSIs) stained with hematoxylin and eosin (H&E) is essential for large-scale artificial intelligence-based biomarker analysis in breast cancer. However, existing public datasets for breast cancer segmentation lack the morphological diversity needed to support model generalizability and robust biomarker validation across heterogeneous patient cohorts. We introduce BrEast cancEr hisTopathoLogy sEgmentation (BEETLE), a dataset for multiclass semantic segmentation of H&E-stained breast cancer WSIs. It consists of 587 biopsies and resections from three collaborating clinical centers and two public datasets, digitized using seven scanners, and covers all molecular subtypes and histological grades. Using diverse annotation strategies, we collected annotations across four classes - invasive epithelium, non-invasive epithelium, necrosis, and other - with particular focus on morphologies underrepresented in existing datasets, such as ductal carcinoma in situ and dispersed lobular tumor cells. The dataset's diversity and relevance to the rapidly growing field of automated biomarker quantification in breast cancer ensure its high potential for reuse. Finally, we provide a well-curated, multicentric external evaluation set to enable standardized benchmarking of breast cancer segmentation models. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2510_02037 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | A Multicentric Dataset for Training and Benchmarking Breast Cancer Segmentation in H&E Slides Lems, Carlijn Tessier, Leslie Bokhorst, John-Melle van Rijthoven, Mart Aswolinskiy, Witali Pozzi, Matteo Klubickova, Natalie Dintzis, Suzanne Campora, Michela Balkenhol, Maschenka Bult, Peter Spronck, Joey Detone, Thomas Barbareschi, Mattia Munari, Enrico Bogina, Giuseppe Wesseling, Jelle Lips, Esther H. Ciompi, Francesco Meeuwsen, Frédérique van der Laak, Jeroen Quantitative Methods Computer Vision and Pattern Recognition Image and Video Processing Automated semantic segmentation of whole-slide images (WSIs) stained with hematoxylin and eosin (H&E) is essential for large-scale artificial intelligence-based biomarker analysis in breast cancer. However, existing public datasets for breast cancer segmentation lack the morphological diversity needed to support model generalizability and robust biomarker validation across heterogeneous patient cohorts. We introduce BrEast cancEr hisTopathoLogy sEgmentation (BEETLE), a dataset for multiclass semantic segmentation of H&E-stained breast cancer WSIs. It consists of 587 biopsies and resections from three collaborating clinical centers and two public datasets, digitized using seven scanners, and covers all molecular subtypes and histological grades. Using diverse annotation strategies, we collected annotations across four classes - invasive epithelium, non-invasive epithelium, necrosis, and other - with particular focus on morphologies underrepresented in existing datasets, such as ductal carcinoma in situ and dispersed lobular tumor cells. The dataset's diversity and relevance to the rapidly growing field of automated biomarker quantification in breast cancer ensure its high potential for reuse. Finally, we provide a well-curated, multicentric external evaluation set to enable standardized benchmarking of breast cancer segmentation models. |
| title | A Multicentric Dataset for Training and Benchmarking Breast Cancer Segmentation in H&E Slides |
| topic | Quantitative Methods Computer Vision and Pattern Recognition Image and Video Processing |
| url | https://arxiv.org/abs/2510.02037 |