Detalle de: High-resolution dataset of core needle prostate biopsies with expert pathologist annotations
| Title | High-resolution dataset of core needle prostate biopsies with expert pathologist annotations |
| Abstract |
We present an expert-labelled image dataset designed to support automated analysis of image-guided prostate core needle biopsies (CNBs) in the context of prostate cancer diagnosis. The dataset includes 37 whole slide images (WSIs), corresponding to 60 CNBs stained with hematoxylin and eosin (H&E), from 32 patients. The dataset contains 15480 manual expert annotations, covering benign glands, tumor, stromal tissue, adipose tissue, nerves, nerve ganglia, blood vessels, muscle, and inflammatory cells. Additional recorded features include stromal retractions, normal and abnormal glandular secretions, and tissue or digital image artifacts. These annotations represent the normal findings that pathologists encounter in routine diagnosis. The annotations were performed by two experienced uropathologists with final labels determined by consensus; immunohistochemical techniques were used as guidance. This dataset provides a foundation for training and benchmarking machine learning models for prostate biopsy assessment and is openly available to facilitate reproducibility and further research. |
| Apa Reference | Calapaquí Terán, A.K., González Bernad, A.A., Sánchez Magdaleno, L., Marcos González, S., Delgado Bolton, R.C., Moustafá Calvo, J., Cobo Cano, M., Lloret Iglesias, L. & Gómez Román, J.J. (2026). High-resolution dataset of core needle prostate biopsies with expert pathologist annotations. Scientific Data. 10.1038/s41597-026-08349-y |
| Doi | 10.1038/s41597-026-08349-y |
| Publication Date | 22 Sep 2026 |
| Publication Type | Article |
| Researcher | Lara Lloret Iglesias |
| Ontology | Artificial Intelligence, Machine Learning in Biology, Benchmarking |
| Research Groups | |
| Tools |
