HistoBreastCAD: an integrated deep learning framework for segmentation and multiclass classification of whole-slide breast histopathology images
4 affiliations
- 1Departamento de Ciencias de La Computación, Universidad Técnica Particular de Loja, Loja, Ecuador
- 2Departamento de Química y Ciencias Exactas, Universidad Técnica Particular de Loja, Loja, Ecuador
- 3Corporación Médica Monte Sinaí-CIPAM (Centro Integral de Patología Mamaria) Cuenca-Ecuador, Facultad de Ciencias Médicas, Universidad de Cuenca, Cuenca, Ecuador
- 4Departamento de Ciencias de La Salud, Universidad Técnica Particular de Loja, Loja, Ecuador
Abstract
Breast cancer remains a major cause of cancer-related morbidity and mortality worldwide, and histopathological examination remains the diagnostic gold standard. This study presents HistoBreastCAD, an integrated deep learning framework that combines tissue-region segmentation, four-class histopathology classification, spatial mapping of predictions, and interactive whole-slide image (WSI) review. Segmentation was evaluated using WSI-grouped five-fold cross-validation, while classification used source-image-grouped five-fold cross-validation to prevent source-level data leakage. The final DeepLabV3+ L0 segmentation configuration achieved an out-of-fold micro-Dice of 0.8390, IoU of 0.7226, precision of 0.8278, and recall of 0.8505. For classification, EfficientNetB0 was the strongest individual model, while a frozen equal-weight soft-voting ensemble of EfficientNetB0 and ResNet18-GAP-attention achieved an accuracy of 0.8283 ± 0.0320 and macro-F1 of 0.8282 ± 0.0316, outperforming EfficientNetB0 in all five validation folds. An exploratory reader study using the 100 microscopy images from the held-out BACH test set showed moderate interobserver agreement between two pathologists (Cohen’s κ = 0.5105; n = 95), whereas model-to-reader concordance was lower. The ensemble did not increase reader concordance relative to EfficientNetB0. HistoBreastCAD therefore demonstrates the feasibility of integrating established deep learning components into a source-separated WSI-oriented digital pathology workflow. However, external multi-institutional validation, adjudicated reader studies, and formal explainability assessment are required before claims of clinical generalizability or utility can be supported.
Related
- Phenotypic Profiling of SOD1-G93A Mice: Integrating Sciatic Nerve Histopathology with Systemic Chemokine Dynamics During ALS Progression
- Histopathology and spatial transcriptomics jointly map myofiber-specific pathological programs in mTORC1-driven myopathy
- Large Language Models for Structured Information Extraction from German Histopathology Reports - Hepatocellular Carcinoma
- MambaGLA-NPC: a 3D framework integrating CNN and state space models for robust multi-center nasopharyngeal carcinoma segmentation
- Segmentation-Less 3D Quantitative Image Analysis of Tissue Architecture with Application to the Localization of Organelles in MDCK Cysts
BibTeX
@article{celivega2026histobreastcad,
title = {{HistoBreastCAD: an integrated deep learning framework for segmentation and multiclass classification of whole-slide breast histopathology images}},
author = {Oswaldo Celi-Vega and Yuliana Jiménez-Gaona and Patricio Corral-Dominguez and Ángel Gutiérrez and Zaida Socola-Barzallo},
journal = {Frontiers in Bioinformatics},
year = {2026},
month = oct,
doi = {10.3389/fbinf.2026.1870283},
url = {https://doi.org/10.3389/fbinf.2026.1870283},
note = {CC BY 4.0},
}
RIS
TY - JOUR
TI - HistoBreastCAD: an integrated deep learning framework for segmentation and multiclass classification of whole-slide breast histopathology images
AU - Oswaldo Celi-Vega
AU - Yuliana Jiménez-Gaona
AU - Patricio Corral-Dominguez
AU - Ángel Gutiérrez
AU - Zaida Socola-Barzallo
PY - 2026
DA - 2026/10/13
JO - Frontiers in Bioinformatics
DO - 10.3389/fbinf.2026.1870283
UR - https://doi.org/10.3389/fbinf.2026.1870283
AB - Breast cancer remains a major cause of cancer-related morbidity and mortality worldwide, and histopathological examination remains the diagnostic gold standard. This study presents HistoBreastCAD, an integrated deep learning framework that combines tissue-region segmentation, four-class histopathology classification, spatial mapping of predictions, and interactive whole-slide image (WSI) review. Segmentation was evaluated using WSI-grouped five-fold cross-validation, while classification used source-image-grouped five-fold cross-validation to prevent source-level data leakage. The final DeepLabV3+ L0 segmentation configuration achieved an out-of-fold micro-Dice of 0.8390, IoU of 0.7226, precision of 0.8278, and recall of 0.8505. For classification, EfficientNetB0 was the strongest individual model, while a frozen equal-weight soft-voting ensemble of EfficientNetB0 and ResNet18-GAP-attention achieved an accuracy of 0.8283 ± 0.0320 and macro-F1 of 0.8282 ± 0.0316, outperforming EfficientNetB0 in all five validation folds. An exploratory reader study using the 100 microscopy images from the held-out BACH test set showed moderate interobserver agreement between two pathologists (Cohen’s κ = 0.5105; n = 95), whereas model-to-reader concordance was lower. The ensemble did not increase reader concordance relative to EfficientNetB0. HistoBreastCAD therefore demonstrates the feasibility of integrating established deep learning components into a source-separated WSI-oriented digital pathology workflow. However, external multi-institutional validation, adjudicated reader studies, and formal explainability assessment are required before claims of clinical generalizability or utility can be supported.
SN - 2673-7647
ER -