EUVIMEDEUVIMED entdecken
Open Access · CC BY-NC 4.0

Molecular subtyping of breast cancer from routine Hematoxylin & Eosin-stained histology images using deep learning

Mehdi Zarei, Masoud Arabfard, Mehdi Raei, Farrokh Modarresi

Cancer Treatment and Research Communications · 2026 · 49 · 101439

DOI: 10.1016/j.ctarc.2026.101439 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Cancer Treatment and Research Communications, 2026. Verfasst von Mehdi Zarei und 3 weiteren Autor:innen. Beteiligte Einrichtung: Student Research Committee, Baqiyatallah University of Medical Sciences, Tehran, Iran. Die Arbeit nennt 45 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY-NC 4.0).

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

INTRODUCTION: Breast cancer is a heterogeneous malignancy comprising distinct molecular subtypes, each with varying therapeutic strategies and prognoses. Conventional molecular subtyping relies on immunohistochemistry and in situ hybridization studies, which may not be universally accessible due to financial and technical limitations. This study evaluates the utility of deep learning for identifying breast cancer molecular subtypes using histopathological images. MATERIALS AND METHODS: A dataset of 1000 H&E-stained breast cancer images (× 40 magnification) was collected and labeled into four molecular subtypes, and a modified EfficientNetB0 architecture was employed. The model was trained using the Adam optimizer with categorical cross-entropy loss over 25 epochs. Model performance was evaluated on a separate test set using accuracy, AUC, and classification metrics. RESULTS: The model reached an overall classification accuracy of 0.81. Luminal A and Basal-like subgroups achieved the best diagnostic performance. The highest F1-score was observed for Luminal A (0.93), followed by Basal-like (0.86), Luminal B (0.76), and HER2-enriched (0.69). In an independent external validation cohort of 50 patients from a separate hospital, the model achieved an overall accuracy of 0.78. CONCLUSION: Our results demonstrate that deep learning-based image analysis can effectively predict breast cancer molecular subtypes from routine histopathological images, offering a valuable adjunct when IHC and molecular tests are unavailable, inconclusive, or cost-prohibitive. Furthermore, this approach could serve as a quality control mechanism by flagging discordant cases where histological prediction and molecular tests are not compatible. The findings provide proof-of-concept evidence that H&E histopathology images contain information that can be used by deep learning to predict breast cancer molecular subtypes.

Schlagwörter

Stichwörter der Autor:innen

Breast cancerConvolutional neural networksDeep learningDigital pathologyHistopathologyImage analysisMolecular classification

Autor:innen und Institutionen

  • Student Research Committee, Baqiyatallah University of Medical Sciences, Tehran, Iran.
  • Artificial Intelligence in Health Research Center, Biomedicine Technologies Institute, Baqiyatallah University of Medical Sciences, Tehran, Iran.
  • Baqiyatallah Research Center for Gastroenterology and Liver Diseases (BRCGL), Baqiyatallah University of Medical Sciences, Tehran, Iran; Faculty of Medicine, Baqiyatallah University of Medical Sciences, Tehran, Iran. El…

Zitieren

APA

Zarei, M., Arabfard, M., Raei, M., & Modarresi, F. (2026). Molecular subtyping of breast cancer from routine Hematoxylin & Eosin-stained histology images using deep learning. Cancer Treatment and Research Communications, 49, 101439. https://doi.org/10.1016/j.ctarc.2026.101439

Vancouver

Zarei M, Arabfard M, Raei M, Modarresi F. Molecular subtyping of breast cancer from routine Hematoxylin & Eosin-stained histology images using deep learning. Cancer Treatment and Research Communications. 2026;49:101439. doi:10.1016/j.ctarc.2026.101439

BibTeX
@article{Zarei2026,
  author = {Zarei, Mehdi and Arabfard, Masoud and Raei, Mehdi and Modarresi, Farrokh},
  title = {Molecular subtyping of breast cancer from routine Hematoxylin & Eosin-stained histology images using deep learning},
  journal = {Cancer Treatment and Research Communications},
  year = {2026},
  volume = {49},
  pages = {101439},
  doi = {10.1016/j.ctarc.2026.101439}
}
Kostenlos starten →