EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Automatic Grade Classification in Prostate Histopathological Images Using EfficientNet and Ordinal Focal Loss

Woshington Valdeci de Sousa Rodrigues iD, Armando Luz iD, José Denes Lima Araújo iD, João Diniz iD, Antonio Oseas Filho iD

Bioengineering · 26.04.2026 · 13 (5) · 503

DOI: 10.3390/bioengineering13050503 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Bioengineering, online erschienen am 26.04.2026. Verfasst von Woshington Valdeci de Sousa Rodrigues und 4 weiteren Autor:innen. Beteiligte Einrichtung: Instituto Federal do Piauí (IFPI), Campus Picos, Picos 64605-500, PI, Brazil. Die Arbeit wurde bisher 1-mal zitiert (Crossref) und nennt 28 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

The automatic classification of ISUP (International Society of Urological Pathology) grade groups in prostate histopathological images remains challenging due to the high similarity between adjacent classes, class imbalance, and label noise. In this work, we propose a deep learning pipeline based on EfficientNet convolutional neural networks combined with a hybrid loss function that integrates ordinal regression and Focal Loss to better capture the ordered nature of ISUP grades. A noise-filtering strategy based on the entropy of predictions from multiple EfficientNet models was first applied to identify and remove high-uncertainty samples from the training set. The problem was then reformulated as an ordinal regression task to explicitly model the hierarchical relationship among grades. Experiments conducted on the PANDA dataset demonstrate that removing noisy samples improved performance from κ=0.826 to κ=0.833. Incorporating ordinal loss further increased performance to κ=0.851. The best configuration, combining ordinal regression and Focal Loss, achieved κ=0.857 and an accuracy of 0.669, while reducing severe misclassifications and concentrating errors among adjacent classes. These results indicate that explicitly modeling ordinal structure and mitigating label noise are effective strategies for improving prostate cancer grading systems.

Schlagwörter

Stichwörter der Autor:innen

EfficientNetPANDA datasetdigital pathologyfocal losshistopathological image classificationordinal regressionprostate cancer

Autor:innen und Institutionen

  • Instituto Federal do Piauí (IFPI), Campus Picos, Picos 64605-500, PI, Brazil.
  • Universidade Federal do Piauí (UFPI), Campus Picos, Picos 64607-760, PI, Brazil.
  • Fábrica de Inovação, Instituto Federal do Maranhão-Campus Grajaú, Grajaú 65940-000, MA, Brazil.

ORCID-Profile: Woshington Valdeci de Sousa Rodrigues, Armando Luz, José Denes Lima Araújo, João Diniz, Antonio Oseas Filho

Zitieren

APA

Rodrigues, W. V. S., Luz, A., Lima Araújo, J. D., Diniz, J., & Filho, A. O. (2026). Automatic Grade Classification in Prostate Histopathological Images Using EfficientNet and Ordinal Focal Loss. Bioengineering, 13(5), 503. https://doi.org/10.3390/bioengineering13050503

Vancouver

Rodrigues WVS, Luz A, Lima Araújo JD, Diniz J, Filho AO. Automatic Grade Classification in Prostate Histopathological Images Using EfficientNet and Ordinal Focal Loss. Bioengineering. 2026;13(5):503. doi:10.3390/bioengineering13050503

BibTeX
@article{Rodrigues2026,
  author = {Rodrigues, Woshington Valdeci de Sousa and Luz, Armando and Lima Araújo, José Denes and Diniz, João and Filho, Antonio Oseas},
  title = {Automatic Grade Classification in Prostate Histopathological Images Using EfficientNet and Ordinal Focal Loss},
  journal = {Bioengineering},
  year = {2026},
  volume = {13},
  number = {5},
  pages = {503},
  doi = {10.3390/bioengineering13050503}
}
Kostenlos starten →