Automatic Prostate Cancer Classification in WSIs Using EfficientNet, Ensemble Learning, and Ordinal Loss Modeling

• Woshington V. S. Rodrigues IFPI
• Antonio O. C. Filho UFPI
• Armando L. Borges UFPI
• Jose D. Araújo UFPI
• João O. B. Diniz IFMA

DOI: https://doi.org/10.5753/sbcas.2026.20665

Abstract

Manual histopathological evaluation of prostate cancer in WSIs is subject to interobserver and intraobserver variability. To mitigate this problem, this work proposes the automatic classification of ISUP grade groups using Convolutional Neural Networks. The methodology used the PANDA dataset, applying entropy-based noise reduction and patch extraction. The EfficientNet models (B0, B3, and B7) were trained with the Ordinal Focal Loss function to preserve disease progression and mitigate class imbalance. Finally, the individual predictions were combined into an ensemble. The simple averaging strategy achieved the best performance, with a Quadratic Weighted Kappa of 0.879 and an accuracy of 0.698. The approach preserved the ordinal integrity of the grades, rendering extreme errors nearly nonexistent and demonstrating its clinical viability.

References

Afifi, A., Alrahmawy, M., and El-Baz, A. (2024). Prostate cancer detection using deep learning models on histopathological image slides: An experimental analysis. International Journal of Advanced Computer Science and Applications, 15(1):xxx–xxx.

Albahri, O. S., Albahri, A. S., Mohammed, K. I., Zaidan, A., Zaidan, B. B., Hashim, M., Salman, O. H., Alaa, M., and Alsalem, M. A. (2022). Expert system research for prostate cancer: A literature review. Computers in Biology and Medicine, 145:105487.

Ali, I., Khan, A., Khan, M., Ahmad, R., and Ullah, I. (2023). Shannon entropy in artificial intelligence and its applications based on information theory. Entropy, 25(2):364.

Alici-Karaca, D. and Akay, B. (2024). An efficient deep learning model for prostate cancer diagnosis. IEEE Access, 12:150776–150790.

Araújo, F. H., Silva, R. R., Medeiros, F. N., Neto, J. F. R., Oliveira, P. H. C., Bianchi, A. G. C., and Ushizima, D. (2021). Active contours for overlapping cervical cell segmentation. International Journal of Biomedical Engineering and Technology, 35(1):70–92.

Bray, F., Laversanne, M., Sung, H., Ferlay, J., Siegel, R. L., Soerjomataram, I., and Jemal, A. (2024). Global cancer statistics 2022: Globocan estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA: A Cancer Journal for Clinicians, 74(3):229–263.

Bulten, W., Pinckaers, H., van Boven, H., Vink, R., de Bel, T., van Ginneken, B., van der Laak, J., Hulsbergen-van de Kaa, C., and Litjens, G. (2020). Automated deep-learning system for Gleason grading of prostate cancer using biopsies: a diagnostic study. The Lancet Oncology, 21(2):233–241.

Campanella, G., Hanna, M. G., Geneslaw, L., Miraflor, A., Silva, V. W. K., Busam, K. J., Brogi, E., Reuter, V. E., Klimstra, D. S., and Fuchs, T. J. (2019). Clinical-grade computational pathology using weakly supervised deep learning on whole slide images. Nature Medicine, 25(9):1301–1309.

Chen, X. (2026). A transfer learning-based deep focal multiclass network for psychological emotion recognition in community-correction populations. Alexandria Engineering Journal, 135:235–242.

Cohen, J. (1968). Weighted kappa: Nominal scale agreement with provision for scaled disagreement or partial credit. Psychological Bulletin, 70(4):213–220.

Dina, A. S., Siddique, A., and Manivannan, D. (2023). A deep learning approach for intrusion detection in internet of things using focal loss function. Internet of Things, 22:100699.

Epstein, J. I., Egevad, L., Amin, M. B., Delahunt, B., Srigley, J. R., and Humphrey, P. A. (2016). The 2014 International Society of Urological Pathology (ISUP) consensus conference on gleason grading of prostatic carcinoma: Definition of grading patterns and proposal for a new grading system. The American Journal of Surgical Pathology, 40(2):244–252.

Kosoko, I., Garg, A., Jain, S., and Hewage, P. (2024). Leveraging deep learning and explainable ai for diagnosis of prostate cancer. In Proceedings of the International Conference on Applied Artificial Intelligence in Medical Imaging, Bolton, United Kingdom.

Lin, T.-Y., Goyal, P., Girshick, R., He, K., and Dollár, P. (2017). Focal loss for dense object detection. In Proceedings of the IEEE International Conference on Computer Vision (ICCV), pages 2980–2988.

PANDA Challenge (2020). Prostate cANcer grade assessment challenge. [link]. Accessed: 2024-01-01.

Siegel, R. L., Giaquinto, A. N., and Jemal, A. (2024). Cancer statistics, 2024. CA: A Cancer Journal for Clinicians, 74(1):12–49.

Xiang, J., Wang, X., Wang, X., Zhang, J., Yang, S., Yang, W., Han, X., and Liu, Y. (2023). Automatic diagnosis and grading of prostate cancer with weakly supervised learning on whole slide images. Computers in Biology and Medicine, 152:106340.

Zhou, Z.-H. (2025). Ensemble methods: foundations and algorithms. Chapman and Hall/CRC.